A309 === Subject: Re: 1972 Garrett 2000 RPN Calculator That is an awesome calculator!. I thought I was a calculator geek but have never heard of a Garrett. My earliest calc was a Sharp 4-function handheld with a nixie tube read out that I purchased in 1970 for $300. It was traded in for a newer Sharp model that was the same but had 1 memory locatio. I gave it to my father-in-law many years ago. I have often wondered if he still has it. It would be a collector's item as this Garrett is. I discovered HP calcs when a professor used an HP 35 in class. I immediately called the Orlando, FL regional office. They had a newly released HP 45 in stock. I drove down and purchased it for $335. It worked perfectly as recently as 1990 when it was consumed in a house fire. :( > I am selling my 1972 Garrett 2000 Calculator on eBay. > This is the first non-HP calculator to run RPN. > http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=5788184270 I am still using a 45 as my work calculator. It is on my desk and I use it for anything not too demanding. However it is older than me. I wish I had known all the excitement of having one memory on a calculator! Arnaud === Subject: Re: 1972 Garrett 2000 RPN Calculator To understand the excitement, you would have to consider the times. Calculators consisted of mechanical adding machines, slide rules, and very large and heavy Mechanical calculators. The rotary calculators were glorified adding machines. Big and heavy doesn't do them justice. They would rumble and perculate doing a calcuation for sometimes minutes. Very unreliable and maintenance heavy. Slide rules were great. I still have a couple. I just wish my vision was still good enough to use them. I still remember the fresman class I attended on how to use the things. It was required for all engineering students at the time. I have always wanted a Curta coffee grinder mechanical calculator. They are an engineering marvel. They were sold out of scientific/technical magazine until sometime in the 70s. I would still like to have one but the prices have gone into the stratosphere. I was a tech rep with Xerox in the early 70s. One of the managers had a very early calc with the innovative liquid crystal display. He purchased it because it was easier to see. You wouldn't believe how slow the display was. You had to watch and wait for it to completely change. I almost forgot. We also had a small computer in the science office. It loaded math programs from a computer grade tape a little smaller than a cassette. The programs would take a couple of minutes to load. (The name will probably come to me just after I compose this message. A senior moment. :) ) It cost $17,000.00 in early 70s dollars. Do the math. I built an Altair in 75 or so. It was programmed by flipping switches. I then purchased my first store bought computer. A TRS-80 Model one for $895 - 4 k memory. I soon upgraded it to Level II and 16K memory for an additional $800. Those were the days! Jeff > I am still using a 45 as my work calculator. It is on my desk and I use it > for anything not too demanding. However it is older than me. I wish I had > known all the excitement of having one memory on a calculator! === Subject: Re: 1972 Garrett 2000 RPN Calculator Bugger - I would have paid thousands for one of those, but alas, you don't ship to New Zealand. Oh well - never mind. > I am selling my 1972 Garrett 2000 Calculator on eBay. > This is the first non-HP calculator to run RPN. > http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=5788184270 === Subject: Re: 1972 Garrett 2000 RPN Calculator >Bugger - I would have paid thousands for one of those, but alas, you don't >ship to New Zealand. Oh well - never mind. If you're really willing to pay thousands for it, why not get a plane ticket to LA to go and pick it up yourself? Such a plane ticket should cost less than thousands, and then you'll still have hundreds left with which to pay the item. >> I am selling my 1972 Garrett 2000 Calculator on eBay. >> This is the first non-HP calculator to run RPN. >> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=5788184270 -- Paul Schlyter, Grev Turegatan 40, SE-114 38 Stockholm, SWEDEN e-mail: pausch at saaf dot se WWW: http://stjarnhimlen.se/ === Subject: Re: 1972 Garrett 2000 RPN Calculator >Bugger - I would have paid thousands for one of those, but alas, you don't >ship to New Zealand. Oh well - never mind. Don't blame me - blame eBay. I checked every will ship to box that was available. Needless to say, I will indeed ship to New Zealand. === Subject: Re: 1972 Garrett 2000 RPN Calculator > Don't blame me - blame eBay. I checked every will ship to box that > was available. Needless to say, I will indeed ship to New Zealand. Hmmm - how about the box that says: Ships to [ ] WORLDWIDE? Seems to work for others. === Subject: Re: 1972 Garrett 2000 RPN Calculator >> Don't blame me - blame eBay. I checked every will ship to box that >> was available. Needless to say, I will indeed ship to New Zealand. >Hmmm - how about the box that says: Ships to [ ] WORLDWIDE? >Seems to work for others. It's times like these that I am glad I save the screen at each step when I do something on eBay. When I set up the auction, there was no worldwide check box. Now there is! I just fixed it. I sure am getting a lot of watchers and page hits; too bad none of then has resulted in a bid... Either I overpriced it or they are all smart enough to know that bidding your highest offer at the last minute is the obtimum strategy... === Subject: Re: 1972 Garrett 2000 RPN Calculator posting-account=boybFQ0AAACcJkbp4Mpsq2-jhOTZrn5e :D me too === Subject: Clock Is Ticking Toward HHC2005 in Chicago posting-account=Go2g5wsAAAAsfBzIGOikqm0lzTSUDSrr Just a reminder that HHC2005, this year's U.S. HP Calculator Conference is approaching. The conference will be held on Saturday and Sunday, September 17th and 18th in Franklin Park, Illinois, a suburb of Chicago, virtually next door to the O'Hare Airport. Check out all the details at http://holyjoe.net/hhc2005/ on the web. Already approximately thirty people have indicated their strong intention to attend, and we're expecting some interesting presentations from the HP Calculator group, headquartered in San Diego. These affairs have been held since the first one in Santa Clara, California in the Summer of 1979, where we celebrated the newly-released HP41C. They've all been interesting, and there's usually a surprise or two which makes each one memorable. They are still soliciting speakers on all topics related to HP handhelds, so please check out the details and volunteer to present your interesting material to the group. Hope to see you in a couple of months.... Jake Schwartz jakes magpage com or jakes pahhc org === Subject: Re: Clock Is Ticking Toward HHC2005 in Chicago posting-account=ZkGXSAwAAABowQ-go75CnX8ACDw_z9sQ > Check out all the details at http://holyjoe.net/hhc2005/ on the web. As an outside observer (I don't have any dogs in this fight), I am curious if Joe Horn's sale of $100 49G+ calculators was ever resolved? I know he had some back injury issues (and a harddrive failure too?), and that was keeping him from responding to folks that hadn't received their calculators. Has this been resolved, or will Joe have some 'splainen (explaining) to do at the conference? To me, this saga has been an intriguing series of events; it's the stuff movies are made from. It reinforces the notion that truth is stranger than fiction. Curious about the ending, mk === Subject: Re: Clock Is Ticking Toward HHC2005 in Chicago > ...As an outside observer (I don't have any dogs in this fight), I am > curious if Joe Horn's sale of $100 49G+ calculators was ever resolved?... > Curious about the ending, > mk On my front, no it has not been resolved, no goods or any dialog. I would be very interested in what really transpired. Best wishes to all HHC2005 goers Rick === Subject: Re: Clock Is Ticking Toward HHC2005 in Chicago posting-account=msh5ow0AAACgGZihasI14P8AUGVn2RWK Sounds good. The CHIP group here is looking forward to it, and Joe Horn has been keeping the web page up to date. I'm well into preparing a talk - a combination of a practical application and programming/user interface techniques - and look forward to HP's participation. Brian Walsh > Just a reminder that HHC2005, this year's U.S. HP Calculator Conference > is approaching. The conference will be held on Saturday and Sunday, > September 17th and 18th in Franklin Park, Illinois, a suburb of > Chicago, virtually next door to the O'Hare Airport. Check out all the > details at http://holyjoe.net/hhc2005/ on the web. Already > approximately thirty people have indicated their strong intention to > attend, and we're expecting some interesting presentations from the HP > Calculator group, headquartered in San Diego. These affairs have been > held since the first one in Santa Clara, California in the Summer of > 1979, where we celebrated the newly-released HP41C. They've all been > interesting, and there's usually a surprise or two which makes each one > memorable. They are still soliciting speakers on all topics related to > HP handhelds, so please check out the details and volunteer to present > your interesting material to the group. > Hope to see you in a couple of months.... > Jake Schwartz > jakes magpage com > or > jakes pahhc org === Subject: Re: Clock Is Ticking Toward HHC2005 in Chicago > Sounds good. The CHIP group here is looking forward to it, and Joe Horn > has been keeping the web page up to date. I'm well into preparing a > talk - a combination of a practical application and programming/user > interface techniques - and look forward to HP's participation. Wild rumors (again) in the air & ear: HPQ will introduce new supermodels (no not chicks, just calcs :) at HHC2005 More information to all participants (after signing a NDA with blood :) at the location. If anyone sues me - I'm out of control and have medication for it... (which seems to increase the symptons so far :) VPN PS: Cyrille - don't try to deny it, you have not been fired therefore you (and the rest of the team) are up to something... Maybe Xpander^2 ? hp 44 Magnum? hp 15c Platinum? or maybe a turbocharged hp 50gx+III++ === Subject: Re: Clock Is Ticking Toward HHC2005 in Chicago : > Sounds good. The CHIP group here is looking forward to it, and Joe Horn : > has been keeping the web page up to date. I'm well into preparing a : > talk - a combination of a practical application and programming/user : > interface techniques - and look forward to HP's participation. : Wild rumors (again) in the air & ear: : HPQ will introduce new supermodels (no not chicks, just calcs :) : at HHC2005 : More information to all participants (after signing a NDA with blood :) : at the location. : If anyone sues me - I'm out of control and have medication for it... : (which seems to increase the symptons so far :) : VPN : PS: Cyrille - don't try to deny it, you have not been fired : therefore you (and the rest of the team) are up to something... : Maybe Xpander^2 ? hp 44 Magnum? hp 15c Platinum? : or maybe a turbocharged hp 50gx+III++ I'm hoping for a new version of the 42s (43s?)!! I don't need a high end calculator anymore but a 42s replacement that fits into a shirt pocket would be worth buying!! -- ------------------- Keep working millions on welfare depend on you === Subject: Re: Clock Is Ticking Toward HHC2005 in Chicago X > : Wild rumors (again) in the air & ear: X > I'm hoping for a new version of the 42s (43s?)!! I don't need a high end > calculator anymore but a 42s replacement that fits into a shirt pocket > would be worth buying!! OK HP will do this for you, just because you asked Since you are the only person to order it, the cost will be 15 million US dollars The money will be collected from you in 15 working days Happy holidays with you new hp 43s :-D VPN === Subject: Re: Clock Is Ticking Toward HHC2005 in Chicago posting-account=d8X8zw0AAAB_Eo2ldj3HF5DzN42pN3GY Or maybe just trying to expand the market, get more retailers to stock the current calculators and get more teachers using HP calcs. . . TW === Subject: Re: Clock Is Ticking Toward HHC2005 in Chicago > Or maybe just trying to expand the market, get more retailers to stock > the current calculators and get more teachers using HP calcs. . . I wish HP would take calculators seriously again. The only way to crack the market is to follow TI -- give free graphing calcs to math teachers and professors across the country. Give them promotional materials expounding the benefits HP products bring to the classroom. The TI indoctrination is nearly complete. Oh, to count the number of times I've had a professor say, you did that on your TI?. No; I've always stuck with HP. Educators hardly know such things exist. At least they have some vague understanding of Casio's as being inferior. -Daniel === Subject: Re: Clock Is Ticking Toward HHC2005 in Chicago posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt > Casio's as being inferior. Watch it! I have a 3400p which is lovely, abeit a bit feature poor. Adrian === Subject: AD: HP 42S Original Owners Manual on Ebay I had an original HP 42S Owners Manual in the closet. I decided to post it on Ebay. It is in very good, almost new condition. The starting bid is $0.99 plus shipping. I'm just learning the ropes at ebay. I have a couple of other HP items I may post in the future. This link should work if anyone is interested: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=58039&item=5788369627 &rd=1&ssPageName=WDVW === Subject: Re: HP 42S Original Owners Manual on Ebay > I'm just learning the ropes at ebay. I'll give you a little eBay tip ... By restricting your sale to USA residents only you've just excluded around 96% of your potential market - including me :( In addition to possibly missing out on more $$$ you also send a subtle message to many (like me) that says your money isn't good enough. I know that probably wan't your intention, but I just can't get over how many people do this - it just doesn't make any sense. === Subject: Re: HP 42S Original Owners Manual on Ebay Hi Colin, I appreciate and understand your tip. I have purchased a couple of items from Great britain with great satisfaction. This listing was my first and an experiment. I have no idea HOW to ship to foreign countries. I'm sure there must be restrictions, tarriffs, taxes, and all other kinds of government red tape. I once tried a non-ebay transaction with an amateur radio operator in Russia, the money I sent went off into the either. Pleas accept my apology for any slight. I have a few additional HP items that I may list. I will research international sales before listing them. > By restricting your sale to USA residents only you've just excluded around > 96% of your potential market - including me :( > In addition to possibly missing out on more $$$ you also send a subtle > message to many (like me) that says your money isn't good enough. I know > that probably wan't your intention, but I just can't get over how many > people do this - it just doesn't make any sense. === Subject: Re: HP 42S Original Owners Manual on Ebay In message , J. Greer > I have no idea HOW to ship to foreign countries. I'm sure there must >be restrictions, tarriffs, taxes, and all other kinds of government red >tape. You take it to the post office and say: I'd like to ship this to and they'll tell you what to do. There's no red tape at your end (the recipient has that pleasure) nor extra charges or taxes above the shipping cost (again the recipient has that pleasure, especially in Europe :-( ). The most you have to do is fill in a little green customs declaration slip that says whether it is a gift or is commercial and you write on the item's value. The government in the UK then uses this to rip us off by charging sales tax on both the value of the item and the postage for non-gifts even though second-hand items sold between individuals in the UK are exempt from sales tax. :-( -- Bruce Horrocks Surrey England @.plus.com -- fix the obvious for email === Subject: Re: HP 42S Original Owners Manual on Ebay I have been checking into international shipping. It isn't too bad. I am unable to modify the existing ebay listing since it already has a couple of bids. Any HP items I offer in the future will have the international option. Jeff === Subject: Re: HP 42S Original Owners Manual on Ebay Yayyyyy. I've converted just one lost soul - my life now has true meaning :) > I have been checking into international shipping. It isn't too bad. I am > unable to modify the existing ebay listing since it already has a couple > of bids. Any HP items I offer in the future will have the international > option. > Jeff === Subject: Re: HP 42S Original Owners Manual on Ebay <2h3wMUBp6Z0CFwHm@horrocks.plus.com> <4vwAe.9642$xB6.3305@trnddc03> posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Why this calculator is still SOOO expensive? === Subject: Re: HP 42S Original Owners Manual on Ebay > Why this calculator is still SOOO expensive? Possibly one of worlds best every days calculator HP make ever and have old HP quality, Quality you not find on HP to days. If you learn to use it, you find more or less impossible to find usable replacement exept buy new used hp42S to replace - and still prepare to pay 2-3 times more a new price for secure working ex of hp42s.... hp42S have easy used complex capability and complex matrix handling, is very, very usefull for elecrical enginering and impossible to find same functionalty in other pocket shirt sized calculator. (hp32sii and hp33s has very bad, hardworked and incomple complex support an cannot replace hp42s - i try!!) My hp42S is very durable and still working perfecty after 16 year everyday using and not reset or loss of programs after hundreds of drop to floor (from shirtpocket...) and quality of keyboards is exellent and no missing keys after 16 year heavy using. you can try calculator with simulator on PC, linux or Palm-T with free42 http://home.planet.nl/~demun000/thomas_projects/free42/ And now using on my Palm-T for save my old hp42S on a job (but is not a same thing...) If not familar with hp42S, you need reading manual minimum one time for more advanced using - and can buy on CD-rom from http://www.hpmuseum.org/cd/cddesc.htm === Subject: Re: HP 42S Original Owners Manual on Ebay Anyone can buy a HP 42S from me. I'm out of luck and money How about 250 euros? VPN >> Why this calculator is still SOOO expensive? > Possibly one of worlds best every days calculator HP make ever and have > old HP quality, Quality you not find on HP to days. > If you learn to use it, you find more or less impossible to find usable > replacement exept buy new used hp42S to replace - and still prepare to > pay 2-3 times more a new price for secure working ex of hp42s.... > hp42S have easy used complex capability and complex matrix handling, is > very, very usefull for elecrical enginering and impossible to find same > functionalty in other pocket shirt sized calculator. (hp32sii and hp33s > has very bad, hardworked and incomple complex support an cannot replace > hp42s - i try!!) > My hp42S is very durable and still working perfecty after 16 year everyday > using and not reset or loss of programs after hundreds of drop to floor > (from shirtpocket...) and quality of keyboards is exellent and no missing > keys after 16 year heavy using. > you can try calculator with simulator on PC, linux or Palm-T with free42 > http://home.planet.nl/~demun000/thomas_projects/free42/ > And now using on my Palm-T for save my old hp42S on a job > (but is not a same thing...) > If not familar with hp42S, you need reading manual minimum one time > for more advanced using - and can buy on CD-rom from > http://www.hpmuseum.org/cd/cddesc.htm === Subject: Re: HP 42S Original Owners Manual on Ebay <2h3wMUBp6Z0CFwHm@horrocks.plus.com> <4vwAe.9642$xB6.3305@trnddc03> posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE the HP-48 and HP-28 are better :p after almost 20 years my HP-28S and HP-28C are working just fine :p same for my HP-48S/SX/GX :p most are 10-15 years old and still working like new :p these were the old good days of hp === Subject: Re: HP 42S Original Owners Manual on Ebay > the HP-48 and HP-28 are better :p > after almost 20 years my HP-28S and HP-28C are working just fine :p > same for my HP-48S/SX/GX :p most are 10-15 years old and still working > like new :p > these were the old good days of hp I could sell my HP-28S as well? Maybe also 250 euros? VPN (in not so deep trouble, but deep enough with no income) === Subject: Re: HP 42S Original Owners Manual on Ebay > the HP-48 and HP-28 are better :p on desktop yes, also laptop with mathcad/mathlab etc. advance, heavy and multi funktional products, but for every day using and shirt pocket size 'simple' calculator, hp42s is unbeatable IMHO. (special if working with eletrical network, radio or RF-network analalyzer) one example to 'wrong' on hp33s is place of LOG10 and LN-button on hp42s is an LOG10 and 10^x on same button and next button for LN and e^x, is natural to use '2nd' if you want anti-funktion of log, ln etc. but on hp33s LN and e^x place of different buttons and primary and LOG10 and 10^x is a secondary, you need always '2nd' before using 10-logarithm funktion, and guess who type of 'logarithm' I use frequently... Small differences but huge irritate if always use LOG10 and 10^x for example dB calculate in RF-path and signal processing and very few people using 'Neper'(LN, e^x) frequently... > after almost 20 years my HP-28S and HP-28C are working just fine :p > same for my HP-48S/SX/GX :p most are 10-15 years old and still working > like new :p these were the old good days of hp. (But I have trouble om my old hp48G - is hard to starting time to time now if not press on specific place - and read is it knows HW error depend of dry out of plastic foam or simular inside...) Agree - is differenses if companys leadership driven with technical guys like orginator Bill Hewlett and David packard and leaving space to experiments and make new intresting item different from main product line or now scared economic oriented leadership and trash down companies to 'unintresting', high volyme, low quality without new ideas, only optimize on exist products for 3 year working time, not more... latest version of calculator give not new great ideas like high resolution screen and improve software, or remake small competent shirtpocket calculator with high resolution screen as hp42S - only repack of old product in new bigger size and low quality shells and build emulator platform for old programs (and make new emulator errors)... ie only spent money on design and connectivity to replace old memory and RS232-link, not on impprove inside with really good new idea... --- I measure display area on my small mp3-player - this have small screen with high resolution to write sing-name etc. and have backgrund light And double size of this X and Y, fits on standard hp42S display format and have resolution to fit hp49+ display resolution ... 2 times!! Make this!!! === Subject: Re: HP 42S Original Owners Manual on Ebay > the HP-48 and HP-28 are better :p This may be so, but none will EVER Beat the HP-41 series - they were just YEARS ahead of their time. === Subject: Re: HP 42S Original Owners Manual on Ebay >> the HP-48 and HP-28 are better :p > This may be so, but none will EVER Beat the HP-41 series - they were just > YEARS ahead of their time. Though they aren't anymore... I have one and I can't see what the fuss is all about. I find the calc very unconvenient. The 15C is to my taste much better Arnaud === Subject: Re: HP 42S Original Owners Manual on Ebay > the HP-48 and HP-28 are better :p >> This may be so, but none will EVER Beat the HP-41 series - they were just >> YEARS ahead of their time. > Though they aren't anymore... I have one and I can't see what the fuss is > all about. I find the calc very unconvenient. The 15C is to my taste much > better The 15C is a great calculator, but can you add memory modules? Or ROM modules? Or a printer? Or a plotter? Or a card reader? Or a bar code reader? Or a tape drive? Or a diskette drive? Or an RS-232C interface? Or a network interface that lets you attach many of these devices at the same time? The HP-41 can do all that. I'd like to have a 15C, but I wouldn't trade my 41CX for it. HP calculator history is a continuing story of adding great new features while eliminating useful old ones and introducing new problems. I think the perfect balance was reached with the HP-41; many more advantages than disadvantages were introduced, compared with previous models. The 42S added some advantages (smaller size, more memory) but sacrificed the expandability of the 41 series, which IMO makes it a nice pocket calc but certainly not a replacement for the 41. The 48 series added more advantages (RPL, more memory, graphing screen) but also quite a few disadvantages (larger size and more weight, still no compatibility with 41 peripherals, no built-in support for simple RPN programming, lousy color scheme for the G models). I think the advantages still outweighed the disadvantages (especially with the addition of terrific user-developed software like MetaKernel and Erable) but it was rapidly approaching the point of the trade-offs not being worth it. I'm convinced that line was crossed with the 49 series, and I don't think anything that has been added to those (and other recent models like the 33s) can ever make up for what has been lost in terms of quality, reliability, design and appearance. HP would need to take a very large step backwards toward the older designs before I would even consider buying a new model, regardless of any so-called improvements. For instance, I can't image any possible features that could make up for the idiotic keyboard on the 33s. -- Wayne Brown (HPCC #1104) | When your tail's in a crack, you improvise fwbrown@bellsouth.net | if you're good enough. Otherwise you give | your pelt to the trapper. e^(i*pi) = -1 -- Euler | -- John Myers Myers, Silverlock === Subject: Re: HP 42S Original Owners Manual on Ebay fbarbaise@yahoo.fr meinte >the HP-48 and HP-28 are better :p >after almost 20 years my HP-28S and HP-28C are working just fine :p >same for my HP-48S/SX/GX :p most are 10-15 years old and still working >like new :p >these were the old good days of hp Any Laptop with respective Software is even better. But seriously, the HP-48 (and 28) easily outperform the HP-42 and are much more versatile. But the HP-42 wins easily when the form factor combined with reasonable capabilities is an issue. My 41CV, 48SX, 48GX have their place on my desk, while the HP-42 was my companion in the shirt pocket until recently. It was, because it is now replaced by a Christoph Gie§elink, and Robert Hildinger) on it. These simulations/emulation are sufficient for me as I don't have a serious need for a real calculator any longer. One argument you really can't ignore! The answer to all questions in the universe is 42 :-) Gru§ G.9fnter === Subject: Re: HP 42S Original Owners Manual on Ebay > By restricting your sale to USA residents only you've just excluded > around 96% of your potential market - including me :( I've often had success mailing US-only sellers asking them if they'd consider mailing to Europe (Denmark in my case). Many have no problem mailing items outside their stated region of sale. They do probably miss out on the very highest bids though (since they limit their potential bidder pool). I don't know if it's because it's hard or impossible (on eBay) to state exactly which countries you'd accept to ship to, or if some American sellers simply regard the home market sufficient. It *is* also a lot simpler to ship nationally. Steen === Subject: Re: HP 42S Original Owners Manual on Ebay Steen> I don't know if it's because it's hard or impossible (on eBay) Steen> to state exactly which countries you'd accept to ship to, or if Steen> some American sellers simply regard the home market Steen> sufficient. It *is* also a lot simpler to ship nationally. If you are a commercial operation (rather than an individual selling some old stuff out of the closet/attic/basement/etc) you probably need to have an export license to ship internationally. Not everyone wants to deal with that kind of red tape. Then there is the issue of fixed vs. variable shipping charges. But I best most of it is of the that's what everyone else is doing form. :-/ -JimC === Subject: Re: HP 42S Original Owners Manual on Ebay > I've often had success mailing US-only sellers asking them if they'd > consider mailing to Europe (Denmark in my case). Many have no problem > mailing items outside their stated region of sale. In practice that's exactly what I do - and to be honest, it's never a problem. I'm really just playing devil's advocate here - there's a message in the selling game that goes like this ... You can NEVER make it TOO easy for someone to do business with you, and yet time and time again I see auction after auction where so-and-so will only sell to someone in the states - they won't accept Paypal if it's linked to a credit card - then they want to charge $25 for p and p - they won't take returns - and delivery is 10 to 14 days etc etc etc. Personally, if I want something then I'll make a bit of an effort to get it, but I have to wonder just how many people like me read some of the auctions and just think Nahh - too hard. Just my 10c worth :) === Subject: =?iso-8859-1?q?I_need_HELP,_can_someone_take_a_look_at_FluidHP=3F_It=B4s_sooo _slow_on_HP49G?= posting-account=dTw6fA0AAAAmz8yj9fY6u3ScMAxClw0N ,,FluidHP. It«s a programme for water and steam calculations. You will find it if you search for ,,FluiHP at hpcalc.org. It«s the same programme for 48G and 49G. I use a 49G and it«s terrible slow. The 48G tooks about 5 seconds to calculate an ethalpy. The same calculation tooks 15 till more than 20 seconds on the 49G. I don«t know a lot about RPL, but I tried to find out the problem. I read something at the comp.sys.hp48-forum about integers and reals. I opened all programme files in aprox mode and saved them in aprox mode. The programme still running slow. I checked all flags and set all flags, which were set on the 48G, but the programme is still running slow. I would be very happy, if you could have a look at FluidHP and tell me, what«s the reason for this big gap to the 48G performance. I hope that it is easy to see the command or loops which make the programme slow on 49G. Kolja Metz (I«m sorry for my bad english) === Subject: Re: I need HELP, can someone take a look at FluidHP? It«s sooo slow on HP49G ,,FluidHP. It«s a programme for water and steam calculations. You will find it if you search for ,,FluiHP at hpcalc.org. It«s the same programme for 48G and 49G. I use a 49G and it«s terrible slow. The 48G tooks about 5 seconds to calculate an ethalpy. The same calculation tooks 15 till more than 20 seconds on the 49G. I don«t know a lot about RPL, but I tried to find out the problem. I read something at the comp.sys.hp48-forum about integers and reals. I opened all programme files in aprox mode and saved them in aprox mode. The programme still running slow. I checked all flags and set all flags, which were set on the 48G, but the programme is still running slow. I would be very happy, if you could have a look at FluidHP and tell me, what«s the reason for this big gap to the 48G performance. I hope that it is easy to see the command or loops which make the programme slow on 49G. Kolja Metz (I«m sorry for my bad english) I'm sorry, but you can't wait a few seconds for a result? Give me a break! What if you had do it with a slide rule (remember those) or by hand, God forbid! I can't believe these people that expect instant results. It is a calculator for God's sake! Get used to it! === Subject: =?utf-8?q?Re:_I_need_HELP,_can_someone_take_a_look_at_FluidHP=3F_It=B4s_sooo_ slow_on_HP49G?= <11d5i0m11mgce3f@corp.supernews.com> posting-account=dTw6fA0AAAAmz8yj9fY6u3ScMAxClw0N I had to do more than 20 calculations for one power-generation-exercise. It«s not possible to spend 10 on waiting for an answer. Kolja === Subject: =?iso-8859-1?q?Re:_I_need_HELP,_can_someone_take_a_look_at_FluidHP=3F_It=B4s_ sooo_slow_on_HP49G?= posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 I looked at it and it seems to run fine to me. I calculated the examples they have in the manual and none took more than a few seconds on my 49g+. Considering the calculation said it needed 6 seconds or more, and it took me about 3.5, that is a standard speed up for the 49g+ when runinng a userRPL program. Can you give a specific example where it runs slower (the formula, and your input)? You can easily convert the whole directory from ZINTS to REALS at once instead of each individual program. Recall the entire folder to the stack , run ->STR and then set it into approximate mode, then run STR->. Then store the folder in a variable (you can't overwrite the old one though). That should do it. Hoever, I doubt the reals/zints is the problem. I would say off the top of my head that it is CAS issues. When things tend to run slower on the 49g, it is generally becuase it is doing a more exact calculation using symbolics. Set your CAS flags to default by using the command CASCFG. Are you running at least ROM 1.19-6? If not, upgrade to that one. We want to help, but need some more info. TW === Subject: =?utf-8?q?Re:_I_need_HELP,_can_someone_take_a_look_at_FluidHP=3F_It=B4s_sooo_ slow_on_HP49G?= posting-account=dTw6fA0AAAAmz8yj9fY6u3ScMAxClw0N The programme is running slow at every formula. For example h = f(p,t,x) and input 20Mpa, 550¡C and x=-1 took 16 seconds. Every other input took 16 seconds, too. On the 48G of a friend the same formula and input took 5 seconds. The calculator is only a 49G, not a 49G+. It«s running at Version HP49-B Revision #1.19-6 (2001). I already changed all integers to reals. I tried the command CASCFG but the programme is still running slow. Maybe it is a help. There are following libs on the 49G: 291 771 877 991 999 1014 1114 1123 1200 1278 1291 1300 1366 1378 1540 1672 1765. Most of them are engineering programmes. 1123 is „SymbToolZ, 1278 is „Additional functions“ and 1540 is „CALC1“. Kolja Metz === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G Hi. I've taken a look at the library too (now that you're pushing so hard ;-). There are several problems in it, and Tim is right, the biggest one (that is simple to fix) is the CAS on the HP49G. FluidHP very often evaluate an algebraic instead of an RPN stream, so the CAS is involved on the HP49G all the time - this is not the case on the HP48 as it has a much simpler way of handling algebraics. The library is also structured rather poorly and could do with a major overhaul, but I'm always respectful of UserRPL programs of this size - the important thing is that it works though! > The programme is running slow at every formula. For example h = > f(p,t,x) and input 20Mpa, 550C and x=-1 took 16 seconds. Every other > input took 16 seconds, too. On the 48G of a friend the same formula > and input took 5 seconds. I spent about an hour optimizing all the apparent sections where the HP49 CAS will slow things down - basically converting all algebraics into RPN streams and all zints into reals. I also couldn't help but improve the footprint somewhat while I went through the program, reducing the code size by almost 1kB - remember I've made no changes to the actual structure of the program whatsoever! The above changes improved calculation time by more than a factor of 3, meaning that the HP49G is now just as fast as the HP48 (and the HP49G+ is 2.5 times faster still). With a proper rewrite the calculations shouldn't take more than a few milliseconds to complete (global vars are used heavily in the code, they should go for instance). It's really nothing more than a function evaluation each time - that shouldn't take several seconds. Put a rewritten version in a library, and I estimate the size could be brought down to something like 5k instead of 25k - maybe even 2-3k. It would be a couple of days work though for an experienced programmer - you got to ask yourself if it's really worth it? Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > ...basically converting all algebraics into RPN streams and > all zints into reals. I must add that converting zints to reals will not necessarily make the HP49 CAS calculate things faster, and often zints use less space than reals too. The number 540 is found in the code for instance - the zint 540 is 7 bytes while the real 540. is 10.5 bytes. The big performance benefit comes from the algebraic explotion. Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 So do you want to rewrite it or shall I play with it during my calculus class? It does seem like a usefull program for fluids. But the organization is horrible. Just opening it up made my head spin. . . ;-) If you don't want spend a little bit doing it, can you send me what you did so far and I'll work on the interface and organization? TW === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 So do you want to rewrite it or shall I play with it during my calculus class? It does seem like a usefull program for fluids. But the organization is horrible. Just opening it up made my head spin. . . ;-) If you don't want spend a little bit doing it, can you send me what you did so far and I'll work on the interface and organization? TW === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 So do you want to rewrite it or shall I play with it during my calculus class? It does seem like a usefull program for fluids. But the organization is horrible. Just opening it up made my head spin. . . ;-) If you don't want spend a little bit doing it, can you send me what you did so far and I'll work on the interface and organization? TW === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 So do you want to rewrite it or shall I play with it during my calculus class? It does seem like a usefull program for fluids. But the organization is horrible. Just opening it up made my head spin. . . ;-) If you don't want spend a little bit doing it, can you send me what you did so far and I'll work on the interface and organization? TW === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > So do you want to rewrite it or shall I play with it during my > calculus class? You should definetely follow your calculus class - I couldn't live with that responsibility ;-) > It does seem like a usefull program for fluids. I wouldn't know, I know almost nothing about fluids (I made calculus programs during my fluid physics classes ;-) > But the organization is horrible. Just opening it up > made my head spin. . . ;-) Yes, it took a few moments to grasp the meaning of it all - but 25k in as many vars isn't something to sneeze at. It has taken great effort to write it in the first place. > If you don't want spend a little bit doing it, can you send me what > you did so far and I'll work on the interface and organization? Hehe, or what about a macrochallenge? See who makes the smallest and best performing rewrite of FluidHP ;-)... Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> <42d2a9fa$0$169$edfadb0f@dtext02.news.tele.dk> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > You should definetely follow your calculus class - I couldn't live with > that responsibility ;-) Oh, this is actually my thrid time through. First my class load was to much and I had to drop it. Second time, I missed the last 1 1/2 months of class (for health related issues) in the 4 month semester, missed the third test and I STILL got a D+ in the class. Unfortunately that didn't count. . . =( So here I am going through it again. I am currently seond in my class of 35 students. . . doing proofs of the intermediate value thm. for the 40th time can get old. . . I need a distraction to keep me sane! =) > I wouldn't know, I know almost nothing about fluids (I made calculus > programs during my fluid physics classes ;-) Well I tend to make calculator programs during all my classes-even the ballroom dance classes. Kind of difficult to maintain a good topline while doing a natural pivot while holding a 49g+ in one hand . . . but I have perfected the technique! ;-) > Yes, it took a few moments to grasp the meaning of it all - but 25k in > as many vars isn't something to sneeze at. It has taken great effort to > write it in the first place. Yeah. I had to rewrite a 120kb useRPL library once. It took a long time. . . I am still working on it in fact. > Hehe, or what about a macrochallenge? See who makes the smallest and > best performing rewrite of FluidHP ;-)... Hmm. Could be interesting. Any rules? TW === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G >> You should definetely follow your calculus class - I couldn't live with >> that responsibility ;-) > Oh, this is actually my thrid time through. First my class load was to > much and I had to drop it. Second time, I missed the last 1 1/2 months > of class (for health related issues) in the 4 month semester, missed > the third test and I STILL got a D+ in the class. Unfortunately that > didn't count. . . =( So here I am going through it again. I am > currently seond in my class of 35 students. . . doing proofs of the > intermediate value thm. for the 40th time can get old. . . I need a > distraction to keep me sane! =) >> I wouldn't know, I know almost nothing about fluids (I made calculus >> programs during my fluid physics classes ;-) > Well I tend to make calculator programs during all my classes-even the > ballroom dance classes. Kind of difficult to maintain a good topline > while doing a natural pivot while holding a 49g+ in one hand . . . but > I have perfected the technique! ;-) >> Yes, it took a few moments to grasp the meaning of it all - but 25k in >> as many vars isn't something to sneeze at. It has taken great effort to >> write it in the first place. > Yeah. I had to rewrite a 120kb useRPL library once. It took a long > time. . . I am still working on it in fact. ********************************************************** I have also rewritten a HP 48GX program for the hp 49g+ It is much faster now, but only the main part is rewritten to be used as a library instead of a UserRPL program I had to move certain items around, boy was it difficult... >> Hehe, or what about a macrochallenge? See who makes the smallest and >> best performing rewrite of FluidHP ;-)... > Hmm. Could be interesting. Any rules? -------------------------------------------------------------- Count me out - too busy,,, VPN === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G Hehe, or what about a macrochallenge? See who makes the smallest and best performing rewrite of FluidHP ;-)... > Hmm. Could be interesting. Any rules? - Well, the obvious should be that it should be possible to solve at least the 12 originally supported equations from FluidHP. - And that it shall work on the HP49G (v1.19.6 onwards). I'd also opt for 49G+ support, but that's almost a given if it works on 49G. - It should be contained in a single flash storable lib (what lib number?). Anything else? I hope SysRPL is allowed - I prefer that when making a lib (then I can use Debug4x). I've been fiddling a bit more with FluidHP, and it's actually a quite user friendly program. I'll try to maintain that user friendly-ness or improve it if possible. Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> <42d2a9fa$0$169$edfadb0f@dtext02.news.tele.dk> <42d2d424$0$170$edfadb0f@dtext02.news.tele.dk> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > - Well, the obvious should be that it should be possible to solve at > least the 12 originally supported equations from FluidHP. Reading in the doc, it says only calculations in the 1st and 2nd are possible. Anyone who knows anything about steam (other than VPN who knows all about hot air ;-): are calculations in the other zones used very often or really needed? If so, where can I get some info to add in support for the other ones. > - And that it shall work on the HP49G (v1.19.6 onwards). I'd also opt > for 49G+ support, but that's almost a given if it works on 49G. Yes; a must. > - It should be contained in a single flash storable lib (what lib > number?). Pick one. > Anything else? I hope SysRPL is allowed - I prefer that when making a > lib (then I can use Debug4x). Yes. Anything goes. Even C should someone feel so inclined. I would say concentrate less on size and more on usability/accetable speed of calcualtions. So making a pure stack interface will most likely loose, even if it is smallest. =) > I've been fiddling a bit more with FluidHP, and it's actually a quite > user friendly program. I'll try to maintain that user friendly-ness or > improve it if possible. Same here. It would be nice if someone can provide some examples for testing the various cases so I can make sure it works. TW === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > Reading in the doc, it says only calculations in the 1st and 2nd are > possible. Yes, region 1 and 2. I meant the 12 *equations* from FluidHP. - It should be contained in a single flash storable lib (what lib number?). > Pick one. How about 1135? That's not used currently it seems... > So making a pure stack interface will most likely loose, > even if it is smallest. =) Of course :-) Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> <42d2a9fa$0$169$edfadb0f@dtext02.news.tele.dk> <42d2d424$0$170$edfadb0f@dtext02.news.tele.dk> <42d39264$0$183$edfadb0f@dtext02.news.tele.dk> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 Yeah. I knew you meant the 12 equations. . . I was just thinking that the other regions might be important as well. So if someone reading this provided the info. . . 1135 sounds good to me. I was planning on making a spiffy inform box as well. =o) I always just hate that initial run through a big userRPL program figuring out what is going on. Seems to be the most difficult part for me. . . =( Once it clicks however, it is usually pretty easy to cut the size in half or more. So should we try and set a date, or just when we get around to it? TW === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> <42d2a9fa$0$169$edfadb0f@dtext02.news.tele.dk> <42d2d424$0$170$edfadb0f@dtext02.news.tele.dk> <42d39264$0$183$edfadb0f@dtext02.news.tele.dk> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 Well, I found this http://www.iapws.org/relguide/IF97.pdf which has all the info. At least I know what all those numbers in the program are now. If this was something I was going to use i might try and expand the capabilities, but for now I am not THAT motivated. TW === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > I was just thinking that the other regions might be > important as well. So if someone reading this provided > the info. . . Of course. That'd be a nice addition if it's usable. > 1135 sounds good to me. Great. > I always just hate that initial run through a big > userRPL program figuring out what is going on. > Seems to be the most difficult part for me. . . =( Yeah, that's the hard part. There's really no easy way to view it on a PC either, so it's one tough edit on the calc. > So should we try and set a date, or just when we get > around to it? I can probably get something done before the 1st of August. Should we aim for that, or is it not enough time? Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > I can probably get something done before the 1st of August. Should we > aim for that, or is it not enough time? I will give it a try as well but don't wait for me. If you get something from me by the 1st be very worried next time you fly above England. What is more we have very nice weather these days so I can practice my favourite sport (1min 37s on saturday connaisseurs will apreciate) Also reading through some of the docs and references to hpfluid, at lunch break, it seems there has been an update to the equations since the software was out. I will have a look. Arnaud === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> <42d2a9fa$0$169$edfadb0f@dtext02.news.tele.dk> <42d2d424$0$170$edfadb0f@dtext02.news.tele.dk> <42d39264$0$183$edfadb0f@dtext02.news.tele.dk> <42d3c901$0$164$edfadb0f@dtext02.news.tele.dk> posting-account=dTw6fA0AAAAmz8yj9fY6u3ScMAxClw0N It«s working 3 times faster than teh original one, but there«s a fault in PINP. If you use ts=f(p) or Ps=f(t) there is a THEN-fault. running right. Kolja === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > there«s a fault in PINP. I found it: In line 20 (18 w/ minifont) you'll have to change 'pp' STO into {pp} STO In line 39 (32 w/ minifont) you'll have to change 'pt' STO into {pt} STO Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > It«s working 3 times faster than teh original one, but there«s a > fault in PINP. > If you use ts=f(p) or Ps=f(t) there is a THEN-fault. I'll take a look at it... Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> <42d2a9fa$0$169$edfadb0f@dtext02.news.tele.dk> <42d2d424$0$170$edfadb0f@dtext02.news.tele.dk> <42d39264$0$183$edfadb0f@dtext02.news.tele.dk> <42d3c901$0$164$edfadb0f@dtext02.news.tele.dk> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > I can probably get something done before the 1st of August. Should we > aim for that, or is it not enough time? Sounds good. TW === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G I can probably get something done before the 1st of August. Should we aim for that, or is it not enough time? > Sounds good. Almost done ;-) Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> <42d2a9fa$0$169$edfadb0f@dtext02.news.tele.dk> <42d2d424$0$170$edfadb0f@dtext02.news.tele.dk> <42d39264$0$183$edfadb0f@dtext02.news.tele.dk> <42d3c901$0$164$edfadb0f@dtext02.news.tele.dk> <42e261a3$0$171$edfadb0f@dtext02.news.tele.dk> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 Sounds good. > Almost done ;-) > Steen My interface is done. . . i just need to go through the calculation part. =) Unfortunately I have less time than I thought, but I still will get it done in time. TW === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > My interface is done. . . i just need to go through the calculation > part. =) Unfortunately I have less time than I thought, but I still > will get it done in time. Hi Tim. I've completed my port of FluidHP. I'm going on vacation in 15 hours or so - and then the next 5 days. I've sent you my lib on your gmail account. The result is 9310 bytes and typically 6 times faster than the original FluidHP lib (worst case 3 times faster, best case ~50 times faster). I'm looking forward to seeing your solution once I get back from vacation :-) Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> <42d2a9fa$0$169$edfadb0f@dtext02.news.tele.dk> <42d2d424$0$170$edfadb0f@dtext02.news.tele.dk> <42d39264$0$183$edfadb0f@dtext02.news.tele.dk> <42d3c901$0$164$edfadb0f@dtext02.news.tele.dk> <42e261a3$0$171$edfadb0f@dtext02.news.tele.dk> <42e56f41$0$170$edfadb0f@dtext02.news.tele.dk> posting-account=dTw6fA0AAAAmz8yj9fY6u3ScMAxClw0N Please send me your faster version of FluidHP. Kolja === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > Sounds good. >> Almost done ;-) >> Steen > My interface is done. . . i just need to go through the calculation > part. =) Unfortunately I have less time than I thought, but I still > will get it done in time. I was going through the IAPWS-IF97 and associated documents thinking the best way was to recode everything with double precision reals. I made a few routines but work took priority and I will certainly not have time to do anything for the next month at least. However I have now a very good understanding of IAPWS-IF97 and if needed am happy to answer any question (by email as I believe the forum is not the right place for those) if this can help. Good luck to Steen and Tim on their port. Arnaud === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > My interface is done. . . i just need to go through the calculation > part. =) I finished the interface in the first couple of days - getting any meaning out of the algorithms used have taken considerable time though (I'm 95% done with that part thankfully). They use more than 130 stack levels in some places. Maniacs! :-) I've also found a couple of mistyped numbers in the original, none of which affected precission too much though. (one place 0.0000611 should've been 0.000611 for instance, and 317.7 should've been 317.763). I think I'll land just below 7k for the complete library. All calculations feel near instantaneous (on a 49G+ anyways). It has been great fun till now, but have taken much more time than I anticipated. I hope someone'll find a use for these ports ;-) Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > All calculations feel near instantaneous (on a 49G+ anyways). Geesh! I must revise this statement after updating the internal structure of my library - a couple of calculations do take up to 4 seconds now! It's still 4-5 times faster than original, but very far from instantaneous. Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> <42d2a9fa$0$169$edfadb0f@dtext02.news.tele.dk> <42d2d424$0$170$edfadb0f@dtext02.news.tele.dk> posting-account=lAaz6QwAAABR-WEJBEnTuXipZ8zCV0L0 I would suggest that running on a 48G would be very welcome as well as it is the calculator that really needs size reduction. Also, before starting, I would suggest contacting the authors. A few publish it. That is when he told me a new version would be out very soon. I wish I had access to email when I was working on it. Arnaud === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > I would suggest that running on a 48G would be very welcome as well as > it is the calculator that really needs size reduction. You are right, but the HP48 doesn't have the new inform engine (which I'd very much like to use). If we get a couple of good ports, maybe one of them can be made to run on the HP48 too. > Also, before starting, I would suggest contacting the authors. Of course. I'll send a mail straight away. Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G Also, before starting, I would suggest contacting the authors. > Of course. I'll send a mail straight away. They never did answer btw. Steen === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G <42d2a391$0$157$edfadb0f@dtext02.news.tele.dk> <42d2a4d8$0$167$edfadb0f@dtext02.news.tele.dk> <42d2a9fa$0$169$edfadb0f@dtext02.news.tele.dk> <42d2d424$0$170$edfadb0f@dtext02.news.tele.dk> posting-account=dTw6fA0AAAAmz8yj9fY6u3ScMAxClw0N I would be happy, if I had a version running as fast as on the 48G, but the faster the better ;-). I want to say something to the need of all programme-formulas: The programme is for calculations of the Industrial Formulation IAPWS-IF 97 and I thnik it would be nice to keep this standard. I would be happy, if you could send me your improved version Steen, because I need the programme till 19.07.2005. I hope the battle for the fastest FLuidHp is over till this date and I get a real fast FluidHP ( ;-) ), but at the moment I take everything I can get. Kolja === Subject: Re: I need HELP, can someone take a look at FluidHP? Its sooo slow on HP49G > I would be happy, if I had a version running as fast as on the 48G, > but the faster the better ;-). Fast and small - the calc community is the only place where this is better ;-) > I would be happy, if you could send me your improved version Steen, > because I need the programme till 19.07.2005. You have mail. > I hope the battle for the fastest FLuidHp is over till this date and I > get a real fast FluidHP ( ;-) ), but at the moment I take everything > I can get. I can almost promise you that my port *won't* be finished till that date, but you have a faster version in your inbox. Steen === Subject: Amazon has the HP49g+ again. $122 nt posting-account=cEKa6Q0AAAC3YlI6nbEywfIpggMw2WOF nt === Subject: Re: Amazon has the HP49g+ again. $122 nt kinda like following gas prices... -- Frank Bachman (Grumpy Aero Guy) > nt === Subject: Re: Amazon has the HP49g+ again. $122 nt posting-account=hYpJhAwAAABbgx9xHixEB8na9TYTFvkV > nt Just curious :) Why is everyone so interested in these price fluctuations? Surely they can have nothing to do with anything HP may be planning? Bhuvanesh. === Subject: Re: Amazon has the HP49g+ again. $122 nt >> nt > Just curious :) Why is everyone so interested in these price > fluctuations? Surely they can have nothing to do with anything HP may > be planning? I suddenly got an idea: It has everything to do with HPW calculator plans! The new hypercalc is on the way! I will use one small nuclear battery and has 1GHz dual-core ARM CPU and a new keyboard with separate key hinges for each key Ddoubles are corrected... VPN PS: I should have not mentioned the corrected double keystrokes now nobody believes this news... :-D === Subject: Re: Amazon has the HP49g+ again. $122 nt posting-account=InIDMwwAAACLZ3gpW5d6vXE1F8VsBdvN 1) To capture a low price. 2) Curious about the behavior of one of the largest online retailers. === Subject: Re: Amazon has the HP49g+ again. $122 nt posting-account=cEKa6Q0AAAC3YlI6nbEywfIpggMw2WOF 3. I have no life. 4. Because of #3, this is my idea of a good time. 5. Thrill of the hunt. Will it drop below $100 again? === Subject: Never on a Saturday...Alarms, if you don't want to read about alarms === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms Ooops, that got away from me. OK. I have finally convinced myself that the 49g+ maintains time as accurately as the 48GX, until you execute CLKADJ from the board or as an alarm. CKLADJ never adjusts the time to better than +/- 4000 ticks. Whats really strange is that I have kept track of an alarm that runs at 11:00 PM everynight to adjust the alarm. Since June 20 it has run twice on the following day right when I turn the calculator on. Nothing new there. BUT it has totally failed to run three times, everytime a Saturday. Is that not strange? I usually don't use the calculator on weekends, but if that had anything to do with it, why does it run on Sunday? I guess the calculator isn't familiar eith the French movie, and got the day wrong. Scott Chapin === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms > Ooops, that got away from me. > OK. I have finally convinced myself that the 49g+ maintains time as > accurately as the 48GX, until you execute CLKADJ from the board or as an > alarm. CKLADJ never adjusts the time to better than +/- 4000 ticks. > Whats really strange is that I have kept track of an alarm that runs at > 11:00 PM everynight to adjust the alarm. Since June 20 it has run twice > on > the following day right when I turn the calculator on. Nothing new there. > BUT it has totally failed to run three times, everytime a Saturday. > Is that not strange? I usually don't use the calculator on weekends, but > if > that had anything to do with it, why does it run on Sunday? I guess the > calculator isn't familiar eith the French movie, and got the day wrong. > Scott Chapin What happens if you have a separate alarm for the Saturday only? BTW: can you be sure that an alarm is NOT executed on Saturdays? VPN === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms Whats really strange is that I have kept track of an alarm that runs at 11:00 PM everynight to adjust the alarm. Since June 20 it has run twice on the following day right when I turn the calculator on. Nothing new there. BUT it has totally failed to run three times, everytime a Saturday. I can confirm that rom 2.0 on hp 49g+ misses daily alarms. Especially when the calc has been turned off for more than 24 hrs it occures that a repeatedly alarms are missed. The alarm might be executed when you switch on the calc.Very strange and buggy behaviour. In case the calc is keeping off, the next alarm will be executed. ***Furthermore the alarms does occur randomly very unprecise! I have noticed alarms, which typically run more than 1 minute later than expected. ***There is still a weightily bug in the alarm management. It has nothing to do with Saturdays but it occures, when the calc is not used for a longer period of time. ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, and you will see it has no impact to the build in clock. There are maybe some other undocumented build in surprises concerning time mangegement? Would be nice to be **at least** documented! The hp 49g+ is still not usefull as an alarm clock! It is unreliable! Heiko Arnemann === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms > Whats really strange is that I have kept track of an alarm that runs > at > 11:00 PM everynight to adjust the alarm. Since June 20 it has run > twice > on > the following day right when I turn the calculator on. Nothing new > there. > BUT it has totally failed to run three times, everytime a Saturday. > I can confirm that rom 2.0 on hp 49g+ misses daily alarms. > Especially when the calc has been turned off for more than 24 hrs it > occures that a repeatedly alarms are missed. The alarm might be > executed when you switch on the calc.Very strange and buggy behaviour. > In case the calc is keeping off, the next alarm will be executed. > ***Furthermore the alarms does occur randomly very unprecise! > I have noticed alarms, which typically run more than 1 minute later than > expected. > ***There is still a weightily bug in the alarm management. > It has nothing to do with Saturdays but it occures, when the calc is > not used for a longer period of time. > ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, > and you will see it has no impact to the build in clock. > There are maybe some other undocumented build in surprises concerning > time mangegement? Would be nice to be **at least** documented! > The hp 49g+ is still not usefull as an alarm clock! It is unreliable! > Heiko Arnemann > ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, > and you will see it has no impact to the build in clock. If it is anything like the 48gx, then the number you give to CLKADJ has to be in multiples of clicks. You have to multiply the number of seconds that you want to adjust the clock by 8192 and then execute CLKADJ. Since 2000<8192 it probably ignores it. Rod === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms > If it is anything like the 48gx, then the number you give to CLKADJ has to > be in multiples of clicks. You have to multiply the number of seconds that > you want to adjust the clock by 8192 and then execute CLKADJ. Since > 2000<8192 it probably ignores it. On the 49G and 48GX Heiko use if CLKADJ is correct. Unfortunately, it's not that easy to emulate a similar behaviour on the 49g+. The 49g+ is using a RTC (real time clock) for the time module which has an accuracy to the second only. There's an extra 8192Hz timer but it's not accurate at all and so far hasn't been used for the timing function. Jean-Yves === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms If it is anything like the 48gx, then the number you give to CLKADJ has to be in multiples of clicks. You have to multiply the number of seconds that you want to adjust the clock by 8192 and then execute CLKADJ. Since 2000<8192 it probably ignores it. 49G is correct, too. On hp 49g+ the input for CLKADJ seem to be truncated to full seconds, e.g. 8200 ticks becomes 8192 and -8200 ticks becomes -8192. Jean-Yves Avenard schrieb > On the 49G and 48GX Heiko use if CLKADJ is correct. may be an if to much, in your sentence? CLKADJ works fine on the 49G and 48GX! > Unfortunately, it's not that easy to emulate a similar behaviour on the > 49g+. The 49g+ is using a RTC (real time clock) for the time module > which has an accuracy to the second only. There's an extra 8192Hz timer > but it's not accurate at all and so far hasn't been used for the timing > function. I assume, the economized 49g+ does not have an 8192 Hz-timer, or is it built in, but not used? By the way my HP 49G (using the 8192 Hz-timer) seem to be more accurate than the hp 49g+ RTC. It is may be not that stable (dependent on temperature or battery voltage...), I have not investigated for that up to know. Using a time keeper, the HP 49G seem to be more precize than a more expensive quarz-clock (which is not that easy to be adjusted)! It could be even better than a cheap radio controlled clock and could somtimes be more accurate than GPS time signals. The hp 49g+ clock has lost this property :-( Heiko === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms Whats really strange is that I have kept track of an alarm that runs at 11:00 PM everynight to adjust the alarm. Since June 20 it has run twice on the following day right when I turn the calculator on. Nothing new there. BUT it has totally failed to run three times, everytime a Saturday. I can confirm that rom 2.0 on hp 49g+ misses daily alarms. Especially when the calc has been turned off for more than 24 hrs it occures that a repeatedly alarms are missed. The alarm might be executed when you switch on the calc.Very strange and buggy behaviour. In case the calc is keeping off, the next alarm will be executed. ***Furthermore the alarms does occur randomly very unprecise! I have noticed alarms, which typically run more than 1 minute later than expected. ***There is still a weightily bug in the alarm management. It has nothing to do with Saturdays but it occures, when the calc is not used for a longer period of time. ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, and you will see it has no impact to the build in clock. There are maybe some other undocumented build in surprises concerning time mangegement? Would be nice to be **at least** documented! The hp 49g+ is still not usefull as an alarm clock! It is unreliable! Heiko Arnemann ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, and you will see it has no impact to the build in clock. > If it is anything like the 48gx, then the number you give to CLKADJ has to > be in multiples of clicks. You have to multiply the number of seconds that > you want to adjust the clock by 8192 and then execute CLKADJ. Since > 2000<8192 it probably ignores it. > Rod Hi Rod, I have never noticed that behavior on either of my 48GXs. When adjusting for a deviation at a time signal, my subsequent error has always been less than 3/100ths of a second. That error would be expected due to human reflexes when pushing TIME on a time signal. Scott === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms > Whats really strange is that I have kept track of an alarm that runs > at > 11:00 PM everynight to adjust the alarm. Since June 20 it has run > twice > on > the following day right when I turn the calculator on. Nothing new > there. > BUT it has totally failed to run three times, everytime a Saturday. > > I can confirm that rom 2.0 on hp 49g+ misses daily alarms. > Especially when the calc has been turned off for more than 24 hrs it > occures that a repeatedly alarms are missed. The alarm might be > executed when you switch on the calc.Very strange and buggy behaviour. > In case the calc is keeping off, the next alarm will be executed. > > ***Furthermore the alarms does occur randomly very unprecise! > I have noticed alarms, which typically run more than 1 minute later > than > expected. > > ***There is still a weightily bug in the alarm management. > It has nothing to do with Saturdays but it occures, when the calc is > not used for a longer period of time. > > ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, > and you will see it has no impact to the build in clock. > > There are maybe some other undocumented build in surprises concerning > time mangegement? Would be nice to be **at least** documented! > > The hp 49g+ is still not usefull as an alarm clock! It is unreliable! > > > Heiko Arnemann > > > ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, > and you will see it has no impact to the build in clock. >> If it is anything like the 48gx, then the number you give to CLKADJ has >> to >> be in multiples of clicks. You have to multiply the number of seconds >> that >> you want to adjust the clock by 8192 and then execute CLKADJ. Since >> 2000<8192 it probably ignores it. >> Rod > Hi Rod, > I have never noticed that behavior on either of my 48GXs. When adjusting > for > a deviation at a time signal, my subsequent error has always been less > than > 3/100ths of a second. That error would be expected due to human reflexes > when pushing TIME on a time signal. > Scott I mis-spoke, I meant to say ticks not clicks. The number you provide to CLKADJ has to be a multiple of ticks, i.e., the number of seconds by which you want to adjust the clock times 8192. Rod === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms > Whats really strange is that I have kept track of an alarm that runs at 11:00 PM everynight to adjust the alarm. Since June 20 it has run twice on the following day right when I turn the calculator on. Nothing new there. BUT it has totally failed to run three times, everytime a Saturday. I can confirm that rom 2.0 on hp 49g+ misses daily alarms. Especially when the calc has been turned off for more than 24 hrs it occures that a repeatedly alarms are missed. The alarm might be executed when you switch on the calc.Very strange and buggy behaviour. In case the calc is keeping off, the next alarm will be executed. ***Furthermore the alarms does occur randomly very unprecise! I have noticed alarms, which typically run more than 1 minute later than expected. ***There is still a weightily bug in the alarm management. It has nothing to do with Saturdays but it occures, when the calc is not used for a longer period of time. ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, and you will see it has no impact to the build in clock. There are maybe some other undocumented build in surprises concerning time mangegement? Would be nice to be **at least** documented! The hp 49g+ is still not usefull as an alarm clock! It is unreliable! Heiko Arnemann > > ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, and you will see it has no impact to the build in clock. > > > If it is anything like the 48gx, then the number you give to CLKADJ has > to > be in multiples of clicks. You have to multiply the number of seconds > that > you want to adjust the clock by 8192 and then execute CLKADJ. Since > 2000<8192 it probably ignores it. > > Rod > > Hi Rod, I have never noticed that behavior on either of my 48GXs. When adjusting for a deviation at a time signal, my subsequent error has always been less than 3/100ths of a second. That error would be expected due to human reflexes when pushing TIME on a time signal. Scott > I mis-spoke, I meant to say ticks not clicks. The number you provide to > CLKADJ has to be a multiple of ticks, i.e., the number of seconds by which > you want to adjust the clock times 8192. > Rod Hi Rod, I understand. The 49g+ does need to make whole second ajustments, but that was not the case with the 48GX, in my experience. Scott === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms > Whats really strange is that I have kept track of an alarm that runs > at > 11:00 PM everynight to adjust the alarm. Since June 20 it has run > twice > on > the following day right when I turn the calculator on. Nothing new > there. > BUT it has totally failed to run three times, everytime a Saturday. > I can confirm that rom 2.0 on hp 49g+ misses daily alarms. > Especially when the calc has been turned off for more than 24 hrs it > occures that a repeatedly alarms are missed. The alarm might be > executed when you switch on the calc.Very strange and buggy behaviour. > In case the calc is keeping off, the next alarm will be executed. > ***Furthermore the alarms does occur randomly very unprecise! > I have noticed alarms, which typically run more than 1 minute later than > expected. > ***There is still a weightily bug in the alarm management. > It has nothing to do with Saturdays but it occures, when the calc is > not used for a longer period of time. > ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, > and you will see it has no impact to the build in clock. > There are maybe some other undocumented build in surprises concerning > time mangegement? Would be nice to be **at least** documented! > The hp 49g+ is still not usefull as an alarm clock! It is unreliable! > Heiko Arnemann Hey Heiko, I am back from vacation, and will power my hp49+s on Monday. Maybe there will be more insight to alarms running on a machine that is left off. My take on the Saturdays is more of an idle time problem, that for me occurs on weekends. My GUESS is that repeat alarms get pended during hibernation and only the last one runs when the calculator is powered up. Scott === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms Whats really strange is that I have kept track of an alarm that runs at 11:00 PM everynight to adjust the alarm. Since June 20 it has run twice on the following day right when I turn the calculator on. Nothing new there. BUT it has totally failed to run three times, everytime a Saturday. I can confirm that rom 2.0 on hp 49g+ misses daily alarms. Especially when the calc has been turned off for more than 24 hrs it occures that a repeatedly alarms are missed. The alarm might be executed when you switch on the calc.Very strange and buggy behaviour. In case the calc is keeping off, the next alarm will be executed. ***Furthermore the alarms does occur randomly very unprecise! I have noticed alarms, which typically run more than 1 minute later than expected. ***There is still a weightily bug in the alarm management. It has nothing to do with Saturdays but it occures, when the calc is not used for a longer period of time. ***CLKADJ is not doing its job. Try several times 2000 CLKADJ, and you will see it has no impact to the build in clock. There are maybe some other undocumented build in surprises concerning time mangegement? Would be nice to be **at least** documented! The hp 49g+ is still not usefull as an alarm clock! It is unreliable! Heiko Arnemann > Hey Heiko, > I am back from vacation, and will power my hp49+s on Monday. Maybe there > will be more insight to alarms running on a machine that is left off. > My take on the Saturdays is more of an idle time problem, that for me occurs > on weekends. My GUESS is that repeat alarms get pended during hibernation > and only the last one runs when the calculator is powered up. > Scott OK, I powered up the hp49g+ that I take home with me. It was last powered up on Aug 5, prior to vacation. Having powered it up today, Aug 14, I see that the daily 2300hr alarm ran on the 5th, 7th, 9th, 11th, and 13th, every other day! I'll let you know what happens with the other one tomorrow. Scott === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms Scott Chapin schrieb > OK, I powered up the hp49g+ that I take home with me. It was last powered up > on Aug 5, prior to vacation. Having powered it up today, Aug 14, I see that > the daily 2300hr alarm ran on the 5th, 7th, 9th, 11th, and 13th, every other > day! I'll let you know what happens with the other one tomorrow. Hello Scott, so, it misses every second day, in case you do not touch the calc. What about the accuracy of the alarms. My alarms where more than one minute later? Heiko === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms > Scott Chapin schrieb OK, I powered up the hp49g+ that I take home with me. It was last > powered up on Aug 5, prior to vacation. Having powered it up today, Aug 14, I see > that the daily 2300hr alarm ran on the 5th, 7th, 9th, 11th, and 13th, every > other day! I'll let you know what happens with the other one tomorrow. > Hello Scott, > so, it misses every second day, in case you do not touch the calc. > What about the accuracy of the alarms. My alarms where more > than one minute later? > Heiko Hi Heiko, Well, I turned the beta machine on at work today and it ran on the 5th, 7th, 9th, 11th, 13th, and 15th (today), but that was this morning and it was actually running the 14th's alarm. I guess it pends until you turn it on, or the next alarm is due. All alarms, when run are 1 minute plus a random tad late (when they run). I was thinking of an power cycling alarm at 600hrs, as I would not be using the calculator then. I'm also wondering whether it is a garbage collection issue. If there were a way to force a collection after the alarm is run, perhaps that would cure it? I know...blue skies. The fact that the alarm execution is delayed 1+ minutes, makes me think that housekeeping might be the culprit. Scott === Subject: Re: Never on a Saturday...Alarms, if you don't want to read about alarms Ooops, that got away from me. OK. I have finally convinced myself that the 49g+ maintains time as accurately as the 48GX, until you execute CLKADJ from the board or as an alarm. CKLADJ never adjusts the time to better than +/- 4000 ticks. Whats really strange is that I have kept track of an alarm that runs at 11:00 PM everynight to adjust the alarm. Since June 20 it has run twice on the following day right when I turn the calculator on. Nothing new there. BUT it has totally failed to run three times, everytime a Saturday. Is that not strange? I usually don't use the calculator on weekends, but if that had anything to do with it, why does it run on Sunday? I guess the calculator isn't familiar eith the French movie, and got the day wrong. Scott Chapin > What happens if you have a separate alarm for the Saturday only? > BTW: can you be sure that an alarm is NOT executed on Saturdays? > VPN My, this is interesting. It's Monday you know......and again my alarm did not run on Saturday. When my alarm runs it puts a one in the stack and apends a row in a matrix with the date and time. I started this with just the date being called and appended to a list. But on 30 June, the list list did not contain 29 June but a 30 June before the 30 June alarm should run, so I figured it ran when I powered up that day (30 June). Sure enough, the next day I had two 30 Junes in the list. Each Monday upon powering up, I have only had two 1s in the stack instead of three, and Saturday's date and run time are not in the matrix, but Friday's and Sunday's are. I don't know what else I can do to be sure. Here's my alarm which is set to run daily at 23:00. << 14500 CLKADJ ALLOG DUP SIZE EVAL DROP 1 + DATE TIME 2 ROW-> SWAP ROW+ 'ALLOG' STO >> ALLOG is a two column matrix. The first cloum is the date and the second is the time. BTW...How was vacation? Scott === Subject: [OT] Holiday riding - Was: Never on a Saturday...Alarms, if you don't want to read about alarms X > BTW...How was vacation? Beam me up, Scott! It was fun riding over 1000 miles in Finland We (me&Nina) visited my parents in Oulu. Our first day we arrived Turku (the ancient capital of Finland) and Givi importer installed to our topcase, the Givi E52 Maxia, a dual brake light system, E97 Stoplight It already incorporated E96 Topcase Rack Also the saddle was molded to our comfort: The rider seat was narrowed and contoured to the butt and the passenger seat got more fill to get it softer and higher, higher for a better knee angle You just can not lower the passenger foot rest on the right side because the muffler is there. You can't lower the muffler because of the ground clearance issue You can't use a high rider because of the side hard cases So we got the passenger seat stuffed When at Oulu, my brother, who was an instrument technician (now a Master of Arts in Guitar) helped me to install a new alarm system, which now seems to operate It took three calendar days because the instructions were wrong at least in point where the color code was Black-Red but when we finally decided it to be Black-White instead (ground) it worked! One wire is still unconnected and I'm not sure whether it is the remote start or engine cut off or both I'm just glad over the tilt&move alarm feature Wonder what happens if the system malfunctions and sets the alarm on while riding - the engine will be cut off in 10 seconds. The rider has better got the clutch pulled before it happens... VPN I will put more in Finnish (use a translator) somewhere on the internet Suzuki Bandit sites during the next week or so. Photos will be available when they finally get scanned from film I still don't own a good digicamera === Subject: Re: [OT] Holiday riding - Was: Never on a Saturday...Alarms, if you don't want to read about alarms BTW...How was vacation? > Beam me up, Scott! > It was fun riding over 1000 miles in Finland > We (me&Nina) visited my parents in Oulu. > Our first day we arrived Turku (the ancient capital of Finland) > and Givi importer installed to our topcase, the Givi E52 Maxia, > a dual brake light system, E97 Stoplight > It already incorporated E96 Topcase Rack > Also the saddle was molded to our comfort: > The rider seat was narrowed and contoured to the butt > and the passenger seat got more fill to get it softer and higher, > higher for a better knee angle > You just can not lower the passenger foot rest on the right side > because the muffler is there. You can't lower the muffler > because of the ground clearance issue > You can't use a high rider because of the side hard cases > So we got the passenger seat stuffed > When at Oulu, my brother, who was an instrument technician > (now a Master of Arts in Guitar) helped me to install a new > alarm system, which now seems to operate > It took three calendar days because the instructions > were wrong at least in point where the color code was Black-Red > but when we finally decided it to be Black-White instead (ground) > it worked! One wire is still unconnected and I'm not sure whether > it is the remote start or engine cut off or both > I'm just glad over the tilt&move alarm feature > Wonder what happens if the system malfunctions and sets > the alarm on while riding - the engine will be cut off in 10 seconds. > The rider has better got the clutch pulled before it happens... > VPN > I will put more in Finnish (use a translator) > somewhere on the internet Suzuki Bandit sites > during the next week or so. > Photos will be available when they finally get scanned from film > I still don't own a good digicamera Give me a link. I'd like to see some pics. Scott === Subject: Re: [OT] Holiday riding - Was: Never on a Saturday...Alarms, if you don't want to read about alarms X > BTW...How was vacation? X >> VPN >> I will put more in Finnish (use a translator) >> somewhere on the internet Suzuki Bandit sites >> during the next week or so. >> Photos will be available when they finally get scanned from film >> I still don't own a good digicamera > Give me a link. I'd like to see some pics. > Scott These are the old pictures, the text (Finnish) is updated http://www.mp-foorumi.com/gallery/gallery.php?user=VPN&gid=958 VPN === Subject: HP 49G+ from Amazon: No keyboard problems so far I purchased one of the HP 49G+ calculators from Amazon last week during the $78 sale, and so far I haven't had it miss any keystrokes. Doubling is a little harder to say -- my fingers can get out of sync enough that sometimes I inadvertently double a keystroke. :-) Anyway, the serial number is 51001506, for those who are tracking these things. For $78 the HP 49G+ is a steal. I had been waiting for Qonos (and was prepared to spend the $300+ for it), but at this point I think it's pretty much a dead project, unfortunately. ---Joel Kolstad === Subject: Re: HP 49G+ from Amazon: No keyboard problems so far posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J Using the calc with moderation, the keys start to loosen past 6 months, maybe 1 month using a lot. === Subject: User RPL and local variables. Can I assign the output of a subroutine in a User RPL program to a local variable? E.g If I have a program that starts like this: << -> v w << v w + >> ... I want to assign << v w + >> (the subroutine) to, say, t for later use in the program. Is it possible to do something like this? I have tried several strategies and they either fail or the all the local variables are masked. Andrew === Subject: Re: User RPL and local variables. posting-account=D_LFpQwAAAD6y8YsWaqWMWI8FrNaVlfr VPN Notes: One can also use a compiled local variable which is a name prepended by backarrow <- -> <-t Hi! For me the use of compiled local variables is THE best solution. You can make your subroutines and then create a library. you can try this: << -> <-v <-w x << S1 100 + <-v <-w x * * + >> In which S1 is your subroutine 1: << <-v <-w + >> S1 STO after you have both programs stored in the calc, you just need to run the the main program and BINGO! Example: If you use 3 4 5 you should get 167 Then you can use a program like LIBKIT and store all your subroutines (if you have more than one subroutine) and the main program in a library. The cool thing of this is that the users of your programs just use the main program and your subroutines are hidden. NOTE: If you try to run the subroutine INSTEAD of the main program you will receive an error message: Undefined Local Name because there isn't any program that has created previously the variables involved. I hope this example helps you, good luck! JMD === Subject: Re: User RPL and local variables. A solution. With the help of James and Virgil's comments, this is the solution that I came up with: It assumes that here are two values on the stack before running the program. << -> V W << @Do something with V, W here. V W + @ If you want to reuse the local variables V and W later in the program do this: V W >> -> T V W << @ Continue with the calculations. V W T >> Of course various combinations of SWAP, ROLL DUP will work, but I was interested in the use of local variables. The scoping of the local variables is interesting. It seems that you must reassign them, as above, after they have been used in a subroutine that assigns a value to a local variable. Andrew === Subject: Re: User RPL and local variables. A solution. > With the help of James and Virgil's comments, this is the solution that I > came up with: > It assumes that here are two values on the stack before running the program. > << -> V W > << @Do something with V, W here. > V W + > @ If you want to reuse the local variables V and W later in > the program do this: > V W > >> -> T V W > << > @ Continue with the calculations. > V W T > >> > Well, that should work, but actually, there's no need to create another pair of local variables V and W. Try this: << @ Start program. -> V W @ Create local variables V and W. << @ Start local variable defining procedure. @ V and W are available within this defining procedure. << @ Start a subroutine to be placed on stack. @ Nested subroutine that can use V and W. >> @ End subroutine. -> T @ Make subroutine a local variable T. << @ Start a nested local variable defining procedure. @ V, W, and T are available during this defining @ procedure. >> @ End nested defining procedure. @ T no longer defined, but V and W still available. >> @ End outer defining procedure. @ V and W no longer defined. >> @ End program. > Of course various combinations of SWAP, ROLL DUP will work, but I was > interested in the use of local variables. The scoping of the local variables > is interesting. It seems that you must reassign them, as above, after they > have been used in a subroutine that assigns a value to a local variable. No, I think you're still misunderstanding them. As long as the defining procedure (a subprogram or an algebraic object) lasts, the local variables are still defined, even if used within a nested procedure. Inner procedures can use local variables defined for outer procedures, but outer procedures can't access local variables defined for inner procedures. Of course a local name can be quoted ('V', 'W', or 'T', for example) to put it on the stack unevaluated, so that a new object can be STOred in the local variable. Try using the DBUG operations to single-step through some of these programs. Note that while single-stepping, you can execute the local variables to see what they are and then DROP them, or see whether they're still defined for that matter. Borrowing from Bill Wickes's Insights book: 1 -> x << 2 -> y << x y + >> x + y + >> returns '4+y' (not 6), as follows: 1 -> x @ Store 1 in local variable x. << @ Start of program in which x is recognized. 2 -> y @ Store 2 in local variable y. << @ Start of program in which y is recognized. x y + @ Add x from outer program to y from inner @ program, returning 3. >> @ End of inner program where y is recognized. x + @ Add x to 3, returning 4. y + @ This y is *not* a local name, because it is @ outside of the program where y is local. It @ therefore names a global variable, which we are @ here assuming to have no current value. The sum @ is therefore '4+y' >> @ End of outer program where x is a local variable. If you rewrite the above sequence as 1 -> x << 2 -> y << x y + x + y + >> moving the final y back inside the program where the local variable y is defined, the sequence then returns the value 6. When two nested local variables structures define local variables with the same name, two separate local variables are created. Any use of the name refers to the most recently created local variable. The fact that there is another local variable with the same name in a previously created local memory does not matter. Thus 1 -> x << 2 -> x << x >> returns 2, whereas 1 -> x << 2 -> x << >> x >> returns 1. Also note that a FOR ... NEXT/STEP structure uses a named local variable for the loop index; the local variable is accessible anywhere within the loop after it's been named. I suppose that a START ... NEXT/STEP stucture would use an unnamed local variable for its loop index. You're welcome again. -- James === Subject: Re: User RPL and local variables. === Subject: Re: User RPL and local variables. > Can I assign the output of a subroutine in a User RPL program to a local > variable? > E.g > If I have a program that starts like this: > << -> v w << v w + >> ... > I want to assign << v w + >> (the subroutine) to, say, t for later use in > the program. > Is it possible to do something like this? I have tried several strategies > and they either fail or the all the local variables are masked. Sure, like this: << << v w + >> -> v w t << Rest of the program >> << Note that calling local variable t just puts the program << v w + >> on the stack; it doesn't execute it. You'll want to do t EVAL. You can also avoid the local variable and just put the program << v w + >> on the stack, and use a variation of PICK (including PICK3, OVER or DUP) and then EVAL when you want to use it, except that for the final use, instead use something in the ROLL family (including ROT or SWAP) followed by EVAL. But if you want to assign the *result* of the subroutine to a local variable, perhaps the easiest strategy is to start out with a dummy local variable. For example, assuming two numbers on the stack, DUP one of them and assign it to a dummy local variable, like this: << DUP -> v w d << v w + 'd' STO Rest of the program >> In this case, when you want to put the sum of v and w on the stack, just call d. If you want to store something else in d, then put whatever it is on the stack, and then do 'd' STO. You're welcome. -- James === Subject: Re: User RPL and local variables. posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo HP screwed up on the EVAL and where to use it . If they had simply not basterdised the TIC !! In Forth there are no such features . Every time you put something on the stack ( at high level ) it is executed , no exceptions . Forth would only need an EVAL after a TIC !!! Simple huh ? there are only 2 operations , run and create . EVERYTHING is RUN , unless TIC'd . HP has not a clue !! Everything in the 49 can be rewritten to run perfectly , w/ fewer problems if they fixed it . === Subject: Re: User RPL and local variables. > << > << > v w + > >> > -> v w t > << > Rest of the program > >> > << Of course that should be a *closing* program delimiter at the end, so make that: << << v w + >> -> v w t << Rest of the program >> -- James === Subject: Re: User RPL and local variables. > Can I assign the output of a subroutine in a User RPL program to a local > variable? > E.g > If I have a program that starts like this: ... > I want to assign << v w + >> (the subroutine) to, say, t for later use in > the program. > Is it possible to do something like this? I have tried several strategies > and they either fail or the all the local variables are masked. > Andrew Perfectly possible. << << v w + >> -> t << @ the rest of your program here >> However note that v and w must be global variables in the above or at least defined prior to defining t. Alternately you might want something like << << -> v w << v w + >> -> t << @ the rest of your program here >> in which case invoking 't' will use the two lowest items on the stack as values for v and w. Or you can call 't' as a function, 't(x,y)' by itself or in an algebraic. === Subject: Re: User RPL and local variables. VPN Notes: One can also use a compiled local variable which is a name prepended by backarrow <- -> <-t << @ you can use <-t now in separate subprograms the <-t will be automatically erased after the defining main program ends like any other local variable, but it is accessible in separate subprograms like any Global Variable Remember always use EVAL after a local variable call 't' or '<-t' will ReCaLl the contents of a local variable to the stack t or <-t will also just do the recall, so you have to add EVAL VPN >> Can I assign the output of a subroutine in a User RPL program to a local >> variable? >> E.g >> If I have a program that starts like this: > ... >> I want to assign << v w + >> (the subroutine) to, say, t for later use in >> the program. >> Is it possible to do something like this? I have tried several strategies >> and they either fail or the all the local variables are masked. >> Andrew > Perfectly possible. > << << v w + >> -> t > << > @ the rest of your program here > >> > > However note that v and w must be global variables in the above or at > least defined prior to defining t. > Alternately you might want something like > << << -> v w << v w + >> -> t > << > @ the rest of your program here > >> > > in which case invoking 't' will use the two lowest items on the stack as > values for v and w. Or you can call 't' as a function, 't(x,y)' by > itself or in an algebraic. === Subject: How to change an INFORM label programmatically (SysRPL) Hi. I have this inform box coded in SysRPL (new inform engine), and would like to change the value of a label in runtime. It seems I can use ^IfSetAllLabelsMessages, but how do I prepare the arguments for that? What do I do with the returned data string afterwards? Is there any easier way to change the text on a single label (are the label values stored in a variable when the inform is running)? A small Steen === Subject: Re: How to change an INFORM label programmatically (SysRPL) posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 I tried doing this oncea long time ago, and after a long time and it not working quite right, I discovered I saved space and time by simply making another label/field visible or not. But if this is for an updated inform builder like I suspect. . . nevermind. =) TW === Subject: Re: How to change an INFORM label programmatically (SysRPL) Hi Steen, I think the standard way would be to override the associated IFM_DISPLABEL message, maybe in conjunction with IFM_DISPCLIENT . Bugt you could also use IFEDispLabel, which takes the client GROB and the LabelId from the stack. You would only have to find out the function or message to manipulate the associated label text. Raymond Steen Schmidt schrieb im Newsbeitrag > Hi. > I have this inform box coded in SysRPL (new inform engine), and would > like to change the value of a label in runtime. > It seems I can use ^IfSetAllLabelsMessages, but how do I prepare the > arguments for that? What do I do with the returned data string > afterwards? > Is there any easier way to change the text on a single label (are the > label values stored in a variable when the inform is running)? A small > Steen === Subject: Re: How to change an INFORM label programmatically (SysRPL) > Hi Steen, > I think the standard way would be to override > the associated IFM_DISPLABEL message, > maybe in conjunction with IFM_DISPCLIENT . > Bugt you could also use IFEDispLabel, > which takes the client GROB and the LabelId from the stack. This is not what Steen is asking at all working with the 49G new InputForm engine. JY === Subject: Re: How to change an INFORM label programmatically (SysRPL) Jean-Yves Avenard schrieb im Newsbeitrag >> Hi Steen, >> [..] >> Bugt you could also use IFEDispLabel, >> which takes the client GROB and the LabelId from the stack. > This is not what Steen is asking at all working with the 49G new InputForm > engine. I'm sure you also have an answer to the original question... Raymond === Subject: Re: How to change an INFORM label programmatically (SysRPL) > I'm sure you also have an answer to the original question... Cyrille would be the best one to answer... There are things I can't answer due to legal agreements. JY === Subject: Re: How to change an INFORM label programmatically (SysRPL) hello, try: Ifl'StringData Label2 FALSE TEXT FLASHPTR IfSetGrob3 where Label2 is the label definition normally created by the input form editor. >> I'm sure you also have an answer to the original question... > Cyrille would be the best one to answer... > There are things I can't answer due to legal agreements. > JY === Subject: Re: How to change an INFORM label programmatically (SysRPL) > hello, > try: > Ifl'StringData Label2 FALSE TEXT FLASHPTR IfSetGrob3 > where Label2 is the label definition normally created by the input > form editor. Steen === Subject: It's me again... replacement No. 6 is broken.. Good things come to those who wait... january I received my 6th 49G+ which was working at least. But now, the keys 1 and 2 have become loose. It took the keyboard about half a year to break... I am glad that we have in germany a two-year-warranty. Maybe I should try another calc. I am thinking about the 33S. What do you think? Christian === Subject: Re: It's me again... replacement No. 6 is broken.. posting-account=IvsPIgwAAACwGSfru_P0ZGT09wgEXKGS What serial number is on your hp49g+. I have had mine for about 3 months now. I have serail number CNA5150...... and cant find any problems with the keyboard. mine is working to my satifaction. mine was replaced because of early keboard problems (CN333......) my nwe hp49g+ was sent as a replacement from California. === Subject: Re: It's me again... replacement No. 6 is broken.. It was a CNA 42.... I asked the HP support about the current s/n. Here in Germany they deliver still the CNA4.... wiltin schrieb im Newsbeitrag > What serial number is on your hp49g+. I have had mine for about 3 > months now. I have serail number CNA5150...... and cant find any > problems with the keyboard. mine is working to my satifaction. mine > was replaced because of early keboard problems (CN333......) my nwe > hp49g+ was sent as a replacement from California. === Subject: Re: It's me again... replacement No. 6 is broken.. Hi Christian, I still suggest to take a look at the HP-48G series. Maybe they're not as fast as the 49g series stuff, but at least an HP-48G/G+/GX is reliable. Raymond Christian Buhmann schrieb im Newsbeitrag > Good things come to those who wait... > january I received my 6th 49G+ which was working at least. > But now, the keys 1 and 2 have become loose. It took the keyboard about > half a year to break... > I am glad that we have in germany a two-year-warranty. Maybe I should try > another calc. I am thinking about the 33S. What do you think? > Christian === Subject: Re: It's me again... replacement No. 6 is broken.. posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 Well, my business parter has been using his out in the field, dragging through dense underbrush, wading through swamps, rivers, and hoards of alien invaders now for two years. He has found it to be more reliable. So I guess it all depends on your personal views. . . === Subject: Re: It's me again... replacement No. 6 is broken.. Of course it depends on personal views. But I am sure in my cases you would have agreed... On some calcs some keys couldn«t be pressed down.. Two others had no USB connection And now two keys became loose I think nobody would / should accept this for a calc at this price.... schrieb im Newsbeitrag > Well, my business parter has been using his out in the field, dragging > through dense underbrush, wading through swamps, rivers, and hoards of > alien invaders now for two years. He has found it to be more reliable. > So I guess it all depends on your personal views. . . === Subject: Re: It's me again... replacement No. 6 is broken.. By the way... one of my colleagues has been using his HP42CY since the early eighties. He said when he bought his HP the were calles the Mercedes of calcs. Sad that this is no more true. Christian Buhmann schrieb im Newsbeitrag > Of course it depends on personal views. But I am sure in my cases you > would have agreed... > On some calcs some keys couldn«t be pressed down.. > Two others had no USB connection > And now two keys became loose > I think nobody would / should accept this for a calc at this price.... > schrieb im Newsbeitrag >> Well, my business parter has been using his out in the field, dragging >> through dense underbrush, wading through swamps, rivers, and hoards of >> alien invaders now for two years. He has found it to be more reliable. >> So I guess it all depends on your personal views. . . === Subject: Re: It's me again... replacement No. 6 is broken.. > one of my colleagues has been using his HP42CY since the early eighties. He > said when he bought his HP the were calles the Mercedes of calcs. > Sad that this is no more true. Yes, but it's not that some calculator manufacturer dethroned HP and forced them to lower the quality. Instead, dirt cheap PCs came in and relegated calculators primarily to the educational market. Folks in that market have nowhere near the free bucks to spend that the more traditional HP calculator markets (surveyors, engineers, etc.) do... You can point to a lot of technologies that, while still ubiquitous, have been somewhat surpassed by newer innovations and as such arguably the best products ever made for them have already come and gone. Examples: Analog FM radios (the ones from the '80s were much better... and much spendier!), cassette decks (when was the last time you saw a boombox where the built-in cassette deck had Dolby noise reduction?), VCRs (SVHS is almost dead; remaining VCRs are regular old VHS, no flying erase heads, editing decks, etc.), analog oscilloscopes (sniff!), etc. Some of these technologies such as VCRs and cassette decks will probably be gone completely in another 10 years, whereas others such as FM radio and scientific calculators will manage to eek out an existance for a somewhat longer period... (since calculators will just slowly morph into slightly modified PDAs, as Qonos was going to be...) I think nobody would / should accept this for a calc at this price.... For $78, the HP-49G is a very good deal. For $150 it's a little harder to say -- the TIs do start to look awfully attractive! ---Joel Kolstad === Subject: Re: It's me again... replacement No. 6 is broken.. Joel you are right... > For $78, the HP-49G is a very good deal. For $150 it's a little harder to > say -- the TIs do start to look awfully attractive! Why I do not like the TI: 1. The keys, they are two soft 2. the notation: If you want the tangent of 30 degrees. You will have to press [shift] [t] [3] [0] [)] [enter] 3. the programming, ok basic is simple but the TI input and output is strange [OT]: On thursday (2005-07-07) I was at the U2 concert in Berlin. God that was great. And thx to the internet I found the whole recording of the show....AND I bought two tickets on ebay for the price of 35? each.... I love the internet.... Joel Kolstad schrieb im Newsbeitrag >> one of my colleagues has been using his HP42CY since the early eighties. >> He >> said when he bought his HP the were calles the Mercedes of calcs. >> Sad that this is no more true. > Yes, but it's not that some calculator manufacturer dethroned HP and > forced > them to lower the quality. Instead, dirt cheap PCs came in and relegated > calculators primarily to the educational market. Folks in that market > have > nowhere near the free bucks to spend that the more traditional HP > calculator > markets (surveyors, engineers, etc.) do... > You can point to a lot of technologies that, while still ubiquitous, have > been > somewhat surpassed by newer innovations and as such arguably the best > products ever made for them have already come and gone. Examples: Analog > FM > radios (the ones from the '80s were much better... and much spendier!), > cassette decks (when was the last time you saw a boombox where the > built-in > cassette deck had Dolby noise reduction?), VCRs (SVHS is almost dead; > remaining VCRs are regular old VHS, no flying erase heads, editing decks, > etc.), analog oscilloscopes (sniff!), etc. > Some of these technologies such as VCRs and cassette decks will probably > be > gone completely in another 10 years, whereas others such as FM radio and > scientific calculators will manage to eek out an existance for a somewhat > longer period... (since calculators will just slowly morph into slightly > modified PDAs, as Qonos was going to be...) > I think nobody would / should accept this for a calc at this price.... > For $78, the HP-49G is a very good deal. For $150 it's a little harder to > say -- the TIs do start to look awfully attractive! > ---Joel Kolstad === Subject: Re: It's me again... replacement No. 6 is broken.. posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE You wanna break the world record in changing HP-49g+? :)) fred. === Subject: Re: It's me again... replacement No. 6 is broken.. posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J Here in Chile, now HP offers to extend a year the warranty for each 2 years warranty too. By the way, you beat me, i only change the calc 3 times :D Christian Buhmann ha escrito: > Good things come to those who wait... > january I received my 6th 49G+ which was working at least. > But now, the keys 1 and 2 have become loose. It took the keyboard about half > a year to break... > I am glad that we have in germany a two-year-warranty. Maybe I should try > another calc. I am thinking about the 33S. What do you think? > Christian === Subject: EQNWRITER HANGS posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J 1. Start equation writer on hp49g+ (rom 2.0?) 2. Type some like: (X^5)/10 <--- in pretty print, of course 3. Select the whole fraction, push +/- twice 4. Press - once, cursor appear in wrong position: ((X^5)/10)-[here] <--- old 49g ((X^[here_overlaped]5)/10) <--- 49g+ 5. then ENTER (mine hangs) === Subject: Re: EQNWRITER HANGS > 1. Start equation writer on hp49g+ (rom 2.0?) > 2. Type some like: > (X^5)/10 <--- in pretty print, of course > 3. Select the whole fraction, push +/- twice > 4. Press - once, cursor appear in wrong position: > ((X^5)/10)-[here] <--- old 49g > ((X^[here_overlaped]5)/10) <--- 49g+ Those overlappings has happened to me many times but I never cared to report them since it was otherwise ok but the hanging thing is a new one VPN > 5. then ENTER (mine hangs) === Subject: Re: EQNWRITER HANGS posting-account=U2_nsg0AAACMehSjdUxQpFmS8kMQCYl9 Mine does not hang however, it only shows ' Invalid Expressions' === Subject: Re: EQNWRITER HANGS posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 Wow! Thats a good one! =) Works with any number; using both + or -. I think the extra + in front of it messes things up. . . TW === Subject: Re: EQNWRITER HANGS > Wow! Thats a good one! =) Works with any number; using both + or -. > I think the extra + in front of it messes things up. . . > TW Yes good find. The funny thing is that it only does it on the 49g+, not the 49g or 48gii... JY === Subject: Re: EQNWRITER HANGS <3jk0v6Fq801jU1@individual.net> posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J Jean-Yves Avenard ha escrito: Wow! Thats a good one! =) Works with any number; using both + or -. I think the extra + in front of it messes things up. . . TW > Yes good find. > The funny thing is that it only does it on the 49g+, not the 49g or 48gii... > JY I dont have old hp's, but emulator (48gx and 49g) does not hangs, maybe in 49g+ NEG is bad rewritten? === Subject: Re: EQNWRITER HANGS <3jk0v6Fq801jU1@individual.net> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > Yes good find. > The funny thing is that it only does it on the 49g+, not the 49g or 48gii... I think the most interesting part is that it seems like something so simple to do, yet it has not been mentioned before. I guess this goes to show how often people push negative TWICE on a regular basis. Other than people trying to do a one sided limit (and thinking that pushing twice will somehow make the little sign move right or something) I guess it just doesn't happend that much. TW === Subject: Re: EQNWRITER HANGS > I think the most interesting part is that it seems like something so > simple to do, yet it has not been mentioned before. I guess this goes As I said it is something that doesn't happen on the 49g or the 48gii. So my guess is that it's related to the new 80 lines screen which was only introduced with ROM 2.0. Probably one reason why it hasn't been found yet. JY === Subject: Re: EQNWRITER HANGS >> I think the most interesting part is that it seems like something so >> simple to do, yet it has not been mentioned before. I guess this goes > As I said it is something that doesn't happen on the 49g or the 48gii. > So my guess is that it's related to the new 80 lines screen which was > only introduced with ROM 2.0. Probably one reason why it hasn't been > found yet. > JY Well, I'll add a report on Bugzilla for this one... -- Stefano Priore | Debian Sarge 3.1r0 --------------------------------+-------------------------------- Video meliora proboque, | Linux Registered User #210152 autem deteriora sequor... | Linux Registered Machine #97752 === Subject: Manual + addenda for HP-45 how much must one pay for HP-45 manual with addenda today? Any suggestions? (No auction on ebay) Marc === Subject: Re: Manual + addenda for HP-45 try http://www.hpmuseum.org/cd/cddesc.htm > how much must one pay for HP-45 manual with addenda today? > Any suggestions? (No auction on ebay) > Marc === Subject: Re: Manual + addenda for HP-45 > try http://www.hpmuseum.org/cd/cddesc.htm >>how much must one pay for HP-45 manual with addenda today? >>Any suggestions? (No auction on ebay) Sorry, but I need it in another direction. I don't want do buy, I want to sale, or in other words - a colleague (HP collector) in my firm want buy this manual with addenda from me and I don't known what price should I call. Marc === Subject: Re: Manual + addenda for HP-45 How about 100 bucks? VPN >> try http://www.hpmuseum.org/cd/cddesc.htm >how much must one pay for HP-45 manual with addenda today? >Any suggestions? (No auction on ebay) > Sorry, but I need it in another direction. I don't want do > buy, I want to sale, or in other words - a colleague (HP > collector) in my firm want buy this manual with addenda from > me and I don't known what price should I call. > Marc === Subject: Re: Manual + addenda for HP-45 > How about 100 bucks? Very funny! :) Marc === Subject: Need help with hp49g+ and USB I have a very basic question about hp49g+. It is so that I em trying to set up a communication between my computer and my hp49g+ via USB to download some programs, but couldn't establish any communication. What's wrong? And what parameters should be stored in IOPAR. I using the communication program HP 48 49 Calculator Connectivity Kit. Michael Numminen, M.Sc.E.E. === Subject: Re: Need help with hp49g+ and USB > What's wrong? And what parameters should be stored in IOPAR. I using the > communication program HP 48 49 Calculator Connectivity Kit. Read the nice help file, especially the section on trouble shooting. Also the help file mentions that IOPAR does not matter -- - - - - - - - - - - - - - - - - Bill Graves RKBA! bgraves@ix.netcom.com === Subject: Re: Need help with hp49g+ and USB posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Just use an SD card and an SD card reader... Fred. === Subject: Re: Need help with hp49g+ and USB posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo > Just use an SD card and an SD card reader... > Fred. ------------------------------------------ Great idea ! I've been trying to make USB work for many months , no luck . Maybe it's for not being real USB !!! === Subject: Re: Need help with hp49g+ and USB posting-account=U2_nsg0AAACMehSjdUxQpFmS8kMQCYl9 If you follow the instruction and all, and it is still not working, there is a flag about connecting to a computer using infrared that you need to uncheck, I'm not sure what flag number that is (maybe someone else here have a better idea). Go to Mode, Flag and scroll down until you see a flag that says Connect using infrared (or some sort). You have to uncheck that. === Subject: Re: Need help with hp49g+ and USB Make sure you use the latest connectivity software. Older version does not work. Go to HP.com Once you install the latest, you can just cut and paste files. I had the same problem unitl I used the latest version. cslim > If you follow the instruction and all, and it is still not working, > there is a flag about connecting to a computer using infrared that you > need to uncheck, I'm not sure what flag number that is (maybe someone > else here have a better idea). Go to Mode, Flag and scroll down until > you see a flag that says Connect using infrared (or some sort). You > have to uncheck that. === Subject: Axes in Fast-3D plot with 49G+ posting-account=U2_nsg0AAACMehSjdUxQpFmS8kMQCYl9 How can I include axes in Fast-3D plot? For example, lets say I want to plot x+y+z=2 surface using Fast-3D, how can I create a plot with the X-axis, Y-axis, and Z-axis shown (with the intersection to the axes)? f4ding === Subject: Re: Axes in Fast-3D plot with 49G+ posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > How can I include axes in Fast-3D plot? Write a new one. . . =) That axis are already there. They aren't huge ones that cover the screen though. They are small tiny things down above the TRACE button. They are even labeled for you nicely. . . ;-) TW === Subject: Re: Axes in Fast-3D plot with 49G+ posting-account=U2_nsg0AAACMehSjdUxQpFmS8kMQCYl9 Haaaa, the obvious answer. LOL. So you are saying there isn't a way to do this (besides the obvious) ?? The tiny one is fine, but there won't be any intersection between the surface and the axis shown. Although trace or edit might be able to give some info on that...O well. By the way, I hate programming ;) How can I include axes in Fast-3D plot? > Write a new one. . . =) > That axis are already there. They aren't huge ones that cover the > screen though. They are small tiny things down above the TRACE button. > They are even labeled for you nicely. . . ;-) > TW === Subject: Do you hate to Store Things...??? posting-account=gXhbegwAAADGqPjeoqi5G78ml6pFC5ww Do you hate to store things...??? i generate a huge mountain of crap whenever i'm using my calculator, and i often want to set something aside, not on the stack, where i might absentmindedly clear it, but somewhere in the current directory, just for a moment or two, and the tedious thing about this is-- thinking up a unique name for it...!!! i don't want to lose my train of thought from what i'm doing to check to see if 'A' is already taken, and often if i did use that scheme, i might have to check to see if 'M' or 'Q' were taken already, or i would have to think of 'what it was' to think up an appropriate label for it...! UGH...!!! so i've created a little ( ? ) Save program that is on my STO key, as a UserKey, and it simply saves whatever is on the stack under a unique name... then if i forget where it is, i can use VIEW to look around for it... Since i've come up with this program, i've developed a more general, transdirectory tool that subliments 'Last Commands' ( which can only hold 4 items )... Here is Save http://transamoebae.blogspot.com/2005/07/save-for-hp48gx.html This also contains the subroutine 'Clic' which i think is invaluable, it allows you to DoubleClic a program to bring up a Prompt, or do something else even...??? The double click means that you can still use the program as a subroutine for other programs, but will remind you of its necessary input if you use it alone...!!! === Subject: SOUND CHECK Sound section of OPEN FIRE updated, have fun ! manjo http://fly.srk.fer.hr/~manjo/openfire === Subject: Re: SOUND CHECK Great... but where/which are the new download ? db4gr3$fqo$1@ss405.t-com.hr... > Sound section of OPEN FIRE updated, > have fun ! > manjo > http://fly.srk.fer.hr/~manjo/openfire === Subject: Re: SOUND CHECK He he :-) nothing new to download :-) -just information ...ooo by the way please people encurage Lilian to release Treize (incredible game) for 49G+ am i the only one who likes to play games ? (of course not -right ?) oo Lilian -did you notice Eric released your conversions on Hpcalc.org as well -greetings to Eric as well !! take care and see you soon ! manjo === Subject: Re: SOUND CHECK db6f28$gjd$1@ss405.t-com.hr... > He he :-) >> ...ooo by the way please people encurage Lilian to release Treize > (incredible game) > for 49G+ But only in French... and the palyers really need to understand the story :( === Subject: Re: SOUND CHECK > ...ooo by the way please people encurage Lilian to release Treize (incredible game) for 49G+ > But only in French... and the palyers really need to understand the story :( Ok, i agree but in that case let's look at it this way: Treize = Great and fun French-learning tool :-) see you later manjo === Subject: Re: SOUND CHECK <42d6a8df$0$25055$8fcfb975@news.wanadoo.fr> posting-account=PUKtNgwAAACqPi8kDc99J0TU9omMMXD5 This is not the place but im curious to know about the progress of dune. How the conversion is doing ? Thx for the work === Subject: Re: SOUND CHECK > This is not the place but im curious to know about the progress of > dune. How the conversion is doing ? Thx for the work I'm glad you asked, Well... Lilian did Dune way back, and it is ready, the problem is that current ROM (rom2.0) has a contrast bug which needs to be fixed. (game gets stuck while fading the screen - Lilian found similar issues with contrast with some other programs too) as soon as ROM is updated (contrast controll bug fixed) we will release dune and awailable for you to play :-) i hope Lilian agrees on this :-) i belive they will update the rom, so don't worry -all what we need at this point is support from you people manjo http://fly.srk.fer.hr/~manjo/openfire Lilian's page http://perso.atsat.com/pigallio === Subject: Re: SOUND CHECK <42d6a8df$0$25055$8fcfb975@news.wanadoo.fr> posting-account=PUKtNgwAAACqPi8kDc99J0TU9omMMXD5 With the release of the new rom, I started having problems with many programs and mainly with games =[ But fortunately, mosf of them don't interfere on the gameplay I'll stay tuned for dune release ^^ === Subject: confused about ARM posting-account=O6knhQ0AAAD8zE21TyMVwtCasYnRdHOm Reading web pages like the ARM Assembly Programming on the HP49G+ i understood that its possible to run ARM code embedding it in saturm asm code as described in that page. Thats necessary because hp49g+ emulates a saturn and then expects a program to be saturn instructions. But reading other document posted here i see there is an ARM mode in the MASD. So what kind of code is generated by this ARM mode MASD, i mean, does it need the same trick to run :-/ ? please be patient as im not a computer scientist and im just trying to understand. === Subject: Re: confused about ARM > Reading web pages like the ARM Assembly Programming on the HP49G+ i > understood that its possible to run ARM code embedding it in saturm > asm code as described in that page. Thats necessary because hp49g+ > emulates a saturn and then expects a program to be saturn instructions. > But reading other document posted here i see there is an ARM mode in > the MASD. So what kind of code is generated by this ARM mode MASD, i > mean, does it need the same trick to run :-/ ? Yes. It's not exactly trivial to get ARM code running. You have to make sure its word-aligned (eg the address of each instruction is a multiple of 4). Then you move the address of the ARM code into the C register, and run ARMSAT. The example code included with the MASD docs does the word-alignment in a slightly different way. It copies the ARM code to 0x80100. This location is unused, word-aligned, and is 250 bytes long. So you can't use it for any non-trivial programs. I don't see much point in programming assembler. It takes too long to do. The only real advantage is small size - but there is quite a lot of storage for programs anyway. You can also use MASD directly on the calculator, if you enjoy entering data on a tiny screen/keyboard :) Al === Subject: Re: confused about ARM <42d6f4c8$0$10837$5a62ac22@per-qv1-newsreader-01.iinet.net.au> posting-account=FCAFHAwAAADaPCjuTWXm3ER18SIzldtO Just a small addition, since I know many people may be interested in the following code snippets, both for the new MASD: 1) The following code is a minimalist launcher, similar to the one found in the docs, but without the limitation in the size of the program (which Al correctly pointed out) when using 80100. Of course, you need enough free ram (not checked here) to contain a copy of the ARM program + 128 bytes. CODE % SMALL ARM LAUNCHER SAVE GOSUB SKIPCODE % INSERT ARM CODE HERE !ARM % TEST CODE - RETURN ONLY MOV PC,LR !ASM *SKIPCODE A=PC A=A-4.A C=RSTK ACEX.A D0=C % END OF BLOCK C=C-A.A % NIBBLES TO COPY D1=(2) 0 % ALIGN MOVEUP CD1EX ARMSAT LOADRPL ENDCODE @ 2) The following is a generic template based on that doc which generates a string to be executed by the ARMToolBox launcher. !ASM !NO CODE CP=0 DCCP #1028 SAT READ DCCP #1028 SAT WRITE DCCP #260 SAT PRIOR DCCP 4 S AL DCCP 4 S AH DCCP 4 S BL DCCP 4 S BH DCCP 4 S CL DCCP 4 S CH DCCP 4 S DL DCCP 4 S DH DCCP 4 S R0L DCCP 4 S R0H DCCP 4 S R1L DCCP 4 S R1H DCCP 4 S R2L DCCP 4 S R2H DCCP 4 S R3L DCCP 4 S R3H DCCP 4 S R4L DCCP 4 S R4H DCCP 4 S D0 DCCP 4 S D1 DCCP 4 S P DCCP 4 S P4 DCCP 4 S P432 DCCP 4 S ST DCCP 4 S CARRY DCCP 4 S HEXDEC DCCP #32 S RSTK DCCP 4 S RSTKPTR STRING { ¢LÒ¥ARM!¢ % THIS IS L, Ò (CHAR 179), THE ANGLE (CHAR 128), ARM! !ARM *ARMSTART % ARM CODE OR DATA HERE *MAIN % ARM ENTRY POINT % MORE CODE OR DATA $(8) 0 % NO SECONDARY E.P. $(8) MAIN-ARMSTART % ENTRY PT $(8) 0 % BASE ADDRESS ¢L3v1¢ % TABLE ID , ASCII STRING L3v1 (CASE IS IMPORTANT!) !ASM } @ I hope this is useful to somebody, Claudio Reading web pages like the ARM Assembly Programming on the HP49G+ i understood that its possible to run ARM code embedding it in saturm asm code as described in that page. Thats necessary because hp49g+ emulates a saturn and then expects a program to be saturn instructions. But reading other document posted here i see there is an ARM mode in the MASD. So what kind of code is generated by this ARM mode MASD, i mean, does it need the same trick to run :-/ ? > Yes. It's not exactly trivial to get ARM code running. You have to make > sure its word-aligned (eg the address of each instruction is a multiple > of 4). Then you move the address of the ARM code into the C register, > and run ARMSAT. > The example code included with the MASD docs does the word-alignment in > a slightly different way. It copies the ARM code to 0x80100. This > location is unused, word-aligned, and is 250 bytes long. So you can't > use it for any non-trivial programs. > I don't see much point in programming assembler. It takes too long to > do. The only real advantage is small size - but there is quite a lot of > storage for programs anyway. You can also use MASD directly on the > calculator, if you enjoy entering data on a tiny screen/keyboard :) > Al === Subject: Re: confused about ARM hello, you might want to add a $80BEE (cash flush) before the ARMSAT, it might help sometimes... Just a small addition, since I know many people may be interested in the following code snippets, both for the new MASD: 1) The following code is a minimalist launcher, similar to the one found in the docs, but without the limitation in the size of the program (which Al correctly pointed out) when using 80100. Of course, you need enough free ram (not checked here) to contain a copy of the ARM program + 128 bytes. CODE % SMALL ARM LAUNCHER SAVE GOSUB SKIPCODE % INSERT ARM CODE HERE !ARM % TEST CODE - RETURN ONLY MOV PC,LR !ASM *SKIPCODE A=PC A=A-4.A C=RSTK ACEX.A D0=C % END OF BLOCK C=C-A.A % NIBBLES TO COPY D1=(2) 0 % ALIGN MOVEUP CD1EX ARMSAT LOADRPL ENDCODE @ 2) The following is a generic template based on that doc which generates a string to be executed by the ARMToolBox launcher. !ASM !NO CODE CP=0 DCCP #1028 SAT_READ DCCP #1028 SAT_WRITE DCCP #260 SAT_PRIOR DCCP 4 S_AL DCCP 4 S_AH DCCP 4 S_BL DCCP 4 S_BH DCCP 4 S_CL DCCP 4 S_CH DCCP 4 S_DL DCCP 4 S_DH DCCP 4 S_R0L DCCP 4 S_R0H DCCP 4 S_R1L DCCP 4 S_R1H DCCP 4 S_R2L DCCP 4 S_R2H DCCP 4 S_R3L DCCP 4 S_R3H DCCP 4 S_R4L DCCP 4 S_R4H DCCP 4 S_D0 DCCP 4 S_D1 DCCP 4 S_P DCCP 4 S_P4 DCCP 4 S_P432 DCCP 4 S_ST DCCP 4 S_CARRY DCCP 4 S_HEXDEC DCCP #32 S_RSTK DCCP 4 S_RSTKPTR STRING { ¢LÒ?ARM!¢ % THIS IS L, Ò (CHAR 179), THE ANGLE (CHAR 128), ARM! !ARM *ARMSTART % ARM CODE OR DATA HERE *MAIN % ARM ENTRY POINT % MORE CODE OR DATA $(8) 0 % NO SECONDARY E.P. $(8) MAIN-ARMSTART % ENTRY PT $(8) 0 % BASE ADDRESS ¢L3v1¢ % TABLE ID , ASCII STRING L3v1 (CASE IS IMPORTANT!) !ASM } @ I hope this is useful to somebody, Claudio Reading web pages like the ARM Assembly Programming on the HP49G+ i understood that its possible to run ARM code embedding it in saturm asm code as described in that page. Thats necessary because hp49g+ emulates a saturn and then expects a program to be saturn instructions. But reading other document posted here i see there is an ARM mode in the MASD. So what kind of code is generated by this ARM mode MASD, i mean, does it need the same trick to run :-/ ? > Yes. It's not exactly trivial to get ARM code running. You have to make > sure its word-aligned (eg the address of each instruction is a multiple > of 4). Then you move the address of the ARM code into the C register, > and run ARMSAT. > The example code included with the MASD docs does the word-alignment in > a slightly different way. It copies the ARM code to 0x80100. This > location is unused, word-aligned, and is 250 bytes long. So you can't > use it for any non-trivial programs. > I don't see much point in programming assembler. It takes too long to > do. The only real advantage is small size - but there is quite a lot of > storage for programs anyway. You can also use MASD directly on the > calculator, if you enjoy entering data on a tiny screen/keyboard :) > Al === Subject: Re: confused about ARM <42d6f4c8$0$10837$5a62ac22@per-qv1-newsreader-01.iinet.net.au> posting-account=FCAFHAwAAADaPCjuTWXm3ER18SIzldtO I advice everybody using my code to revise it as follows: ... MOVEUP CD1EX $80BEE ARMSAT LOADRPL ... But... this is for 49G+ only. Remove the $80BEE for use in HP48GII or HP39G+. Claudio === Subject: Re: confused about ARM > But... this is for 49G+ only. Remove the $80BEE for use in HP48GII or > HP39G+. Hum.. More accurately, it's only for ROM 2.0 and above. On the other hand, for the other ARM based calculator it should do nothing in theory JY === Subject: Re: confused about ARM <42d6f4c8$0$10837$5a62ac22@per-qv1-newsreader-01.iinet.net.au> <3k3brkFs1pteU1@individual.net> posting-account=FCAFHAwAAADaPCjuTWXm3ER18SIzldtO But... this is for 49G+ only. Remove the $80BEE for use in HP48GII or HP39G+. > Hum.. More accurately, it's only for ROM 2.0 and above. > On the other hand, for the other ARM based calculator it should do > nothing in theory I see, that's excellent! I assumed that an invalid opcode would crash. I'll use it for all calcs then. Claudio === Subject: Re: confused about ARM I haven't had time to look at it yet but what is the $80BEE doing? Arnaud === Subject: Re: confused about ARM hello, typically, to run ARM code, you do as follow: copy the ARM code at a known 4byte allign address jump to the ARM code... now, if you dig in more details, copying the ARM code is a memory write operation to execute the ARM code, the ARM cpu needs to read memory for execution if you dig even deeper, you will see that the ARM has cache memory, I do not know if you are familiar with cache or not, so here is a breif explanation: Cache memory is a trick used to speed up memory access by having a copy of memory data in a small memory area which is located close to the CPU. think about it as a pantry in which you have a small amount of often used products (salt, shugal, flour...) so you do not need to go to the store every time you need one of them.... or like a trash can, so you can drop stuff easely without have to go all the way to the recycle bin outside your house... well, the ARM has 2 such caches, one is for memory access, one if for instruction (ie: if you do a memory wread or write, the Data cache (DCache) is used, if you execute code, the Instruction cache is used).... the ARM Dcache works as follow: when you write data in memory, if the data was already in the cache, the data in the cache gets updated, else, the memory itsenf is updated. When you read data from memory, if the data was in the cache, you get the data from there, if the data was not in the cache, a space in the cache is cleaned up by copying the data back in the main memory, and then the new data is loaded in the cache and to the CPU.... (old data is choosen to be cleaned)... now, the problem should be more visible, when you write your ARM code in memory, only the DCache is modified, not the actuall memory, then when you try to execute the ARM code, the ICache will be loaded FROM MAIN MEMORY, the cache did not need to be cleaned yet), causing you to execute something different than what you thought you were executing! the $80BEE instruciton causes the ARM to clean all of the D-Cache and to invalidate the I-Cache in order to be sure that the memory is up to date.... hope that this helps... > I haven't had time to look at it yet but what is the $80BEE doing? > Arnaud === Subject: Re: confused about ARM just read the assembly coding part where it was not obvious there was a cache) Arnaud === Subject: Re: confused about ARM <42d6f4c8$0$10837$5a62ac22@per-qv1-newsreader-01.iinet.net.au> <3k3brkFs1pteU1@individual.net> posting-account=FCAFHAwAAADaPCjuTWXm3ER18SIzldtO > just read the assembly coding part where it was not obvious there was a > cache) > Arnaud Just to complete the thread, and to help other people to avoid struggling with cache bugs, here is a demo (a piece of code is worth more than a thousand words) of what kind of effects (besides headaches) you can expect from cache problems: (you need to have the development library attached) << #08041110h 1234567890 POKEARM #08041110h #5h PEEKARM >> store it for further tests: 'TESTPROG' STO The program above pokes 1234567890 to an ARM address, and then reads the same address to see what number was stored there. It should (obviously) return 1234567890 to the stack. If you run it once you see the right result. Now run it repeatedly: << 0. 100. FOR I TESTPROG NEXT >> (or execute TESTPROG directly from the keyboard several times) Look at the results on the stack (scroll up) to find that not all the results are 1234567890. There's nothing wrong with the UserRPL program, also both POKEARM and PEEKARM are correct, bug-free, etc. and still they fail from time to time. If you add: CODE $80BEE RPL ENDCODE before POKEARM and PEEKARM in the example, the bug goes away. The problem is exactly as Cyrille described, and things like this can happen with the small launcher I posted earlier if you don't add the $80BEE. This is a kind of bug that can give you nightmares, so it's better to avoid. Claudio === Subject: Re: confused about ARM <42d6f4c8$0$10837$5a62ac22@per-qv1-newsreader-01.iinet.net.au> <3k3brkFs1pteU1@individual.net> posting-account=O6knhQ0AAAD8zE21TyMVwtCasYnRdHOm so the launcher that comes in ARMtoolbox in hpgcc will need this fix too? === Subject: Re: confused about ARM <42d6f4c8$0$10837$5a62ac22@per-qv1-newsreader-01.iinet.net.au> <3k3brkFs1pteU1@individual.net> posting-account=FCAFHAwAAADaPCjuTWXm3ER18SIzldtO > so the launcher that comes in ARMtoolbox in hpgcc will need this fix > too? Yes, but only in some cases (the fix is coming with the next release anyway). If (and only if): 1) you use it to launch your own ARM assembler programs (I mean non-hpgcc) and 2) you are using the program as a string in the stack (program is not stored in a variable or in a port) and 3) you try to run 2 different programs one immediately after another then you may encounter something like this. What I experienced (with early versions of the Toolbox) was that sometimes programs returned immediately without (apparently) being executed. It only happens with small programs executed rapidly in sequence. Large programs don't have problems because the processor cache is very small, and the problem occurs when there is a cache hit between two different programs. The problem cannot happen with hpgcc programs, because they flush the caches at startup. Claudio === Subject: Re: confused about ARM <42d6f4c8$0$10837$5a62ac22@per-qv1-newsreader-01.iinet.net.au> <3k3brkFs1pteU1@individual.net> posting-account=FCAFHAwAAADaPCjuTWXm3ER18SIzldtO > I haven't had time to look at it yet but what is the $80BEE doing? > Arnaud It flushes the ARM processor caches. It avoids problems related to the particular behavior of the ARM cache, which cause applications to act erratically sometimes, with possible unexpected, hard-to-explain crashes. Claudio === Subject: Re: confused about ARM > CODE > % SMALL ARM LAUNCHER > SAVE > GOSUB SKIPCODE > % INSERT ARM CODE HERE > !ARM > % TEST CODE - RETURN ONLY > MOV PC,LR > !ASM > *SKIPCODE [crazy code snipped] This is the best receipe to a pretty nasty crash and memory loss. Ever wondered what would happen to the 128 bytes located *after* your program's end? Either you allocate memory and then do your manipulation or you make your program 128 bytes bigger than what it should (in which case it will only work once anyway being self modifying and is therefore a really bad idea). Jean-Yves === Subject: Re: confused about ARM <42d6f4c8$0$10837$5a62ac22@per-qv1-newsreader-01.iinet.net.au> <3jqtplFrav29U1@individual.net> posting-account=FCAFHAwAAADaPCjuTWXm3ER18SIzldtO <...> > This is the best receipe to a pretty nasty crash and memory loss. > Ever wondered what would happen to the 128 bytes located *after* your > program's end? > Either you allocate memory and then do your manipulation or you make > your program 128 bytes bigger than what it should (in which case it will > only work once anyway being self modifying and is therefore a really bad > idea). > Jean-Yves Basically, the launcher copies the ARM code below the calculator stack. The 128 bytes I mentioned is simply because I word-align the program by using D1(2)=0, which will waste up to 128 bytes. Let's see with more detail... GOSUB SKIPCODE ... *SKIPCODE A=PC A=A-4.A (get the address of the *skipcode label = end of ARM code) C=RSTK (get the address of the start of the ARM code) ACEX.A D0=C % END OF BLOCK (D0=end of ARM CODE) C=C-A.A % NIBBLES TO COPY (=end-start) D1=(2) 0 % ALIGN (D1 IS POINTING TO THE CALCULATOR STACK, ROUND DOWN TO WORD-ALIGN) (SO FAR, HERE I HAVE D0=END OF SOUCE, D1=END OF DESTINATION, C=NUMBER OF NIBBLES, SO EVERYTHING LOOKS GOOD) MOVEUP (NOW HERE D1=START OF ARM CODE AT DESTINATION ADDRESS, LAST ADDRESS CD1EX (C=START OF ARM CODE) ARMSAT I still can't see anything wrong with my code, and I used it several times. I think you misinterpreted what the code does, or I have a big mistake I'm not aware of. If you see something wrong, please point me where the problem is and I'll fix it. Claudio === Subject: Re: confused about ARM <42d6f4c8$0$10837$5a62ac22@per-qv1-newsreader-01.iinet.net.au> posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo Why not just move one level up from machine code/sysRPL or higher in order to get to generic suedo code and then write it in Forth ARM ? 1) get a very basic Forth running under ARM / 49G+ 2) write the Saturn stuff into Forth . === Subject: Re: confused about ARM posting-account=lAaz6QwAAABR-WEJBEnTuXipZ8zCV0L0 If you use the ARM mode of MASD, it will generate native ARM code. However the system is not designed to cope directly with native ARM code, it always emulate saturn code. To run native ARM code you need to locate it at a byte aligned address and then use a special saturn instruction to jump there and treat the code as ARM. So if you generate some native ARM code, you will have to either embed it in a SATURN code that will do the job described above or use an ARM launcher to do it for you. Arnaud === Subject: ClckAdjst v. 2.8 Hello together, The automatic adjustment in ClckAdjst v. 2.8 is not working accurately. I found the bug and I am testing a new version. For those who are interested in automatic clock adjustment, time zone management and automatic daylight saving time adjustment for HP 49G(+), please e-mail me. Heiko === Subject: Help, defined integral... posting-account=eR28NA0AAADbgJXkhd6TIRlsyrXFXlWO Hello.... Somebody can follow the links of this discussion and to help me... http://www.adictoshp.org/modules/newbb/viewtopic.php?topic_id=3401&forum=12& post_id=21844#forumpost21844 === Subject: Does HP-48SX SOLVER have the IF function for algebraic formulas posting-account=XZBzkw0AAACvm4tUXfyLZUkjp5CZRheR 1. The short version of the question is does the HP-48SX (ROM version E) support the IF function in algebraic formulas? It doesn't appear to. It does appear to support LET L and GET G with LET using comma for the delimiter, but these don't appear to be documented. Up until this point I've mainly used the SOLVER on the HP-17BII. It seems hard to believe that it supports IF and the HP-48SX does not. If it makes any difference I am trying to port Tony Hutchins Black-Scholes package that runs on the HP-200LX. It's clear from looking at the equations in Tony's package that the HP-200LX SOLVER has other changes. It allows identifiers that start with '#'. The syntax for IF uses semi-colon as the delimiter. LET appears to accept both comma and semi-colon as delimiters and I don't know if there is a difference between the two. 2. Is there any kind of documentation that describes the differences between the various versions of the SOLVER? leb === Subject: Re: Does HP-48SX SOLVER have the IF function for algebraic formulas you can use IFTE in algebraics, the short form of 'IF-THEN-ELSE' . The parameters and usage should be covered in the HP-48 manual. Silly example: 'IFTE(IP(SIN(90)),0,10)' In DEG mode the result will be 0, in RAD mode it will be 10. Raymond schrieb im Newsbeitrag > 1. The short version of the question is does the HP-48SX (ROM version > E) support the IF function in algebraic formulas? It doesn't appear > to. > It does appear to support LET L and GET G with LET using comma for > the delimiter, but these don't appear to be documented. > Up until this point I've mainly used the SOLVER on the HP-17BII. It > seems hard to believe that it supports IF and the HP-48SX does not. > If it makes any difference I am trying to port Tony Hutchins > Black-Scholes package that runs on the HP-200LX. It's clear from > looking at the equations in Tony's package that the HP-200LX SOLVER has > other changes. It allows identifiers that start with '#'. The syntax > for IF uses semi-colon as the delimiter. LET appears to accept both > comma and semi-colon as delimiters and I don't know if there is a > difference between the two. > 2. Is there any kind of documentation that describes the differences > between the various versions of the SOLVER? > leb === Subject: Experience with changing softkeys in INFORM? Hi. I have another small problem with the new inform engine. I handle message 12 in the inform box message handler - with this I replace the three right-most softkeys in the inform box. The problem is that my new key definitions are only active when field zero is selected. When other fields are selected DoBadKey is executed instead. Why is this? It seems that message 12 is never sent to the field's individual message handlers, so I can't handle it in other places than in the inform box's message handler. Steen === Subject: Re: Experience with changing softkeys in INFORM? > The problem is that my new key definitions are only active when field > zero is selected. When other fields are selected DoBadKey is executed > instead. Why is this? I replied to this one already: TakeOver is the key > It seems that message 12 is never sent to the > field's individual message handlers, so I can't handle it in other > places than in the inform box's message handler. The menu definition is a system-wide call, so it's only called when you have to redisplay the whole inform box. If you want to have the menu that changes per field, you will have to make the menu a program that will make the content change depending let say a local variable. Like you menu will look like: { :: TakeOver menuvalue@ #1= ITE Menu1 Menu2 ; :: TakeOver menuvalue@ #1= ITE action1 action2 ; } You would then change within the field's individual message handler the value of menuvalue (like in the even this field is selected). This will allow the menu's content to change depending on which field the cursor is located. Jean-Yves === Subject: Re: Experience with changing softkeys in INFORM? posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > The problem is that my new key definitions are only active when field > zero is selected. When other fields are selected DoBadKey is executed > instead. Why is this? It seems that message 12 is never sent to the > field's individual message handlers, so I can't handle it in other > places than in the inform box's message handler. That does seem strange. Unless this is something you don't want me to see, you could send it and I can take a look at it. There may be some little thing hidden in plain view that is messing things up. . . =( TW === Subject: Re: Experience with changing softkeys in INFORM? > That does seem strange. Unless this is something you don't want me to > see, you could send it and I can take a look at it. A simple example will illustrate something similar: :: Label BINT1 BINT10 'DROPFALSE BINT1 BINT20 BINT15 BINT6 BINT1 { BINT0 } BINT5 NULL$ MINUSONE BINT5 %0 %0 ONEONE ' :: BINT12 #=case :: { { A :: ; } { B :: ; } { C :: Test xMSGBOX ; } } TRUE ; FALSE ; Title FPTR2 ^IfMain @ A, B & C does show up in the softmenu, but pressing C has the same effect as pressing A & B (DoBadKey, instead of the message box I expect). Steen === Subject: Re: Experience with changing softkeys in INFORM? A, B & C does show up in the softmenu, but pressing C has the same > effect as pressing A & B (DoBadKey, instead of the message box I > expect). > Steen Because you want the menu to be executed without having a command line running. Add a TakeOver at the beginning on the program like: { A :: TakeOver ; } Otherwise the program will only run if you have a command line Jean-Yves === Subject: Re: Experience with changing softkeys in INFORM? > Because you want the menu to be executed without having a command > line running. Add a TakeOver at the beginning on the program like: > { A :: TakeOver ; } Steen === Subject: Conn4x Issues posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 I've recently been having some conn4x issues. When I connect the calculator to the computer, it detects it just fine with no problems at all. When I open the program, it detect my 49g+ port. When I go to connect, it flashes the commands one the screen, and then freezes. I can't push CANCL to quit until I pull the USB plug out of the calc. Then I can cancel it just fine. I really just need to connect so I can get screen captures since I never use it to transfer files. However, the screen capture locks up as well. It seems to transfer a tiny bit of the screen, and then stops. Yes I am set to to print via wire and transfer via wire. Now here is the strange part. It sometimes WORKS! I connect and start taking screenshots immediately. I take a few, and then it might die again or continue to work. restarting the computer helps, or it doesn't. I've reinstalled the driver and program, but this doesn't seem to help. This is driving me nuts! Has anyone experienced anything even remotely similar to this? TW === Subject: Re: Conn4x Issues > Now here is the strange part. It sometimes WORKS! I connect and start > taking screenshots immediately. I take a few, and then it might die > again or continue to work. restarting the computer helps, or it > doesn't. I've reinstalled the driver and program, but this doesn't > seem to help. This is driving me nuts! > Has anyone experienced anything even remotely similar to this? Try new batteries? -- - - - - - - - - - - - - - - - - Bill Graves RKBA! bgraves@ix.netcom.com === Subject: Re: Conn4x Issues posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > Try new batteries? First thing I tried. Rechargables like I normally use, and new AAA alkalines. =( TW === Subject: Re: Conn4x Issues posting-account=UO9F1g0AAADizWS3xVT-crLKDUmoJI-f If you turn on the clock. Turn it off then try again. === Subject: Re: Conn4x Issues > If you turn on the clock. Turn it off then try again. You could always turn around 3 times then spit over your left shoulder and see if it works any better. Jean-Yves === Subject: Re: Conn4x Issues <3jqt69Fr47g3U2@individual.net> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > You could always turn around 3 times then spit over your left shoulder > and see if it works any better. TW === Subject: Re: Conn4x Issues <3jqt69Fr47g3U2@individual.net> posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE You may use an SD card with an SD card reader, it's the fastes way to transfer programs. Have you tried my HPConn on hpcalc? (http://www.hpcalc.org/details.php?id=6238). It's quiet simple and not super genial GUI, but maybe it does what you need. Fred. === Subject: Re: Conn4x Issues <3jqt69Fr47g3U2@individual.net> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 I have two SD readers. . . unfortunately I just need to get screenshots. Does that capture them? TW === Subject: Debug4x and Directories - change is coming For Directories on the calculator, Debug4x in current release does not set/debug with breakpoints because the addresses are all wrong. In the next release, I plan to remove the Project is a Directory option and replace the capability with a SASM Macro. The new way will allow full debug for directories. It has these limitations: replaced with a Dir macro reference. 2. The entire directory must be contained in one source file - INCLUDE may be used to separate some source. 3. Any file that is INCLUDED cannot be debugged automatically from the source listing (all other parts of that program can still be debugged with source). I plan to remove the option Project is a Directory shortly. Please email me details if that makes a problem for you. ------------------------------------------------------------- For those of you who understand SASM, here is the macro. Your comments are appreciated. The macro does rely on a few behaviors of the SASM = statements. DirMacro.s ************************************************************** * Directory building macro * * Invoke this way: * ASSEMBLE * Dir name * saturn code (if any) * RPL * :: stuff ; (if any) * * All directory entries must be in the same file * All material in the file must be part of the same * directory. ************************************************************** ASSEMBLE Dir MACRO IFNDEF $dir$st CON(5) =DORRP CON(3) #7FF * Number of attached libraries REL(5) $dir$st * Offset to last directory entry CON(5) 0 * constant 00000 ELSE CON(5) (*)-$dir$st * size of the previous entry ENDIF $dir$st = * * used by next call to macro CON(2) (($dir$nm$0)-2-(*))/2 NIBASC $1 $dir$nm$0 = * * must be unique because of fwd reference CON(2) ((*)-2-$dir$st)/2 ENDM * * Sample Usage to Build a complete directory * with two verbs * ASSEMBLE Dir ABC ( name of the variable ) RPL :: ABC ( just put ABC on the stack ) ; ASSEMBLE Dir 456 RPL :: 456. ( just put 456. on stack ) ; ---------------------------------------------------------------------------- -------------------- -- - - - - - - - - - - - - - - - - Bill Graves RKBA! bgraves@ix.netcom.com === Subject: GBASIS corrupting HP49G+ memory? posting-account=hYpJhAwAAABbgx9xHixEB8na9TYTFvkV I just tried a Groebner basis computation on my 49G+ (in RPN mode). Everything seemed to be going ok at first (the area just below the header showed GBASIS 3 PAIRS 3/3), but in a minute or so the display got messed up (for example, the softkey menu moved into the middle of the screen), and the computation, which should have been fairly easy even for lexicographic ordering, appears to be hanging (the hourglass is displayed in the header area, and the calculator does not respond to the ON key). Has anyone seen this sort of behavior (either GBASIS hanging or a math command messing up the display)? Bhuvanesh. === Subject: Re: GBASIS corrupting HP49G+ memory? posting-account=hYpJhAwAAABbgx9xHixEB8na9TYTFvkV I just came across another issue: ['X'] ['X', 'Y'] GBASIS gives a weird array [External ... ] that hangs the calculator if I press [F2] (VIEW). I expected just [X] as the result. Bhuvanesh. === Subject: Re: GBASIS corrupting HP49G+ memory? > I just came across another issue: > ['X'] > ['X', 'Y'] > GBASIS > gives a weird array [External ... objects, etc.>] that hangs the calculator if I press [F2] (VIEW). I > expected just [X] as the result. === Subject: Re: GBASIS corrupting HP49G+ memory? posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J About the solver integrated in the hp49g+.....why when I put: [EQNs.....] [vars] if I dont put all the vars in the equations upper, the solver auto check the variables??? for example [f(x,y,t)=a, g(x,y,t)=b, h(x,y,t)=c] [x,t] when I SOLVE, hp search solution for x,y,t always????!! (so, why ask for vars, if they aren't used?) === Subject: Re: GBASIS corrupting HP49G+ memory? > About the solver integrated in the hp49g+.....why when I put: > [EQNs.....] > [vars] > if I dont put all the vars in the equations upper, the solver auto > check the variables??? > for example > [f(x,y,t)=a, g(x,y,t)=b, h(x,y,t)=c] > [x,t] > when I SOLVE, hp search solution for x,y,t always????!! (so, why ask > for vars, if they aren't used?) If there are more variables than equations, the calculator needs to know which variables to try to solve for. === Subject: Re: GBASIS corrupting HP49G+ memory? > Has anyone seen this sort of behavior (either GBASIS hanging or a math > command messing up the display)? What arguments did you use? Jean-Yves === Subject: Re: GBASIS corrupting HP49G+ memory? <3jqo20Frg98nU1@individual.net> posting-account=hYpJhAwAAABbgx9xHixEB8na9TYTFvkV Has anyone seen this sort of behavior (either GBASIS hanging or a math command messing up the display)? > What arguments did you use? > Jean-Yves OK, now it's just hanging. I'm not sure what caused the display issue earlier, but I'll try to reproduce it. The polynomials are: {x-I*v1, x-v1^2, 1-x*y} (I=sqrt(-1)) The variable ordering is: {v1, y, x} I quoted everything: ['X-i*V1','X-V1^2','1-X*Y'] ['V1','Y','X'] GBASIS Bhuvanesh. === Subject: Re: GBASIS corrupting HP49G+ memory? <3jqo20Frg98nU1@individual.net> posting-account=hYpJhAwAAABbgx9xHixEB8na9TYTFvkV > Has anyone seen this sort of behavior (either GBASIS hanging or a math > command messing up the display)? What arguments did you use? Jean-Yves > OK, now it's just hanging. I'm not sure what caused the display issue > earlier, but I'll try to reproduce it. > The polynomials are: {x-I*v1, x-v1^2, 1-x*y} (I=sqrt(-1)) > The variable ordering is: {v1, y, x} > I quoted everything: > ['X-i*V1','X-V1^2','1-X*Y'] > ['V1','Y','X'] > GBASIS > Bhuvanesh. Can anyone else reproduce this? Is it a known issue? By the way, is there a Groebner basis implementation for other monomial orderings (for the HP49G(+), of course)? Bhuvanesh. === Subject: Re: GBASIS corrupting HP49G+ memory? > > Has anyone seen this sort of behavior (either GBASIS hanging or a > > math > > command messing up the display)? > What arguments did you use? > > Jean-Yves >> OK, now it's just hanging. I'm not sure what caused the display issue >> earlier, but I'll try to reproduce it. >> The polynomials are: {x-I*v1, x-v1^2, 1-x*y} (I=sqrt(-1)) >> The variable ordering is: {v1, y, x} >> I quoted everything: >> ['X-i*V1','X-V1^2','1-X*Y'] >> ['V1','Y','X'] >> GBASIS >> Bhuvanesh. > Can anyone else reproduce this? Is it a known issue? > By the way, is there a Groebner basis implementation for other monomial > orderings (for the HP49G(+), of course)? > Bhuvanesh. http://www.hpcalc.org/hp48/math/symbolic/ See the ALG48 stuff (including source code) http://www.hpcalc.org/details.php?id=1320 includes GROBNER.PS documentation The old 48 stuff is almost the same as the 49 integrated system VPN === Subject: Re: GBASIS corrupting HP49G+ memory? > OK, now it's just hanging. I'm not sure what caused the display issue > earlier, but I'll try to reproduce it. > The polynomials are: {x-I*v1, x-v1^2, 1-x*y} (I=sqrt(-1)) > The variable ordering is: {v1, y, x} > I quoted everything: > ['X-i*V1','X-V1^2','1-X*Y'] > ['V1','Y','X'] > GBASIS > Bhuvanesh. It was coded by Mika, but I guess the problem could be the i since ALG48 code worked for real polynomials. I don't think there is a check for the presence of imaginary numbers in the arguments of GBASIS. Does it work with a real value instead of i? === Subject: Re: GBASIS corrupting HP49G+ memory? <3jqo20Frg98nU1@individual.net> <42d915d3$0$3135$8fcfb975@news.wanadoo.fr> posting-account=hYpJhAwAAABbgx9xHixEB8na9TYTFvkV OK, now it's just hanging. I'm not sure what caused the display issue earlier, but I'll try to reproduce it. The polynomials are: {x-I*v1, x-v1^2, 1-x*y} (I=sqrt(-1)) The variable ordering is: {v1, y, x} I quoted everything: ['X-i*V1','X-V1^2','1-X*Y'] ['V1','Y','X'] GBASIS Bhuvanesh. > It was coded by Mika, but I guess the problem could > be the i since ALG48 code worked for real polynomials. > I don't think there is a check for the presence > of imaginary numbers in the arguments of GBASIS. > Does it work with a real value instead of i? Yes, it seems to work for real values as far as I can tell. Is the real polynomials restriction due to data structure(s) used in the implementation for efficiency? Bhuvanesh. === Subject: Re: GBASIS corrupting HP49G+ memory? > Yes, it seems to work for real values as far as I can tell. > Is the real polynomials restriction due to data structure(s) used in > the implementation for efficiency? Most certainly. ALG48 code has internal dense recursive representation for polynomials with integer coefficients. I added code for Gaussian integers but only for basic arithmetic. I never tried to port Gaussian integers inside Mika's assembler Groebner basis code. It is most certainly not worth the effort since this kind of computation require much more processor speed and memory that is available on the 49(+). === Subject: Another Hp49G+ bug? Please see if you get it. posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt The applied part ---------------- I just came across what looks like a bug on my HP49G+ (CNA41, ROM2.0, RPN mode). Can anyone else see it? Create a function using define, something simple like H(X)=X+1 and then H appears as a soft key. Build a list of numbers {1 2 3} for example, pop it on the stack and then press the H key what I think I should get is {2 3 4} what I get is {2 3 4 1} I tried it with several combinations of function and list, and it seems that the list is having a 0 appended to the end which the function is then processing. The philosophical part ---------------------- I am wondering if I can actually keep using this calculator. With the dddouble ppress issue I need to be contantly vigilant that the calculator isn't doing things I don't want. For comparison - I have a CASIO fx3400p which has NEVER missed a key or had a double press in 15 years. I am aware of the FIX bug, and now I think there is yet another problem as above. What else is lurking in this machine? Can I ever actually trust what it is doing? How to I explain to my investors that we wasted $50,000 making a silicon chip and it doesn't work, because my calculator got the maths wrong? My casio was primitive in comparison but I absolutely trust any answer it gives me. HP - what has happened to you. How can you go from such heights to this level of crummy product? Do you still have a QA department? You should be ashamed. Adrian === Subject: Re: Another Hp49G+ bug? Please see if you get it. > The applied part > ---------------- > I just came across what looks like a bug on my HP49G+ > (CNA41, ROM2.0, RPN mode). Can anyone else see it? > Create a function using define, something simple like H(X)=X+1 > and then H appears as a soft key. > Build a list of numbers {1 2 3} for example, pop it on the stack > and then press the H key > what I think I should get is {2 3 4} > what I get is {2 3 4 1} Try replacing any plus signs in your function by the ADD command. The HP interpretes plus signs sort of like append commands when dealing with lists, but has provided the ADD command which does not mess things up when used with lists. === Subject: Re: Another Hp49G+ bug? Please see if you get it. >> The applied part >> ---------------- >> I just came across what looks like a bug on my HP49G+ >> (CNA41, ROM2.0, RPN mode). Can anyone else see it? >> Create a function using define, something simple like H(X)=X+1 >> and then H appears as a soft key. >> Build a list of numbers {1 2 3} for example, pop it on the stack >> and then press the H key >> what I think I should get is {2 3 4} >> what I get is {2 3 4 1} > Try replacing any plus signs in your function by the ADD command. The HP > interpretes plus signs sort of like append commands when dealing with > lists, but has provided the ADD command which does not mess things up > when used with lists. Have anyone tried -- that is: subtract the negative (or minus & change sign) Veli-Pekka === Subject: Re: Another Hp49G+ bug? Please see if you get it. > The applied part > ---------------- > I just came across what looks like a bug on my HP49G+ > (CNA41, ROM2.0, RPN mode). Can anyone else see it? > Create a function using define, something simple like H(X)=X+1 > and then H appears as a soft key. > Build a list of numbers {1 2 3} for example, pop it on the stack > and then press the H key > what I think I should get is {2 3 4} > what I get is {2 3 4 1} { 1 2 } 3 + and you get: { 1 2 3 } { 1 2 } 3 ADD and you get. { 4 5 } The calculator is acting as expected: + add the element to the list. If you want to add the value to all arguments of the list: use ADD like H(X)=X ADD 1 , DEFINE If you rely on a $130 calculator machine to make business decision worth $50,000 then I would question your working practice really... Jean-Yves === Subject: Re: Another Hp49G+ bug? Please see if you get it. <3jqt43Fr47g3U1@individual.net> posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt > The calculator is acting as expected I can only assume this is meant to amuse? :-) X+1 does not give 2,3,4 when X = 1,2,3? I am at home suffering from a fever (honestly) and my mind is racing and acting a little odd, but I didn't think I was this bad. I better go back to bed and take some more pills. I'll drop in again on Monday. Adrian === Subject: Re: Another Hp49G+ bug? Please see if you get it. > I can only assume this is meant to amuse? :-) Not at all. > X+1 does not give 2,3,4 when X = 1,2,3? That's parallel list processing - an entirely different feature. Steen === Subject: Re: Another Hp49G+ bug? Please see if you get it. <3jqt43Fr47g3U1@individual.net> posting-account=msh5ow0AAACgGZihasI14P8AUGVn2RWK > If you rely on a $130 calculator machine to make business decision worth > $50,000 then I would question your working practice really... > Jean-Yves WTF kind of reply is that?! NASA used the HP-41 on missions to the moon, for Pete's sake! Jean-Yves, I'm sick of your arrogant, less than helpful attitude! Get a grip! === Subject: Re: Another Hp49G+ bug? Please see if you get it. >>If you rely on a $130 calculator machine to make business decision worth >>$50,000 then I would question your working practice really... >>Jean-Yves > WTF kind of reply is that?! NASA used the HP-41 on missions to the > moon, for Pete's sake! Jean-Yves, I'm sick of your arrogant, less than > helpful attitude! Get a grip! Yeah, but they didn't use it as the *sole* means of calculating stuff, did they? Now, whenever I do important calculations I always always use three different devices and at least two methods, if available. Never, ever, do I rely solely on Excel, although I haven't had a major booboo in Excel. Check, check and check again! HOWEVER, I have to agree with Adrian in that if I cannot trust a piece of kit (any kit really) then I have to think carefully why I should continue to use it. I mean, come on! It's not as if calculators are so damnably complicated that they can't be done right. I have a couple of TIs and Casios and they don't give me any problems; what's so special about HP that they think they can fob people off with shoddy products! :wave: JasonG === Subject: Re: Another Hp49G+ bug? Please see if you get it. <3jqt43Fr47g3U1@individual.net> posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt True enough. I don't ever rely on a single point of failure whenever I can help it, normally things go through at least one circuit simulator, often two. But its not always possible. For example, I am calculating the inductance of a spiral antenna, ideally suited to a calculator because it is a manual one-off calculation. That will probably never go though a simulator because we don't have the necessary field solver. So the answers my calculator will give will go into the product and output drivers will be sized accordingly. But dammit, this calculator isn't a $2 lump of crap out of my breakfast flakes. In Europe these things retail at around $230! For that price I can get a low end laptop and open source SW to do a spread sheet. Maybe I should? Ironically enough - I tried a linear regression recently under excel, and got an asnwer that looked a little off. I ran the numbers though the HP49G+ and the old casio - and guess who was wrong of the 3? Excel. You have to laugh. Adrian === Subject: Re: Another Hp49G+ bug? Please see if you get it. > True enough. > I don't ever rely on a single point of failure whenever I can help it, > normally things go through at least one circuit simulator, often two. > But its not always possible. > For example, I am calculating the inductance of a spiral antenna, > ideally > suited to a calculator because it is a manual one-off calculation. That > will probably never go though a simulator because we don't have the > necessary field solver. So the answers my calculator will give will go > into the product and output drivers will be sized accordingly. > But dammit, this calculator isn't a $2 lump of crap out of my breakfast > flakes. In Europe these things retail at around $230! For that price I > can > get a low end laptop and open source SW to do a spread sheet. Maybe I > should? > Ironically enough - I tried a linear regression recently under excel, > and got > an asnwer that looked a little off. I ran the numbers though the HP49G+ > and the old casio - and guess who was wrong of the 3? Excel. You have > to laugh. > Adrian This is why I don't trust my HP Handheld PC with Excel I use the 49g+ kkeeyybbooaarrdd instead VPN PS: I'm not laughing, but could I have your Casio, please?! :) === Subject: Re: Another Hp49G+ bug? Please see if you get it. <3jqt43Fr47g3U1@individual.net> posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt Sell my Casio? NO WAY MAN!!!! it is the lovely fx-3400P. Sadly no longer sold by casio. Adrian === Subject: Re: Another Hp49G+ bug? Please see if you get it. >>True enough. >>I don't ever rely on a single point of failure whenever I can help it, >>normally things go through at least one circuit simulator, often two. >>But its not always possible. >>For example, I am calculating the inductance of a spiral antenna, >>ideally >>suited to a calculator because it is a manual one-off calculation. That >>will probably never go though a simulator because we don't have the >>necessary field solver. So the answers my calculator will give will go >>into the product and output drivers will be sized accordingly. >>But dammit, this calculator isn't a $2 lump of crap out of my breakfast >>flakes. In Europe these things retail at around $230! For that price I >>can >>get a low end laptop and open source SW to do a spread sheet. Maybe I >>should? >>Ironically enough - I tried a linear regression recently under excel, >>and got >>an asnwer that looked a little off. I ran the numbers though the HP49G+ >>and the old casio - and guess who was wrong of the 3? Excel. You have >>to laugh. >>Adrian > This is why I don't trust my HP Handheld PC with Excel > I use the 49g+ kkeeyybbooaarrdd instead > VPN > PS: I'm not laughing, but could I have your Casio, please?! :) Excel is make for bean counters, not for (so much) advanced math in scientic field (but many people try nevertheless) ( i remeber old time Excel 1.0 on mac, and diagrams view is impossible to using logaritm scale on Y-axe but possible on X-axe - and I calculate electrical filters... Ghaaaaaaa) and later excel version (98?) find in formula with heavy using sin, cos, tan. nl and log etc. and result not be comparable and wide away from simular program make in pascal (using same formula) - and after counting value number, come to insight excel using only single precision float inside in least one unknow step in calculate chain of formula... after this is not using excel for more advanced problems... I don't trust excel and untrusted tool is a usuless tools. /TE === Subject: Re: Another Hp49G+ bug? Please see if you get it. > Excel is make for bean counters, not for (so much) advanced math in > scientic field (but many people try nevertheless) I don't classify myself has a bean counters and I use Excel almost every day for many different things > I don't trust excel > and untrusted tool is a usuless tools. Well I do, it's the best all-around application in the computer world IMO. I've done almost everything with it from generating invoices to analyzing survey's result or re-calculating my mobile phone's invoice. Jean-Yves === Subject: Re: Another Hp49G+ bug? Please see if you get it. >WTF kind of reply is that?! NASA used the HP-41 on missions to the >moon, No, not to moon. Justz for space shuttle missions. One for steering the robotic arm, two others they had on board for computer-replacement if necessary. Volker -- Besides, i'm of the opinon, that TCPA has to be stopped === Subject: Re: Another Hp49G+ bug? Please see if you get it. >>WTF kind of reply is that?! NASA used the HP-41 on missions to the >>moon, > No, not to moon. Justz for space shuttle missions. One for steering > the robotic arm, two others they had on board for computer-replacement > if necessary. Was it the HP-65 or 67 that was used on Apollo-Soyuz missions? Tom Lake === Subject: Re: Another Hp49G+ bug? Please see if you get it. >> No, not to moon. Justz for space shuttle missions. One for steering >> the robotic arm, two others they had on board for computer-replacement >> if necessary. >Was it the HP-65 or 67 that was used on Apollo-Soyuz missions? That' a point i'm not aware of. I don't even know, _if_ they had used calcs on those missions. Volker -- Besides, i'm of the opinon, that TCPA has to be stopped === Subject: Re: Another Hp49G+ bug? Please see if you get it. > No, not to moon. Justz for space shuttle missions. One for steering > the robotic arm, two others they had on board for computer-replacement > if necessary. >>Was it the HP-65 or 67 that was used on Apollo-Soyuz missions? > That' a point i'm not aware of. I don't even know, _if_ they had used > calcs on those missions. I looked it up. It was the HP-65 http://www.hpmuseum.org/adverts/sa65spc.htm Tom Lake === Subject: Re: Another Hp49G+ bug? Please see if you get it. > WTF kind of reply is that?! NASA used the HP-41 on missions to the > moon, for Pete's sake! Jean-Yves, I'm sick of your arrogant, less than > helpful attitude! Get a grip! I can't speak for Jean-Yves, but if I were he, as a (the?) major contributor to the 49g+ ROM, I'd be getting more then just a little pissed off at people constantly shouting BUG BUG BUG (which is tantamount to a direct criticism of the quality of his work), when 9 times out of 10 the software is acting as it should, and it's the user who doesn't understand how to use the product properly. I for one have learned that it's very easy to criticise - especially when I didn't know the full story. === Subject: Re: Another Hp49G+ bug? Please see if you get it. > I can't speak for Jean-Yves, but if I were he, as a (the?) major contributor > to the 49g+ ROM, I'd be getting more then just a little pissed off at people > constantly shouting BUG BUG BUG (which is tantamount to a direct criticism > of the quality of his work), when 9 times out of 10 the software is acting > as it should, and it's the user who doesn't understand how to use the > product properly. > I for one have learned that it's very easy to criticise - especially when I > didn't know the full story. Perhaps design fault would be more accurate than bug, unless you're a backwards-compatibility freak. (How many people have list-processing User-RPL programs which they still need, don't have the source for and cannot easily reprogram?) It is certainly an annoyance which should not be present. === Subject: Re: Another Hp49G+ bug? Please see if you get it. > Perhaps design fault would be more accurate than bug, unless you're a > backwards-compatibility freak. (How many people have list-processing > User-RPL programs which they still need, don't have the source for and > cannot easily reprogram?) > It is certainly an annoyance which should not be present. Ok.. Let's take a big breath... What about all the user programs that are adding an object to a list (which in my experience happens more often that doing any list processing). This is the way it was done on the hp48s/sx before list processing was ever introduced. How would you add an object to a list? There's no design fault here, that is how the calculator has been behaving since it was first introduced in 1989. Jean-Yves === Subject: Re: Another Hp49G+ bug? Please see if you get it. > Ok.. Let's take a big breath... Can I breathe out yet? I'm asphyxiating. > What about all the user programs that are adding an object to a list > (which in my experience happens more often that doing any list processing). > This is the way it was done on the hp48s/sx before list processing was > ever introduced. Overloading the + operator in this way was a poor design decision. Breaking backwards compatibility (in a way that is extremely easy to fix, just change << list object + >> to << list object ADD >>) would have been better than introducing this horrible inconsistency for the future. === Subject: Re: Another Hp49G+ bug? Please see if you get it. Borris schrieb im Newsbeitrag >> Ok.. Let's take a big breath... > Can I breathe out yet? I'm asphyxiating. >> What about all the user programs that are adding an object to a list >> (which in my experience happens more often that doing any list >> processing). >> This is the way it was done on the hp48s/sx before list processing was >> ever introduced. > Overloading the + operator in this way was a poor design decision. > Breaking backwards compatibility (in a way that is extremely easy to fix, > just change << list object + >> to << list object ADD >>) would have been > better than introducing this horrible inconsistency for the future. Maybe the original list 'adding' feature (+) seems to be inconsistent now, but actually it's very consistent. It always yields the same type of result for any object type combinations, whereas ADD only works for a subset of object types. Take a list with arbitrary objects in level 2, like { ABC 'Foo' 2 [2 3] } and another object in level 1, like a number, or worse, a string. Take for example the string Bar. 2: { ABC 'Foo' 2 [2 3] } 1: Bar What would be the most consistent way to combine the two objects? Would you like any included list object to be broken by adding Bar to it? This would give s.t. like { ABCBar FooBar 2Bar [2 3]Bar }, at least if it worked that way... I guess this won't be the desired result, apart from the fact that some object types can't be combined in an easy way. In the above hypothetical example, most of the list objects would have been trashed after an attempt to 'add' the level 1 object to them. However, to get around this limitation, kinda list processing was added in the HP-48 G series, to be able to really 'add' an object to each list object, leading to the result shown in the example. For numerical values it will make sense, but other object types, like arrays or identifiers, will either be converted to strings, or still produce an error message, if the objects can't be combined. Example: 2: { ABC 'Foo' 2 [2 3] } 1: 2 Performing ADD will lead to an error, since 'Foo' can't be combined with the real number 2 . So maybe the original design of + for list and object wasn't optimal from list processing view, but back then it was the way to go, and still is consistent, as + 'adds' an object to the list, which then becomes a list member. Raymond === Subject: Re: Another Hp49G+ bug? Please see if you get it. <3jqt43Fr47g3U1@individual.net> <_GpCe.1181$PL5.153131@news.xtra.co.nz> <1121604031.a94ff53bd1fcbda236f18bb07bf5764f@teranews> <3jv7jsFr7bs8U1@individual.net> <1121681632.8fa4b3b312c894375433a5fe48a72d9e@teranews> posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt Your question : > 2: { ABC 'Foo' 2 [2 3] } > 1: Bar > What would be the most consistent way to combine the two objects? My answer: A function called APPEND (ADD, whatever) will take Bar and add it to the list. A function called + will return an error because Bar cannot be added to an integer (to name one error among several). This is consistent with everyone elses definition of what + means. This is consistent with the fact that adding strings or structured datasets is not to be done with a simple mathematical operator, unless you want to look at error codes. What else makes sense? The present overloading of + amounts to a bear trap for users, where a very common operator takes a different meaning within a context where it would be expect to perform its usual function, or return an error as in your examples. Its all semantics in the end I suppose. If the functions of + and ADD were simply swapped around then it would all work fine. No bear trap with + and still the ability to append to lists with ADD. Did you see what I posted about the difference between X+1 and X--1? Look at that and tell me you think the present scheme is a sane one. Adrian === Subject: Re: Another Hp49G+ bug? Please see if you get it. > Your question : >> 2: { ABC 'Foo' 2 [2 3] } >> 1: Bar >> What would be the most consistent way to combine the two objects? > My answer: > A function called APPEND (ADD, whatever) will take Bar and add it > to the list. > A function called + will return an error because Bar cannot be added to > an integer (to name one error among several). > This is consistent with everyone elses definition of what + means. > This is consistent with the fact that adding strings or structured > datasets > is not to be done with a simple mathematical operator, unless you > want to > look at error codes. > What else makes sense? > The present overloading of + amounts to a bear trap for users, where a > very common operator takes a different meaning within a context where > it would be expect to perform its usual function, or return an error as > in your > examples. > Its all semantics in the end I suppose. If the functions of + and ADD > were simply swapped around then it would all work fine. No bear > trap with + and still the ability to append to lists with ADD. > Did you see what I posted about the difference between X+1 and X--1? > Look at that and tell me you think the present scheme is a sane one. > Adrian The problem was that the backward consistency and the old behavior with the add to the list were in the way when list processing was ADDed. I would have used & from the beginning for the concatenation of lists. BUT how about a vote from this small group? => a separate thread VPN === Subject: Re: Another Hp49G+ bug? Please see if you get it. <3jqt43Fr47g3U1@individual.net> <_GpCe.1181$PL5.153131@news.xtra.co.nz> <1121604031.a94ff53bd1fcbda236f18bb07bf5764f@teranews> <3jv7jsFr7bs8U1@individual.net> <1121681632.8fa4b3b312c894375433a5fe48a72d9e@teranews> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > What would be the most consistent way to combine the two objects? How does one do it in Java? They oviously thought that it made sense to concatanate strings. . . This discussion has popped up almost once every year or two for as long as I've been using HPs. I was originally thinking the same way you are. However, I really DO want to add things to a list much more than I want to add a number to them. I would rather be able to add a number to a list quickly instead of searching for it in a menu. Here's the great thing. . . you can change it by assigning a program if that is what you want. Remember the old blue friend that came before goldenrod? The ' key was shifted. . . but that didn't make the calculator useless. TW === Subject: Re: Another Hp49G+ bug? Please see if you get it. <3jqt43Fr47g3U1@individual.net> <_GpCe.1181$PL5.153131@news.xtra.co.nz> <1121604031.a94ff53bd1fcbda236f18bb07bf5764f@teranews> <3jv7jsFr7bs8U1@individual.net> <1121681632.8fa4b3b312c894375433a5fe48a72d9e@teranews> posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt Agreed. === Subject: Re: Another Hp49G+ bug? Please see if you get it. <3jqt43Fr47g3U1@individual.net> <_GpCe.1181$PL5.153131@news.xtra.co.nz> <1121604031.a94ff53bd1fcbda236f18bb07bf5764f@teranews> <3jv7jsFr7bs8U1@individual.net> posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt Hi again, well I am back, and the fever has gone. First of all - my initial post. I must apologise for the structure of it. I started asking if it is a bug, then assumed it was. That was not what I should have done, but running a temperature of 103F didn't help my clarity of thought. Clearly the calculator is acting as designed, but not as expected I would suggest. Certainly not as I expected. The + operator is one of the most fundamental operators in maths. To re-define its function in this manner is not the right thing to have done, it seems to me. I understand the need to have an operator to add an element to a list, but not + surely?? To then need ADD to do what + should do, it just seems odd. To allow + and ADD with these functions to exist in a context where they will both be freely used can only lead to this type of confusion. What I would suggest is that + does summation and a command like PLUS or ADD or APPEND or whatever... is used to add an element to a list. Adding a list element is a non standard maths function so I would expect it to have a command that I need to go looking for. I found myself wondering what -, * and / will do in a list context? If + adds items then, - presumably removes the last item from the list /n (repeated subtraction) removes the last n items *n (repeated addition), adds n copies of something? ^n (repeated multiplication), I am really struggling here ... Obviously I am being facetious here, but I hope I am also making a valid point. If we mess about with the basic operators then what we have left is an inability to interchange information between each other anymore. Anyway now I understand the use of ADD. So I can continue. Adrian === Subject: Re: Another Hp49G+ bug? Please see if you get it. X > Anyway now I understand the use of ADD. So I can continue. Did you try: X--1 VP === Subject: Re: Another Hp49G+ bug? Please see if you get it. <3jqt43Fr47g3U1@individual.net> <_GpCe.1181$PL5.153131@news.xtra.co.nz> <1121604031.a94ff53bd1fcbda236f18bb07bf5764f@teranews> <3jv7jsFr7bs8U1@individual.net> posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt now I see what you mean, sorry. When I press the +/- button X-1 turns to X+1 on the display, so it will do what I first reported. If I add some brakets then I get X-(-1) which then looks like X--1 when on the stack. Define that then run, {1 2 3} becomes {2 3 4} So now we see that X+1 and X-(-1) do different things. That again seems counter intuitive to me. n.b. while X--1 is on the stack, before I do the DEFINE, If I press EVAL then we get X+1. This of course has now changed the functionality to the append to list function This is getting too confusing .... :-) My head hurts. Adrian === Subject: Re: Another Hp49G+ bug? Please see if you get it. <3jqt43Fr47g3U1@individual.net> <_GpCe.1181$PL5.153131@news.xtra.co.nz> <1121604031.a94ff53bd1fcbda236f18bb07bf5764f@teranews> <3jv7jsFr7bs8U1@individual.net> posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt can't do that, I get a syntax error. J(X)=X-1 acts as I would have expected i.e. {1 2 3} becomes {0 1 2} Adrian === Subject: Re: Another Hp49G+ bug? Please see if you get it. > How would you add an object to a list? > There's no design fault here, that is how the calculator has been > behaving since it was first introduced in 1989. Forgot to add that the way the list processing vs adding element to a list works has been working the exact same way since the 48GX was introduced over 11 years ago. JY === Subject: Re: Another Hp49G+ bug? Please see if you get it. >> If you rely on a $130 calculator machine to make business decision worth >> $50,000 then I would question your working practice really... >> Jean-Yves >WTF kind of reply is that?! NASA used the HP-41 on missions to the >moon, for Pete's sake! Jean-Yves, I'm sick of your arrogant, less than >helpful attitude! Get a grip! Apollo 17 moon landing 11 Dec 1972. HP-41C introduction 1 July 1979. === Subject: Re: Another Hp49G+ bug? Please see if you get it. http://www.nasm.si.edu/research/dsh/artifacts/GC-hewlett.htm === Subject: Re: Another Hp49G+ bug? Please see if you get it. > If you rely on a $130 calculator machine to make business decision worth > $50,000 then I would question your working practice really... > Jean-Yves >>WTF kind of reply is that?! NASA used the HP-41 on missions to the >>moon, for Pete's sake! Jean-Yves, I'm sick of your arrogant, less than >>helpful attitude! Get a grip! > Apollo 17 moon landing 11 Dec 1972. > HP-41C introduction 1 July 1979. Yeah, but you're forgetting about the time dilation effects due to the spacecrafts' speed. Bob === Subject: Re: Another Hp49G+ bug? Please see if you get it. windows-nt) >> If you rely on a $130 calculator machine to make business decision worth >> $50,000 then I would question your working practice really... >> Jean-Yves > WTF kind of reply is that?! NASA used the HP-41 on missions to the > moon, for Pete's sake! Jean-Yves, I'm sick of your arrogant, less than > helpful attitude! Get a grip! It was a serious reply. Using a $130 calculator for which you haven't even read the instructions completely to make decisions of that magnitude is irresponsible. Get a grip. Best, -Al === Subject: Re: Another Hp49G+ bug? Please see if you get it. > It was a serious reply. Using a $130 calculator for which you haven't > even read the instructions completely to make decisions of that > magnitude is irresponsible. Get a grip. I read all the instructions for everything I use, but I consider this to be a form of OCD. Devices should generally operate as expected, and not require a steep learning curve to avoid gotchas. This particular point of backwards compatibility was a very poor design decision. === Subject: Re: Another Hp49G+ bug? Please see if you get it. > WTF kind of reply is that?! NASA used the HP-41 on missions to the > moon, for Pete's sake! Jean-Yves, I'm sick of your arrogant, less than > helpful attitude! Get a grip! You're right.. let's focus on the *bug* reported by the original poster then... JY === Subject: Re: Another Hp49G+ bug? Please see if you get it. > The applied part > ---------------- > I just came across what looks like a bug on my HP49G+ > (CNA41, ROM2.0, RPN mode). Can anyone else see it? > Create a function using define, something simple like H(X)=X+1 > and then H appears as a soft key. > Build a list of numbers {1 2 3} for example, pop it on the stack > and then press the H key > what I think I should get is {2 3 4} > what I get is {2 3 4 1} > I tried it with several combinations of function and list, and it seems > that the list is having a 0 appended to the end which the function is > then processing. This is not a bug. It's just more confusing behavour included for historical reasons. + does not add a number to each element, it instead adds a new element to the list. There is another function for adding to each element. Al === Subject: how to store a variable's name posting-account=O6knhQ0AAAD8zE21TyMVwtCasYnRdHOm How do i store a variable's name, instead of its value, in User RPL? for example, suppose global variables var and Y dont exist yet, then i do: 'Y' 'var' STO if i put var on the stack, its all good, it shows up 'Y' but then i do: 0 'Y' STO now when i put var on the stack , i get 0. i would like it to still return 'Y', even if 'Y' has a value . === Subject: Re: how to store a variable's name posting-account=O6knhQ0AAAD8zE21TyMVwtCasYnRdHOm I still dont understand why my method doesnt work. the variable var contains 'Y' as its value, so when i put var on the stack, the evaluation of it should return its value, that is, 'Y', but it returns the value of its value, that is 0. obviusly my reasoning is wrong coz it doesnt work that way, but i dont understand why === Subject: Re: how to store a variable's name > I still don't understand... > the variable 'var' contains 'Y' as its value, > so when i put var on the stack, > the evaluation of it should return its value, that is, 'Y', The rule is that evaluating the name of a variable (if it exists in the current or a higher directory) evaluates what that variable contains. But if the *content* of the first variable is again a variable name, this causes *recursive* application of the same rule, which in this case again evaluates the content of the *second* variable, exactly as designed (and a very handy and useful design that is!) Thus, to recall the content of a variable without risking re-evaluation of that content, you must do what JYA said, e.g. 'var' RCL, in place of just the unquoted name [var]. [r->][OFF] === Subject: Re: how to store a variable's name posting-account=O6knhQ0AAAD8zE21TyMVwtCasYnRdHOm yes now i understand, example: Ç 0 'a' STO 'a' 'b' STO 'b' 'c' STO 'c' 'd' STO d 'd' RCL @ 'c' RCL @ 'b' RCL @ 'a' RCL @ 0 IF == THEN oh i got it 3 DISP ELSE i give up 3 DISP END 3 FREEZE { a b c d } PURGE È :O) === Subject: Re: how to store a variable's name posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE As jean-yves said, when storing 'Y' in 'var' and calling 'var', 'var' will evaluate its content, in your case: 'Y'. if 'Y' does not exist in HOME or in the directory you're evaluating 'var', it will return 'Y' as a name. That's why you could see 'Y' on the 1st level of the stack. if 'Y' exists, then 'var' will evaluate its content, as 'var' contents 'Y', it will evaluate 'Y', as you asked for. You made the experience yourself already :) if the content of 'var' is the following: << 'Y' >> In this case, Y is surrounded by single quote the evaluation will return: 'Y' no matter if 'Y' exists. if the conent of 'var' is the following: << Y >> In this case, Y is not surrounded by single quote the evaluation will return: 1) 'Y' if 'Y' is not defined in the directory or parent directory where 'var' is executed 2) will execute the content of 'Y' if 'Y' exists in the directory or parent directory where 'var' is executed. In fact, you're calling a (global) routine named 'Y' and 'Y' can contain what ever you want. I hope our explanation were clear. fred. === Subject: Re: how to store a variable's name posting-account=O6knhQ0AAAD8zE21TyMVwtCasYnRdHOm === Subject: Re: how to store a variable's name > i would like it to still return 'Y', even if 'Y' has a value . That's because you are evaluating var. What you should do is enter var within quote like 'var' so it is not evaluated. Then use RCL 'var' RCL The will recall the content of var rather than evaluate it. Jean-Yves === Subject: Re: how to store a variable's name posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Instead of storing directly 'Y' in 'var', store << 'Y' >> in 'var'. Fred. === Subject: bug in user commancd CLKADJ; 49g+ ROM v2.0 Hello together, there seem to be a bug in CLKADJ, as Scott has allready reported. Make a small exercise: 1. Adjust your clock, by using 3. Set time, date.. from the time menu This adjustment does not need to be that precise, just note the deviation, you have afer adjustment. 2. put 8192 (8192 ticks should be 1 s) into the stack and 3. run the following program << 1 10 START DUP DUP CLKADJ DUP NEG CLKADJ NEXT This program adds and substract 8192 ticks 10 times to/from your system clock. 4. Check your clock. Up to here, everything should be ok; the clock should have the same deviation than before running the above program. 5. but when putting 12288 (1.5 seconds) into the stack and running the above porgram, the system clock will go 10 seconds ahead. 1.6 seconds give the same result -> 10 seconds ahead. I have tried it with 2 s which seem to be ok, but fractions of a second seem to be problem. If you are using a time keeper you should not try to adjust with fractions of a second :-) With Wolfgangs Timeman you should not try to adjust by +/-0.1 s, but +/- 1 s should work. Improper working of CLKADJ on hp 49g+ has been reported more than one year ago (ROM v1.23). It would be nice to get more light into this subject. Heiko === Subject: [HP49G+] RND oddity Hi. 'RND(310.999488065_oC,-6.)' (oC is the degrees celsius unit) EVAL returns '-3.9243E1_oC'. That is obviously on the wrong side of comforting, huh? Steen === Subject: Re: [HP49G+] RND oddity > 'RND(310.999488065_oC,-6.)' (oC is the degrees celsius unit) EVAL > returns '-3.9243E1_oC'. Forget about this - it was my own mistake. Sorry to have wasted your time :-/ Steen === Subject: Re: RND oddity posting-account=Kc05ZA0AAADq7YTEHoMU9Aer5AhqUt9P After entering the above expression and pressing EVAL twice(?), I get 310.999000_oC. === Subject: Re: RND oddity > After entering the above expression and pressing EVAL twice(?), I get > 310.999000_oC. Yes, I cannot reproduce it currently either - I don't know if it was a flag setting or something :-/ I'll post again if I find a way to reproduce it. Steen === Subject: Unexpected behavior in input forms - hp49 posting-account=YX-43w0AAACHvk30VnVcLC64muZRYvn8 I discovered that for any fields in most input forms with the edit softkey option, you can call up some choose box/soft menus: eg, calc, alg, s.slv, etc. This means you can call up the calculus menu while highlighting 1. the independent variable field in CAS MODES or 2. the enable beep check box in CALCULATOR MODES or 3. the independent variable name in PLOT SETUP. However I can't recreate this behavior in any of the original input forms that existed in the hp48 such as in the statistics or set date and time input form. I also discovered you can access the MODES input form while editing in the statistics input form. You can't do this in the hp48. === Subject: Re: Unexpected behavior in input forms - hp49 posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > I also discovered you can access the MODES input form while editing in > the statistics input form. You can't do this in the hp48. Well, being a different calculator one would expect differences. Is there some sort of problem you are experiencing, or are you just saying it is different that what you thought? TW === Subject: Re: Unexpected behavior in input forms - hp49 posting-account=YX-43w0AAACHvk30VnVcLC64muZRYvn8 Actually, you can call up the mode menu in any editable fields by first going back to the stack using CALC. For graphing I guess there isn't really any modes to change that would affect graphing except the angle measure. In setting time and alarm, you'd need to be able to toggle the beep. I guess the mind set in using a computing machine is you want to open up an application and be able to access all the needed functions to accomplish a task without backtracking, then turn the machine off. The hp49 software is so feature rich and carefully thought out that it takes some thought to get why the calculator behaves contrary to my own instincts. I'm glad that all the input forms open up so fast. === Subject: Re: Unexpected behavior in input forms - hp49 posting-account=YX-43w0AAACHvk30VnVcLC64muZRYvn8 Well, it's weird that I could change modes in while editing some input form, but not in any of the graphing related forms like graph window or equation entry... it would be convenient to have that shortcut there. Weird also is you can access modes menu in setting time and date or setting alarms, which is useless. === Subject: Does HP-48SX/GX SOLVER have the S function for algebraic formulas? posting-account=XZBzkw0AAACvm4tUXfyLZUkjp5CZRheR S function when using algebraic formulas in the SOLVER. 1. Is this true? 2. Is it supported on the HP 48GX? leb === Subject: Re: Does HP-48SX/GX SOLVER have the S function for algebraic formulas? > S function when using algebraic formulas in the SOLVER. > 1. Is this true? > 2. Is it supported on the HP 48GX? > leb === Subject: Re: Does HP-48SX/GX SOLVER have the S function for algebraic formulas? do you mean the integral symbol, which looks similar to an S? Raymond schrieb im Newsbeitrag > S function when using algebraic formulas in the SOLVER. > 1. Is this true? > 2. Is it supported on the HP 48GX? > leb === Subject: Re: Does HP-48SX/GX SOLVER have the S function for algebraic formulas? posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE What do you mean by S function? Can you show an example please? Fred. === Subject: Re: Does HP-48SX/GX SOLVER have the S function for algebraic formulas? posting-account=XZBzkw0AAACvm4tUXfyLZUkjp5CZRheR Sorry for the delay. The S function allows one to test for the variable being solved for. It's almost always used with the IF function. I know that at least the HP-17BII, the HP-19B and HP-27S, and the HP-200LX have the S function. A partial example from the Technical Applications manual for the HP-19B and HP-27S is below: VECTOR: IF(S(CROSS):CROSS: IF(S(DOT):DOT: IF(S(ANG):ANG:MAG)))= 0XL(M:SQRT(SQ(X1)+ SQ(Y1)+SQ(Z1))) +IF(S(DOT) OR S(ANG): 0XL(C:X1xX2+Y1xY2+Z1xZ2) +IF(S(ANG): ACOS(G(C)/(G(M)x SQRT(SQ(X2)+SQ(Y2)+SQ+(Z2)))):G(C)) +IF(S(MAG):G(M): 0xL(A:Z1xX2-X1xZ2)x L(B:X1xY2-Y1xX2)+ L(X1:Y1xZ2-Z1xY2)+ 0xL(Y1:G(A))+ 0xL(Z1:G(B)))) === Subject: Re: Does HP-48SX/GX SOLVER have the S function for algebraic formulas? posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J You mean conditional in equations? maybe using the hp internal notation for functions (graphs)... << -> vars << 'function_in_alg' >> Try use it in graphs, this work for conditioned graphs, but i dont try this solving equations === Subject: HP49g+ USB on Linux occasionally we see search engine queries regarding USB on Linux in the hpgcc.org server logs. I would like to remind you, that Daniel Herring has compiled a Linux-USB-Mini-HOWTO, which is online available at: http://hpgcc.org/hpgcc/contrib/Linux-USB-Mini-HOWTO.txt best, ibl === Subject: Re: HP49g+ USB on Linux > Daniel Herring has compiled a > Linux-USB-Mini-HOWTO The latest 2.6 kernels feature a specific hp4x driver, instead of modprobing usbserial with vendor and product options, you can compile the hp4x module and modprobe hp4x directly. The driver does absolutely nothing different from calling usbserial with the right options. -- A.H. === Subject: Debug4x New Release Available You may obtain a new Debug4x Release from: http://debug4x.home.netcom.com/ Always read the VersionInfo.txt in the Debug4x directory after installation. Abbreviated Summary of Changes ---------------------------------------------------------------------------- --------------------------- -DIRECTORIES The Direct example has been changed. A macro to assist in building directories has been added to the source. This new structure can be debugged with the normal breakpoint features of Debug4x. Please stop using the older Project is a Directory feature. It has too many bugs and cannot use breakpoints! -WARNING Project Data tab option The project is a directory will be removed in the next release. -EMU EMU has been updated. The 49G+ version no longer crashes on Win98 and (sometimes) WinXP. Still does not support direct ARM code but does emulate the full 49G+ except Port 2 and 3. -Inform Editor Can now use the bigger calculator screen (49G+). The screen area now has a gray background to assist Windows OS styles with white backgrounds in their windows. -Source Edit There is a new File Menu option to Open Calculator Entry Point file. This will display/edit the ROM definitions file. If this file is saved the matching object file will be deleted and then recompiled during the next build. You must do a build before debugging if the file has been deleted. -- - - - - - - - - - - - - - - - - Bill Graves RKBA! bgraves@ix.netcom.com === Subject: Battery Type or Spec posting-account=U2_nsg0AAACMehSjdUxQpFmS8kMQCYl9 (Mainly for 49g+) To those who are using rechargeable batteries, does it really matter if the current is less than the regular batteries? Does this affect the calculator in any way at all? For example, the regular non-rechargeable batteries (AAA) is rated 1.5V, but a close inspection of the rechargeable batteries shows 1.2V. Is this going to matter at all? Also does it matter what kind of batteries used (Duracell, Energizer,...)? === Subject: Re: Battery Type or Spec > (Mainly for 49g+) To those who are using rechargeable batteries, does > it really matter if the current is less than the regular batteries? You'll see the battery warn symbol earlier when you use rechargeable batteries, especially if the calculator is calculating for a longer time, let's say a minute or so. But it still works as it should. ;-) So you can use rechargeable batteries, but you'll have to recharge them more often than you replace normal batteries. My HP 49G+ is doing fine with three AAA with 700 mAh each, I have to recharge them about every month or so. wupperbayer === Subject: Re: Battery Type or Spec One could try to expand the battery case to hold four AAA's :NiMH:'4*1.2_V' => 4.8_V ~ 4.5_V (:alkalines:'3*1.5_V') The over/undervoltage at these limits doesn't matter much... V-P >> (Mainly for 49g+) To those who are using rechargeable batteries, does >> it really matter if the current is less than the regular batteries? > You'll see the battery warn symbol earlier when you use rechargeable > batteries, especially if the calculator is calculating for a longer time, > let's say a minute or so. But it still works as it should. ;-) > So you can use rechargeable batteries, but you'll have to recharge them > more often than you replace normal batteries. My HP 49G+ is doing fine > with three AAA with 700 mAh each, I have to recharge them about every > month or so. > wupperbayer === Subject: Re: Battery Type or Spec <70GCe.5666$6J1.1896@reader1.news.jippii.net> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 I did succesfully convert a 49g+ to 3 AA batteries. . . was quite a job, but it worked fine. Unfortunately the calculator had a rather unfortunate encounter a few weeks later with the pavement as I rode my bike down a rather steep hill. . . =( However, I can say that the 49g+ has a voltage regulator. Our custom external circuit board coincidentally had the same exact one installed on it. I think it can handle up to 7.5~ volts witout a problem. One less part for us to pay for! =) TW === Subject: Re: Battery Type or Spec Photos? Instructions? VN >I did succesfully convert a 49g+ to 3 AA batteries. . . was quite a > job, but it worked fine. Unfortunately the calculator had a rather > unfortunate encounter a few weeks later with the pavement as I rode my > bike down a rather steep hill. . . =( > However, I can say that the 49g+ has a voltage regulator. Our custom > external circuit board coincidentally had the same exact one installed > on it. I think it can handle up to 7.5~ volts witout a problem. One > less part for us to pay for! =) > TW === Subject: Re: Battery Type or Spec <70GCe.5666$6J1.1896@reader1.news.jippii.net> <9uGCe.5688$mN1.2519@reader1.news.jippii.net> posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 Photos were lost in an SD corruption. Basically take the calculator apart. The cut out the ENTIRE battery compartment, build a new one with some foam, aluminum foil and glue. Use a dremel to make the battery cover about 1/3 th thickness it is currently. Then it will fit. Your batteries will be completely up against a peice of material covering the back of the cicuit board and it still caused the battery cover to bulge slightly. Much more trouble than it is worth. . . :-/ TW === Subject: Re: Battery Type or Spec > Photos were lost in an SD corruption. Basically take the calculator > apart. The cut out the ENTIRE battery compartment, build a new one > with some foam, aluminum foil and glue. Use a dremel to make the > battery cover about 1/3 th thickness it is currently. Then it will > fit. Your batteries will be completely up against a peice of material > covering the back of the cicuit board and it still caused the battery > cover to bulge slightly. Much more trouble than it is worth. . . :-/ > TW BUT does it then weight more firm in your hand? VN === Subject: Re: Battery Type or Spec <70GCe.5666$6J1.1896@reader1.news.jippii.net> <9uGCe.5688$mN1.2519@reader1.news.jippii.net> posting-account=hjxs3Q0AAABam4kWUQsVc-uTekTFjQEs Seems HP has delagated the choice of batteries for their calculators to some rather un-creative people (to put it mildly). The HP-41 is a prime example of that with their size N batteries. (which were expensive, hard to find and had low capacity) Things have not improved much with the HP48-49's. I'm sure it would have been possible to modify the innards of the HP49 and HP49g+ to accept 3 AA instead of 3 AAA without any significant (if any) change in the external dimensions. AA's are usually cheaper than AAA's and they have twice the capacity. How sad! === Subject: Re: Battery Type or Spec > Seems HP has delagated the choice of batteries for their calculators to > some rather un-creative people (to put it mildly). The HP-41 is a > prime example of that with their size N batteries. (which were > expensive, hard to find and had low capacity) > Things have not improved much with the HP48-49's. I'm sure it would > have been possible to modify the innards of the HP49 and HP49g+ to > accept 3 AA instead of 3 AAA without any significant (if any) change in > the external dimensions. AA's are usually cheaper than AAA's and they > have twice the capacity. > How sad! Sad it is! I have proposed before the use of AA's in the hp 49g+ Guess what happened...no go - you go !! VPN )-`: === Subject: 1972 Garrett 2000 RPN Calculator I am selling a 1972 Garrett 2000 RPN Calculator on ebay,ending today. This is the first non-Hewlett-Packard RPN calculator ever made. http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=5788184270 === Subject: Re: 1972 Garrett 2000 RPN Calculator posting-account=boybFQ0AAACcJkbp4Mpsq2-jhOTZrn5e we are all HP fans here, RPN is not enough to sell a brick no matter how old it is :) === Subject: Re: 1972 Garrett 2000 RPN Calculator posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J the brick looks funny! hahaha === Subject: Where i can buy the last CNA5 for hp 49g+? === Subject: Re: Where i can buy the last CNA5 for hp 49g+? posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt These folks have CNA5 and will ship overseas. They are about the cheapest too. Be sure to ask for a CNA5, they still have CNA4 in stock. http://www.samsoncables.com/catalog/index.cfm If you can wait 1 week - I have a CNA5 on order. I'll get it a week today. I can tell you what differences I can see. Adrian === Subject: Re: Where i can buy the last CNA5 for hp 49g+? > If you can wait 1 week - I have a CNA5 on order. I'll get it a > week today. I can tell you what differences I can see. === Subject: Re: Where i can buy the last CNA5 for hp 49g+? posting-account=hjxs3Q0AAABam4kWUQsVc-uTekTFjQEs A friend of mine just got some from hpmarketplace.com on ebay. I don't know if they ship overseas. Luis === Subject: 49G+ - Lost font darkness in display hi I was using my calculator and it was fine. used it again after about a 2 minute break, and the contents in the screen are very light/faint. The batteries are fine, and i have swaped with a new set to make sure - the problem is still there. This is the first time this has happened in almost 2 yrs since new. Has this happened to the 49G+ before? It is running rom 1.23 (2 isn't compatible with winHP) === Subject: Re: 49G+ - Lost font darkness in display Such is life - after a short while it has come good touch wood === Subject: Exchange list + and ADD Vote VPN: I vote for NO === Subject: Re: Exchange list + and ADD Vote > VPN: I vote for NO No, no, no, a thousand times no! Scott Chapin === Subject: Re: Exchange list + and ADD Vote >> VPN: I vote for NO > No, no, no, a thousand times no! Sorry, Scott, you have only one vote (-; VPN === Subject: Re: Exchange list + and ADD Vote >VPN: I vote for NO I vote NO as well. Bob === Subject: Re: Exchange list + and ADD Vote I vote NO too reth VPN: I vote for NO > I vote NO as well. > Bob === Subject: Re: Exchange list + and ADD Vote posting-account=hYpJhAwAAABbgx9xHixEB8na9TYTFvkV > VPN: I vote for NO I vote for will never be implemented :D Bhuvanesh. === Subject: Re: Exchange list + and ADD Vote YES People who want backwards-compatibility should use an electromechanical punched card reader. === Subject: Re: Exchange list + and ADD Vote > YES > People who want backwards-compatibility should use an electromechanical > punched card reader. I use my fingers (teacher told me to put my hands into my pockets) ...oops - I got eleven again (adding 4 and 6) VPN === Subject: Re: Exchange list + and ADD Vote posting-account=cLlfiw0AAAChMJg5QaiY1LBfFgQqnnrf I vote for the buttons doing whatever you think they're going to do, without regard to if you have read the documentation or not. It seems that would fix most of the bugs posted to this forum, as they seem to be people saying I intuited that this should do something or other, but it does something else, so it's wrong. Where do I vote for that? === Subject: Re: Exchange list + and ADD Vote posting-account=FCAFHAwAAADaPCjuTWXm3ER18SIzldtO If for you it's just a matter of buttons, you can exchange the behavior of the '+' key by assigning the following SysRPL program: :: DUPTYPELIST? 3PICK TYPELIST? ORcase xADD x+ ; @ But if you want to write UserRPL programs you need to use it the way it is. Claudio === Subject: Re: Exchange list + and ADD Vote posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt You vote YES, here. === Subject: Re: Exchange list + and ADD Vote >I vote for the buttons doing whatever you think they're going to do, > without regard to if you have read the documentation or not. > It seems that would fix most of the bugs posted to this forum, as > they seem to be people saying I intuited that this should do something > or other, but it does something else, so it's wrong. > Where do I vote for that? Create a new thread How about using the assign featured or a custom menu? VPN === Subject: Re: Exchange list + and ADD Vote posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Why exchange the name? Who has had this STUPID idea? I vote NO too. Reason: keep backward compatibility. '+' has always added an element to a list for the dawn of HP calculators... I always have seen '+' acting this way. There is no reason for changing this. Fred. === Subject: Re: Exchange list + and ADD Vote posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt > Who has had this STUPID idea? I did :-) There is a post recently, which I won't repeat here if you want the details, === Subject: Re: Exchange list + and ADD Vote posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J Veli-Pekka Nousiainen ha escrito: > VPN: I vote for NO Any explanation to this? I can't understand === Subject: Re: Exchange list + and ADD Vote Avatar_e schrieb im Newsbeitrag > Veli-Pekka Nousiainen ha escrito: VPN: I vote for NO > Any explanation to this? I can't understand Because it is documented, at least for the HP 48G-Series. Documentation is really an argument :-) Heiko === Subject: Re: Exchange list + and ADD Vote We could see if people *REALLY* want to exchange the name of these commands...not that HPQ would actually do anything,. but just to take a vote to see the results in this small group I just that your vote would be: Avatar_e: YES :-D VPN > Veli-Pekka Nousiainen ha escrito: >> VPN: I vote for NO > Any explanation to this? I can't understand === Subject: Serial connection for 49g+ under rom 2.0? posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J Why in IO options appear the following options: USB Serial IrDA ??????? when i select SERIAL, CONN4x does not link with my hp using the provided usb cable. Can 49g+ manage usb port like a serial? :O or is just another bug? maybe 49g+ can connect to COM using SERIAL option in IO and serial cable provided with 48gII??? === Subject: Re: Serial connection for 49g+ under rom 2.0? 48gII > Why in IO options appear the following options: > USB > Serial > IrDA > ??????? > when i select SERIAL, CONN4x does not link with my hp using the > provided usb cable. Can 49g+ manage usb port like a serial? :O or is > just another bug? > maybe 49g+ can connect to COM using SERIAL option in IO and serial > cable provided with 48gII??? === Subject: Re: Serial connection for 49g+ under rom 2.0? posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J Could the 49g+ connect to some like hyperterminal, or any kermit client like olds hp?? (and hp48gII too) === Subject: Re: Serial connection for 49g+ under rom 2.0? posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > Could the 49g+ connect to some like hyperterminal, or any kermit client > like olds hp?? (and hp48gII too) Unless you can think of some way of making USB signals turn into RS-232 with a regular wire, no. TW === Subject: Re: Serial connection for 49g+ under rom 2.0? posting-account=hjxs3Q0AAABam4kWUQsVc-uTekTFjQEs How about some Windows software that is the equivalent of Kermit but that works through the USB port and connects using Xmodem. I mean some software that could allow control of the HP49g+ from the PC with commands equivalent to REMOTE HOST and so forth. Luis Could the 49g+ connect to some like hyperterminal, or any kermit client like olds hp?? (and hp48gII too) > Unless you can think of some way of making USB signals turn into RS-232 > with a regular wire, no. > TW === Subject: Re: Serial connection for 49g+ under rom 2.0? > How about some Windows software that is the equivalent of Kermit but > that works through the USB port and connects using Xmodem. > I mean some software that could allow control of the HP49g+ from the PC > with commands equivalent to REMOTE HOST and so forth. That works today (on linux). Basically, the 49g+ uses RS232 over USB at an arbitrary baud rate determined by the USB connection; with the appropriate software, you can use kermit or xmodem to talk to your calculator. For linux machines, I've documented the basics in the Mini-HOWTO that was mentioned recently. For MS Windows machines, HP's software does the same basic thing, but I don't know how easily you can send commands with it. Its possible that the modifications wouldn't be too difficult. Just to clarify: the RS232 which goes over the USB cable is embedded in USB packets; it conforms to the USB specs and is not a true RS232 signal. Later, Daniel weekend and should update things shortly. === Subject: Re: Serial connection for 49g+ under rom 2.0? posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J But, usb does not work as a client-server comunication? I want to use ms-kermit.exe for example like with hp48gii direcly talking send CMD >> receive STRING. :( there is no way to use the hp49gx.sys (kinpo hp49gp usb driver) in vb or c#? === Subject: hp 49g+; ->TIME - bug, CLKADJ - bug Hello together, the user-command ->TIME has a bug, it adjust only with an acuracy of +/- 1 s. The command in the time menu, acessible with [RS] [9] Set time, date.. seem to make a proper job! So, if sombody tells you, it is not possible to adjust the clock of the hp 49g+ better then +/ 1 s it is a fairy taile. hp 49g+ ROM 1.23 and 2.0 are effected. Heiko === Subject: file,lib tranfers from PC to hp 48GX posting-account=UZpa1g0AAAAGK-b1uuwdq2SDcb2fyNNB I'm new to transfering lib and other program from PC to hp 48GX. I have memory card and every thing . The connection is fine . All that is === Subject: Re: file,lib tranfers from PC to hp 48GX posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE To connect your HP-48GX to the PC, be sure that: - you have selected the correct COM port and plugged the calc on it - PC and calc use the same transfer protocol (KERMIT or XMODEM) - the parameters of the protocol are the same on both parts (PC & calc) - have a preference for binary transfer for both protocol (flag -35 set) - make sure that you're using wire transfer (flag -33 cleared) Use SEND or RECV for KERMIT transfers Use XSEND or XRECV for XMODEM transfers files from HP calcs to PC. Take a look: http://www.hpcalc.org/details.php?id=6238. HPConn uses XModem protocol. Fred. === Subject: Re: file,lib tranfers from PC to hp 48GX posting-account=UZpa1g0AAAAGK-b1uuwdq2SDcb2fyNNB It is working how files are been transfered. I don know which dir has the file or programs . help please > To connect your HP-48GX to the PC, be sure that: > - you have selected the correct COM port and plugged the calc on it > - PC and calc use the same transfer protocol (KERMIT or XMODEM) > - the parameters of the protocol are the same on both parts (PC & calc) > - have a preference for binary transfer for both protocol (flag -35 > set) > - make sure that you're using wire transfer (flag -33 cleared) > Use SEND or RECV for KERMIT transfers > Use XSEND or XRECV for XMODEM transfers > files from HP calcs to PC. Take a look: > http://www.hpcalc.org/details.php?id=6238. HPConn uses XModem protocol. > Fred. === Subject: Re: file,lib tranfers from PC to hp 48GX posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE What do you mean? I don't understand your question. On the calculator, you can press the [VAR] key to display the files in the menu. To know where you are, you have the instruction PATH. It will push the current path on the 1st level of the stack. The path is also displayed as a list in the status area (top left of the screen). To know how files are being transferred, well, you have to understand XModem and/or Kermit protocols. For XModem (easiest to implement): http://www.commonsoftinc.com/Babylon_Cpp/Documentation/Res/yModem.htm For Kermit (not obvious to implement): http://www.columbia.edu/kermit If you have Windows installed on your PC, you can also use the HyperTerminal for transferring files from HP-48 to PC using the serial communications. Fred. === Subject: Re: file,lib tranfers from PC to hp 48GX > If you have Windows installed on your PC, you can also use the > HyperTerminal for transferring files from HP-48 to PC using the serial > communications. And last (but not least), you can also use EMU48 on your Windows PC, emulating the same or any other compatible calculator on the computer itself; when you act as if to transfer from one calculator to another, EMU48 is able to use the computer's own serial port as the serial port of the emulated calculator. To transfer between the emulated calculator (stack level 1) and files on its computer host, use Emu48's Load/Store. To perform ascii translations of special non-ascii characters (like program delimiters, etc.) in UserRPL programs, the same as occurs during ascii transfer via Kermit, within the emulated calculator itself: [r->][OFF] === Subject: Hewlett-Packard to Lay Off 14,500 to Save $1.9 Billion As reported by The New York Times: Hewlett-Packard to Lay Off 14,500 to Save $1.9 Billion SAN FRANCISCO, July 19 - Hewlett-Packard said today it would lay off 14,500 employees, almost 10 percent of its staff, in the next year and a half and freeze its pension plan for new and existing workers who have not vested in it yet. Hewlett will dissolve its Customer Solutions Group, a business unit that sold to businesses and public entities, and distribute its work to the company's three product groups. The company said it would protect employees in sales and research and development from the brunt of the layoffs, which will mostly affect support positions in information technology, human resources and finance. http://www.nytimes.com/2005/07/19/technology/19cnd-hp.html?ex=1279425600&en= 719bd3b66587113b&ei=5088 http://www.forbes.com/technology/ebusiness/feeds/ap/2005/07/19/ap2145637.htm l http://www.businessweek.com/ap/financialnews/D8BEFLN00.htm [r->][OFF] === Subject: Hewlett-Packard fails to utilize 14,500 people, Cuts budget by $1.9 Billion > As reported by The New York Times: > Hewlett-Packard to Lay Off 14,500 to Save $1.9 Billion Hewlett-Packard fails to utilize 14,500 people, and cuts budget by $1.9 Billion. (Top executives set longer Golf reservations?) Also : Apple withdraws I-pod rebranding arrangment. Loyal hp Users await inevatable battery failure. Looking closly, I note my Pavillion laptop is actually an Acer laptop. Looking closly, I note that my hp printer is actually a re-shelled Epson. Looking closly, I note that 3 of my HP Calculators are Kim-Po units. Old News: Compaq servers are shipping as HP products. Looking closly? Buy the brand. L.A. === Subject: Re: Hewlett-Packard to Lay Off 14,500 to Save $1.9 Billion posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt That's not good news for the HP staff. I lived through a similar period with Motorola. On the other hand - as long as the HP49G+ keyboard designer gets it in the pants .... Adrian === Subject: Re: Hewlett-Packard to Lay Off 14,500 to Save $1.9 Billion > That's not good news for the HP staff. I lived through a similar period > with Motorola. > On the other hand - as long as the HP49G+ keyboard designer gets it > in the pants .... You mean Chinese Kinposh, not HPQ? VPN === Subject: HHC2005 posting-account=Go2g5wsAAAAsfBzIGOikqm0lzTSUDSrr I don't think the attendees will be disappointed. Jake Schwartz === Subject: Re: HHC2005 I hope you will all give HP a very very hard time over the 49G+ keyboard. >I don't think the attendees will be disappointed. > Jake Schwartz === Subject: Re: HHC2005 posting-account=BhOLwQ0AAAAGxLzmZMrNIhvjYvNMjBnc We did last year. I for one am expecting other announcements of a more exciting nature. Well I guess not expecting. . . just hoping. New 51xxx keyboards are equivalent to the 48 in feel (comparing them side by side). The only difference in feel was key travel. The new keyboard has a longer travel. As for durability. . . only time will tell. TW === Subject: Re: HHC2005 posting-account=lAaz6QwAAABR-WEJBEnTuXipZ8zCV0L0 I have a CN51001296. No misses, no doubles, very comfortable keys. However, the shift keys are a bit wobbly because they are on their own, the paint job is not as sharp as on my CN3 and the key labels look a bit fuzzy. And the screen is not as clear as on the CN3 (but just a bit, I guess manufacturing tolerances). If I had the look of the CN3 and the behaviour of the CN5 it would be perfect. It is however very good enough and I wouldn't even think about using my 48 anymore. Arnaud === Subject: Re: HHC2005 posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 I am going. Taking a survey robot and will be doing demonstations as well as handing out overlays to attendees. TW === Subject: Re: HHC2005 >I am going. Taking a survey robot and will be doing demonstations as > well as handing out overlays to attendees. A bit far away for me but I hope to see the reports... Arnaud === Subject: HP49G+ and Old Serial Number?? I just received today a new HP49G+ from an on-line vendor. I purchased it on the basis that most of the complaints posted about keyboards on the G+ looked on the outside of the blister for a serial number sticker and there was none. I took that for a good sign thinking perhaps they no longer did that on newer units and opened the package. The serial number on the back of the case was CN402XXXXX. Based on this I have a few questions. 1. It appears, based on other postings, that I have a new unit that was that still correct or have they changed date codes? 2. Should there have been a sticker with the serial number on the outside of the blister pack and is it possible the vendor removed it on purpose to avoid disclosure until the seal was broken. I know I am cynical, but I am also not very happy person. 3. Have they made substantial physical changes since January of 04 (if that is indeed the manufacturing date) to the 49G+ to make it worth while complaining or asking for an exchange? I have not put in the batteries yet or unpacked any thing else, but the keys are definitely stiffer and they don't feel as nice as on my HP41CX by a country mile. I suspect the ROM will be out of date but that should be correctable. I am more concerned about improvements made in the last 18 months that cannot be corrected except with a newer unit. === Subject: Re: HP49G+ and Old Serial Number?? One more question. The vendor is now indicating to me that he has no knowledge of the date code being embedded in the serial number and that he shipped me a unit that is less than a month old. Is there any reference, other than newsgroup postings, I can point to support the first three digits === Subject: Re: HP49G+ and Old Serial Number?? > One more question. The vendor is now indicating to me that he has no > knowledge of the date code being embedded in the serial number and that he > shipped me a unit that is less than a month old. Is there any reference, > other than newsgroup postings, I can point to support the first three digits Have a read of the final paragraph of ... http://www.hpcalc.org/hp48/docs/faq/48faq-2.html It might help === Subject: Re: HP49G+ and Old Serial Number?? posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 > One more question. The vendor is now indicating to me that he has no > knowledge of the date code being embedded in the serial number and that he > shipped me a unit that is less than a month old. Is there any reference, > other than newsgroup postings, I can point to support the first three digits > after the country code are year and week of manufacture? Well, considering customer support reads this newsgroup and gets much of its information from here, this newsgroup is about as official as it gets. However, the calculator may be new to the vendor in that they didn't recieve it until a month ago. If the vendor doesn't know how HP does the serial numbers, they obviously don't know enough about HP calculators to be a worthwhile calculator vendor. TW === Subject: Re: HP49G+ and Old Serial Number?? posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt I have recently had a CNA451 (with sticker on the case) and a CNA412 (no sticker). I can't honestly say there was any difference in the tactile nature of the keyboards. The CNA412 has a slightly softer touch to it, but that might just be the normal manufacturing tolerances. They didn't look any different either. I have a CNA5 arriving on Monday, and I plan to post a comparison in a few days after that. Adrian === Subject: Re: HP49G+ and Old Serial Number?? I am looking forward to reading about your impressions between the various dates and your newest acquisition. > I have recently had a CNA451 (with sticker on the case) and a CNA412 > (no sticker). I can't honestly say there was any difference in the > tactile nature of the keyboards. The CNA412 has a slightly softer touch > to it, but that might > just be the normal manufacturing tolerances. They didn't look any > different either. I have a CNA5 arriving on Monday, and I plan to post > comparison in a few days after that. > Adrian === Subject: Re: HP49G+ and Old Serial Number?? posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J Some older blister pack does not include the serial number outside. My first hp49g+ CN31xxxxxxx have written the serial directly in the case, not in a sticker. My second Hp, serial CN42xxxxx has the better buttons of all, but 3 button breaks after 6 months and golden paint (incluiding the printed button command references) goes out. My last HP49g+, serial CNA452xxxxx has the same buttons of the first (one button click with a louder clack) As a conclusion, keyboard quality is a random factor. === Subject: Re: HP49G+ and Old Serial Number?? > 2. Should there have been a sticker with the serial number on the outside > of the blister pack and is it possible the vendor removed it on purpose to > avoid disclosure until the seal was broken. I know I am cynical, but I am > also not very happy person. I believe the newer models have the serial number on the blister pack, the older ones don't so it seems you got that backward. Tom Lake === Subject: Re: HP49G+ and Old Serial Number?? I'm afraid to say that your interpretation of the date code is correct. Off memory I think that the 402 series were supposed to have an improved keyboard, but in practice we had the same old issues with them. The good news is that many of the issues were fixed with the ROM 2.00C upgrade - the biggest issues with the keyboard on your machine were the fact that the key hinges break if the unit is used a lot, and the tactile feedback from them is lousy. I've heard a rumour that the 515 serial number items are much improved. === Subject: Re: HP49G+ and Old Serial Number?? what was improved on the units with a 515 code or later. If not, do you by chance remember where you might have read about it so I could research it on for your help. > I'm afraid to say that your interpretation of the date code is correct. > Off memory I think that the 402 series were supposed to have an improved > keyboard, but in practice we had the same old issues with them. The good > news is that many of the issues were fixed with the ROM 2.00C upgrade - > the > biggest issues with the keyboard on your machine were the fact that the > key > hinges break if the unit is used a lot, and the tactile feedback from them > is lousy. > I've heard a rumour that the 515 serial number items are much improved. === Subject: Re: Need help for hp42 Hi! I don't have HP42s calculator, but i wish to use Emu42. Please email me rom image for 42s. ZOleg === Subject: Re: Need help for hp42 > Hi! > I don't have HP42s calculator, but i wish to use Emu42. > Please email me rom image for 42s. The ROM image is copyrighted so you can't have it if you don't buy it (I wish I had it too...). However you have a very good alternative there: http://home.planet.nl/~demun000/thomas_projects/free42/ Arnaud === Subject: attacting Libs and other programs... posting-account=UZpa1g0AAAAGK-b1uuwdq2SDcb2fyNNB hello I recently download some lib and other programs to my 48GX. Do you know how I go about running those programs and Libs? I can see them but I don know how to excutes them in otther to run them. === Subject: Re: attacting Libs and other programs... posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J concrete_jungle ha escrito: > hello > I recently download some lib and other programs to my 48GX. Do you > know how I go about running those programs and Libs? I can see them > but I don know how to excutes them in otther to run them. You must install libraries. In RPN, put the lib in the stack, type the number of a port, 0 for example and STO. Then setup the library with ON+F3 (This runs the inicialization commands for that library) === Subject: userRPL and equation writer posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt when writing a userRPL I want to bung an equation in it. Can I use the equation writer to do this, or do I have to use the more ugly non-pretty print entry? I want the equation to be in the userRPL rather than be in a separate define or something. Adrian === Subject: Re: userRPL and equation writer > when writing a userRPL I want to bung an equation in it. > Can I use the equation writer to do this, or do I have to > use the more ugly non-pretty print entry? I want the equation > to be in the userRPL rather than be in a separate define or > something. You are making a UserRPL program and want an equation in that program's runstream? You can create an equation or expression woth the EQW and leave that on the stack. When editing your userRPL program, you just press the HIST hardkey to access the stack (up- and down-arrows select stack level), then press the ECHO softkey (F1) to insert the contents, of whatever stack level you're pointing at, into your program. ENTER or CANCEL hardkeys brings you back to the program editor. Your program will usually perform better if you create your equation as an RPL stream instead of an algebraic object. As VPN points out, for inspiration look at the ->PRG or ->LST commands in the development lib (256 MENU to access that). Steen === Subject: Re: userRPL and equation writer <42df83ed$0$183$edfadb0f@dtext02.news.tele.dk> posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt > You are making a UserRPL program and want an equation in that program's > runstream? You got it. still learning my way around the machine. In 10 years I'll have it licked, no problem :-) The equations I am writing do not tax the machine. They run in well under a second, and only get used once or twice a week. I have deliberately not used RPL because I find it very opaque as code when I come back a few weeks later to modify it. Algebraic is much better for that (for my eyes anyway), and when I can pop it into the equation writer it is even better. I do use RPL for manual calcs, which is a large part of why I got a HP in the first place. If I really need to squeeze the performance I'll go for hpgcc. My only reservation with that is I need a PC handy to modify the code and re-compile, which sometimes is a bit inconvenient. Adrian === Subject: Re: userRPL and equation writer I assume that what you want to do is as follow: start entering a user RPL program, insert an equation, but you want to enter it through the EQW cause it's better, continue entering the user RPL program.... in this case, do as follow: start typing your program, in the tool menu, do next next, halt, EQW, enter your equation, validate, continue (shift CONT), HIST echo ... and you get your equation in the command line... after that, you can always select your equation in your command kline (through beg and END, and use the edit button in the tool menu to re-edit the equation in the EQW... > when writing a userRPL I want to bung an equation in it. > Can I use the equation writer to do this, or do I have to > use the more ugly non-pretty print entry? I want the equation > to be in the userRPL rather than be in a separate define or > something. > Adrian === Subject: Re: userRPL and equation writer posting-account=KQWjnw0AAACANm1kNgAD2CRArdu2Ufkt Perfect! Merci Cyrille. Adrian === Subject: Re: userRPL and equation writer > when writing a userRPL I want to bung an equation in it. > Can I use the equation writer to do this, or do I have to > use the more ugly non-pretty print entry? I want the equation > to be in the userRPL rather than be in a separate define or > something. > Adrian I don't quite get it but there is that Development Library Command ->PRG === Subject: SUDOKU solver for HP48 posting-account=fPS2ZA0AAAD9dPZJcwFmMSwR7RmVmk0L More info on SUDOKU and algorithm: http://www.eddaardvark.co.uk/sudokusolver.html thx. === Subject: Re: SUDOKU solver for HP48 > More info on SUDOKU and algorithm: > http://www.eddaardvark.co.uk/sudokusolver.html I've got one, but I'm trying to improve its performance. It takes about ten minutes on my HP48GX. Scott -- Scott Hemphill hemphill@alumni.caltech.edu This isn't flying. This is falling, with style. -- Buzz Lightyear === Subject: Re: SUDOKU solver for HP48 posting-account=fPS2ZA0AAAD9dPZJcwFmMSwR7RmVmk0L Scott, would you consider sharing it with the group? Perhaps the joined brains of this group can come up with the best solution. Stefaan Scott Hemphill schreef: More info on SUDOKU and algorithm: http://www.eddaardvark.co.uk/sudokusolver.html > I've got one, but I'm trying to improve its performance. It takes about > ten minutes on my HP48GX. > Scott > -- > Scott Hemphill hemphill@alumni.caltech.edu > This isn't flying. This is falling, with style. -- Buzz Lightyear === Subject: Re: SUDOKU solver for HP48 > Scott, would you consider sharing it with the group? Perhaps the > joined brains of this group can come up with the best solution. Sure. My approach was to write the simplest program I could think of that was consistent with rapid calculation. I think my algorithm shows promise, because I tested it against the one at: http://www.users.waitrose.com/~nihilist/sudoku.html I used the fairly difficult problem (sample9.txt) from the above website: .2....... ...6....3 .74.8.... .....3..2 .8..4..1. 6..5..... ....1.78. 5....9... .......4. (or, if you prefer) 0 2 0 0 0 0 0 0 0 0 0 0 6 0 0 0 0 3 0 7 4 0 8 0 0 0 0 0 0 0 0 0 3 0 0 2 0 8 0 0 4 0 0 1 0 6 0 0 5 0 0 0 0 0 0 0 0 0 1 0 7 8 0 5 0 0 0 0 9 0 0 0 0 0 0 0 0 0 0 4 0 On a PC, I compiled my C program and the one from this website, and found that I could solve this puzzle 100 times in 0.54 seconds, while the other program solved it 100 times in 4.46 seconds. The main differences between my algorithm and the other one are: . mine is much simpler (I think) . if there are no forced choices, I will always pick the square with the fewest number of choices. (This is very important to reduce the size of the search tree.) . precomputed relationships between squares in the others array. . precomputed count of bits in a bitmask in count array. . economical implementation. If my program is compiled normally, it will find *all* the solutions for the given puzzle. If compiled with -DONEONLY, it will stop after it finds the first solution. Now about my RPL version: The ten minutes I reported was for a simple puzzle. The complicated puzzle above would take a lot more time. There's still a lot of hope for improvement, though, because my RPL version is just a prototype, and is a very slow implementation. I used global variables where I could keep things (like the bitmask) on the stack. I also use some very slow list manipulations. But I don't have a lot of time right now, so I'm happy to let other people hack on my code, if they feel so inclined. Note: The INIT function is meant to be called once to set up COUNT and OTHERS. Then you don't have to call it again. Sorry I don't have a native HP48 source file. Note that I used =/ for notequal. ------------------------------------------------------------------------ * -- Scott Hemphill, 26 July 2005 */ #include #include int grid[81]; /* The puzzle grid: contains numbers 1 */ /* through 9, or 0 if a number hasn't */ /* been chosen. */ int bits[81]; /* Bitmap for each square: 0x01 means */ /* that 1 is not possible, 0x02 means */ /* that 2 is not possible, 0x04 means */ /* that 3 is not possible, ..., 0x100 */ /* that 9 is not possible. */ int count[512]; /* Count of number of 1-bits in a bitmap */ int others[81][24]; /* For each square, an array of 20 other */ /* locations affect by this square. For */ /* example, if there is a 6 in square 0, */ /* there are 20 other squares we know */ /* can't contain a six: */ /* 1,2,3,4,5,6,7,8 (horizontal row) */ /* 9,18,27,36,45,54,63,72 (vertcal column) */ /* 1,2,9,10,11,18,19,20 (3x3 square) */ /* That's a total of 24 square numbers, */ /* but 4 of them are duplicates. The */ /* others array is big enough to hold */ /* 24 numbers while it is being built, but */ /* will only 20 numbers will be used when /* it is complete. */ void usage(char *prog) { fprintf(stderr, usage: %s [sudoko input file]n, prog); exit(1); } /* * Used only in initialization, addgroup takes an array of 9 associated * square numbers, and adds then to each others entries in the others * array. A group consistes of the nine square numbers in a row, column * or 3x3 square. No check is made for duplicates; that will be taken * care of when the others array is sorted. The others array is assumed * to be initialized to negative numbers; valid square numbers are 0 through * 80. */ void addgroup(int group[9]) { int i, j, k; for (i = 0; i < 9; i++) { for (j = 0; others[group[i]][j] >= 0; j++) ; for (k = 0; k < 9; k++) { if (k != i) others[group[i]][j++] = group[k]; } } } /* * Used only in initialization, sortgroups (should be called sortothers, * actually) sorts the 81 arrays which are elements of the others array. * Each of these subarrays contains 24 square numbers, although only * 20 of them are unique. Each time a duplicate is found, it is changed * to 99, which makes it sort to the end of the array, where it won't * be used. */ void sortgroups(void) { int i, j, k, tmp; for (i = 0; i < 81; i++) { for (j = 0; j < 23; j++) { for (k = j+1; k < 24; k++) { if (others[i][j] == others[i][k]) { others[i][k] = 99; } else if (others[i][j] > others[i][k]) { tmp = others[i][j]; others[i][j] = others[i][k]; others[i][k] = tmp; } } } } } /* Used to initialize the count array and the others array. count * contains the number of 1-bits in a bitmask. Consider the bitmask * 0x153 (which is 101010011 in binary and 339 in decimal). Therefore * the entry count[339] will contain the value 5. The others array is * initialized to -1s, then groups of 9 associated square numbers are * added one at a time, using the addgroup function. */ void init(void) { int group[9]; int i, j; for (i = 1; i < 512; i<<=1) { for (j = 0; j < 512; j++) { if (i & j) count[j]++; } } for (i = 0; i < 81; i++) { for (j = 0; j < 24; j++) { others[i][j] = -1; /* negative numbers mean unused */ } } for (i = 0; i < 9; i++) { for (j = 0; j < 9; j++) { group[j] = 9*i + j; } addgroup(group); /* add a row of square numbers */ } for (i = 0; i < 9; i++) { for (j = 0; j < 9; j++) { group[j] = 9*j + i; /* add a column of square numbers */ } addgroup(group); } for (i = 0; i < 9; i++) { for (j = 0; j < 9; j++) { group[j] = 27*(i/3) + 9*(j/3) + 3*(i%3) + (j%3); } addgroup(group); /* add a 3x3 square of square numbers */ } sortgroups(); /* sort others array, removing duplicates */ } /* * setbox is used to place a number in the grid, whether it is known, * or a guess. It also updates all the bitmasks for all the associated * squares to indicate that they can't contain this number. */ void setbox(int n, int x) { int i, j; grid[n] = x; j = 1 << (x-1); for (i = 0; i < 20; i++) bits[others[n][i]] |= j; } /* * unsetbox is used to remove a number from the grid. Zero is stored * to indicate an unused square. It is too complicated to calculate * what the bitmask ought to now contain, so it is up to the calling * program to restore the bitmask from a previously saved value. */ void unsetbox(int n, int x) { grid[n] = 0; // it's up to calling program to restore bits } /* * readgrid is used to read the initial data for the grid from a file. * The initial bitmask is calculated at the same time (by calling setbox). */ void readgrid(FILE *f) { char buf[BUFSIZ]; int i, j; for (i = 0; i < 9; i++) { if (!fgets(buf, BUFSIZ, f)) { fprintf(stderr, premature end of inputn); exit(1); } if (strlen(buf) < 10) { fprintf(stderr, invalid input line %dn, i+1); exit(1); } for (j = 0; j < 9; j++) { if ('1' <= buf[j] && buf[j] <= '9') setbox(9*i+j, buf[j]-'0'); } } for (i = 0; i < 81; i++) { if (!grid[i]) continue; if (bits[i] & (1<<(grid[i]-1))) { fprintf(stderr, inconsistent input data for box (%d,%d)n, (i/9)+1, (i%9)+1); exit(1); } } } static long long int num = 0; /* count of number of solutions found */ void printsolution(void) { int i, j; if (num++) printf(n); /* if more than one solution, separate them */ for (i = 0; i < 9; i++) { for (j = 0; j < 9; j++) { printf(%2d, grid[9*i+j]); } printf(n); } } /* selectbox is the heart of the algorithm. It works this way: * First, the grid is searched for an empty square. If there isn't one, * then the grid contains a solution, so it is printed. If an empty * square is found, then it is compared with all other empty squares to * see which one of them has the maximum number of 1-bits in its bitmask. * That square is the one selected, and each of the possible values are * tried in return, calling selectbox recursively to search for a * solution. Note that if a square has 9 1-bits in its bitmask, then * there are no possible values, and selectbox will return without any * further recursion. Otherwise, if a square has 8 1-bits in its bitmask, * then it has a forced choice, and only that choice will be taken at this * level in the recursion. * * Note: compile with -DONEONLY to stop recursion if a solution has already * been found. */ void selectbox(void) { int bitsave[81]; int i, j; int best, bestcount; #ifdef ONEONLY if (num == 0) { /* allow recursion only if no solutions so far */ #endif for (best = 0; best < 81; best++) { /* see if there are any empty squares */ if (!grid[best]) break; } if (best == 81) { /* 81 indicates no empty squares */ printsolution(); return; } bestcount = count[bits[best]]; /* count of bits for the square we found */ for (i = best+1; i < 81; i++) { /* find the biggest count over all squares */ if (grid[i]) continue; /* only consider empty squares */ j = count[bits[i]]; if (j > bestcount) { best = i; bestcount = j; } } for (i = 0; i < 9; i++) { /* Try all numbers for this square */ j = 1 << i; if ((bits[best] & j) == 0) { /* if the number is possible, recurse */ memcpy(bitsave, bits, sizeof bits); /* save bitmask */ setbox(best, i+1); /* place number in square, update bitmask */ selectbox(); /* recurse */ unsetbox(best, i+1); /* unplace number in square */ memcpy(bits, bitsave, sizeof bits); /* restore bitmask */ } } #ifdef ONEONLY } #endif } /* * main program. Calls init to initialize count and others arrays. * Calls readgrid to read initial grid from a file, and calculate initial * bitmask. Calls selectbox to recursively enumerate solution(s). Outputs * summary to stderr. */ int main(int argc, char *argv[]) { FILE *f = stdin; if (argc > 2) usage(argv[0]); if (argc == 2) f = fopen(argv[1], r); if (!f) usage(argv[0]); init(); readgrid(f); selectbox(); #ifdef ONEONLY if (num == 0) fprintf(stderr, no solutions foundn); #else fprintf(stderr, %lld solution%s foundn, num, num==1?:s); #endif return 0; } ------------------------------------------------------------------------ SORTGROUPS: << 1 81 FOR i OTHERS i GET SORT @ {2,2,3,3,...} {} 0 {} 0 WHILE ROT DUP SIZE @ {} 0 {2,2,3,3,...} REPEAT DUP TAIL 4 ROLLD HEAD @ {2,3,3,...} {} 0 2 IF DUP2 == THEN DROP ELSE ROT OVER + ROT DROP SWAP END END DROP2 NEXT 81 ->LIST 'OTHERS' STO ADDGROUP: << -> g << OTHERS {{..} {..} .. {..}} 1 9 FOR i g i GET DUP2 GET {{..} {..} .. {..}} 73 {..} 1 9 FOR j IF i j =/ THEN g j GET + END NEXT PUT NEXT 'OTHERS' STO >> INIT: << #0 WHILE DUP #512d < REPEAT 0 IF OVER #1d AND #0 == THEN 1 + END IF OVER #2d AND #0 == THEN 1 + END IF OVER #4d AND #0 == THEN 1 + END IF OVER #8d AND #0 == THEN 1 + END IF OVER #16d AND #0 == THEN 1 + END IF OVER #32d AND #0 == THEN 1 + END IF OVER #64d AND #0 == THEN 1 + END IF OVER #128d AND #0 == THEN 1 + END IF OVER #256d AND #0 == THEN 1 + END SWAP #1 + END DROP 512 ->LIST 'COUNT' STO 1 81 START {} NEXT 81 ->LIST 'OTHERS' STO 1 81 FOR i i NEXT 1 9 START 9 ->LIST ADDGROUP NEXT 0 8 FOR i 0 8 FOR j j 9 * i + 1 + NEXT NEXT 1 9 START 9 ->LIST ADDGROUP NEXT 0 8 FOR i 0 8 FOR j i 3 / IP 27 * j 3 / IP 9 * + i 3 MOD 3 * + j 3 MOD + 1 + NEXT NEXT 1 9 START 9 ->LIST ADDGROUP NEXT SORTGROUPS SETBOX: << DUP2 GRID 3 ROLLD PUT 'GRID' STO OTHERS ROT GET SWAP 1 - 2 SWAP ^ R->B @ {k} bit BITS @ {k} bit BITS WHILE ROT DUP SIZE REPEAT DUP TAIL 4 ROLLD HEAD @ {..} bit BITS k DUP2 GET 4 PICK OR PUT @ {..} bit BITS END DROP 'BITS' STO DROP UNSETBOX: << GRID SWAP 0 PUT 'GRID' STO READGRID: << IF DUP TYPE 3 == THEN IF DUP SIZE {9 9} == THEN OBJ-> DROP 81 ->ARRY 'GRID' STO 1 81 START #0 NEXT 81 ->LIST 'BITS' STO 1 81 FOR i i GRID i GET IF DUP THEN SETBOX ELSE DROP2 END NEXT ELSE Array not 9x9 HALT END ELSE No array specified HALT END ADDSOLUTION: << GRID OBJ-> DROP {9 9} ->ARRY 'SOLUTION' STO+ SELECTBOX: << 0 1 DO IF GRID OVER GET NOT THEN SWAP DROP 82 ELSE 1 + END UNTIL DUP 81 > END DROP IF DUP NOT THEN DROP ADDSOLUTION ELSE BITS OVER GET B->R 1 + COUNT SWAP GET OVER 1 + WHILE DUP 81 <= REPEAT IF GRID OVER GET NOT THEN BITS OVER GET B->R 1 + COUNT SWAP GET IF 3 PICK OVER < THEN 4 ROLL 4 ROLL DROP2 OVER ELSE DROP END END 1 + END DROP2 0 WHILE DUP 9 < REPEAT IF BITS 3 PICK GET 2 3 PICK ^ R->B AND #0 == THEN DUP2 BITS 3 ROLLD 1 + SETBOX SELECTBOX 'BITS' STO OVER UNSETBOX END 1 + END DROP2 END MAIN: << READGRID {} 'SOLUTION' STO SELECTBOX SOLUTION IF DUP SIZE 1 == THEN 1 GET END ------------------------------------------------------------------------ Scott -- Scott Hemphill hemphill@alumni.caltech.edu This isn't flying. This is falling, with style. -- Buzz Lightyear === Subject: SUDOKU solver for HP48 posting-account=fPS2ZA0AAAD9dPZJcwFmMSwR7RmVmk0L More info on SUDOKU and algorithm: http://www.eddaardvark.co.uk/sudokusolver.html thx. === Subject: Re: SUDOKU solver for HP48 >More info on SUDOKU and algorithm: >http://www.eddaardvark.co.uk/sudokusolver.html If you join HPCC (www.hpcc.org) then you will receive our (print only) journal Datafile, the May/June edition of which contains not one but two Sudoku solvers - one fast one for the HP-71B by Valentin Albillo and a much slower one for the HP-12C Platinum by Tony Hutchins as a reminder that the 12C is a pretty capable machine despite being a business model. -- Bruce Horrocks Surrey England @.plus.com -- fix the obvious for email === Subject: Re: SUDOKU solver for HP48 I think it would be more interesting with a sudoku problem generator, that allows problems to be solved on the calculator. Of course, this program would have to include a sudoku solver. -- Daniel === Subject: Re: SUDOKU solver for HP48 Lidstr.9am a .8ecrit dans le message de > I think it would be more interesting with a sudoku problem generator, that > allows problems to be solved on the calculator. Of course, this program > would have to include a sudoku solver. Generating valid solved grids is fairly easy (see for example) > -- > Daniel === Subject: Re: SUDOKU solver for HP48 posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J I dont find anything. Sodoku becomes popular in this year, maybe is too soon. :D But with a few knows about userRpl is very easy do a solver :) === Subject: Re: SUDOKU solver for HP48 posting-account=fPS2ZA0AAAD9dPZJcwFmMSwR7RmVmk0L Avatar_e schreef: > I dont find anything. Sodoku becomes popular in this year, maybe is too > soon. :D > But with a few knows about userRpl is very easy do a solver :) How would you do it? Start with a LIST containing 81 LISTs with the numbers 1 to 9. Or would you use a matrix? === Subject: Re: SUDOKU solver for HP48 stefanmo_hp48 schrieb im Newsbeitrag > Avatar_e schreef: >> I dont find anything. Sodoku becomes popular in this year, maybe is too >> soon. :D >> But with a few knows about userRpl is very easy do a solver :) > How would you do it? Start with a LIST containing 81 LISTs with the > numbers 1 to 9. Or would you use a matrix? Lists are too memory consuming and too slow to access. I use matrices (in the still not finished 1st version) so far, but the final version will use flat streams of bytes for the data items. This reduces memory usage significantly. A real 9x9 matrix consumes about 663 bytes, whereas a corresponding byte stream will consume about 81 bytes in the byte-word case, or 40.5 bytes in the nibble-word case, each stream also has a small header of a few nibs. It'll take some time to finish the 2nd version, since I don't have much time currently:-( Raymond === Subject: Re: SUDOKU solver for HP48 Hi all, here's a small but nice update. My HP-48 port of the SuDoKu Solver algo from Valentin runs fine (and fast) now. Timings on a real HP-48G : [Test case : seconds] #01 : 18.81 #13 : 86.93 #14 : 334.90 Test cases were published by Valentin Albillo in HPCC Datafile V24 N3 P29 ff. Test case #14 is kinda 'diabolical example', which runs about quarter of an hour on a real HP-71B, and now in about 5.5 minutes on a normal HP-48G;-) Again, these values are in seconds, and made with a real HP-48G, not Emu48 ! On the HP-48G series, you'll just have the input matrix on stack level 1, and then edit it or key the values in when the program prompts you with the MatrixWriter. Then you'll be presented with a custom input form, where you can accept or adjust the usual parameter settings, like width, depth, and verbose. Since the HP-48 S series (S,SX) doesn't have an input form engine built-in, you'll be prompted for the parameter settings if using an HP-48 S series machine. The program will be on www.hpcalc.org within the next few days. If I find the time, then maybe there'll be an even faster version in the near future. Currently the program size is just below 3kb . The program uses almost only supported entry points, and the code is clean, so it should be easy (relatively) to convert it to run on the newer machines, like the rubber duck, or the broken gold tooth. With a suitable TOL it should also run on the 38G/39G(+) type machines Before you ask: I won't do the conversion, since the newer machines are not of much interest for me. Raymond (HPCC #1120) Raymond Del Tondo schrieb im Newsbeitrag > stefanmo_hp48 schrieb im Newsbeitrag >> Avatar_e schreef: > I dont find anything. Sodoku becomes popular in this year, maybe is too > soon. :D > But with a few knows about userRpl is very easy do a solver :) >> How would you do it? Start with a LIST containing 81 LISTs with the >> numbers 1 to 9. Or would you use a matrix? > Lists are too memory consuming and too slow to access. > I use matrices (in the still not finished 1st version) so far, > but the final version will use flat streams of bytes for the data items. > This reduces memory usage significantly. > A real 9x9 matrix consumes about 663 bytes, > whereas a corresponding byte stream will consume > about 81 bytes in the byte-word case, > or 40.5 bytes in the nibble-word case, > each stream also has a small header of a few nibs. > It'll take some time to finish the 2nd version, > since I don't have much time currently:-( > Raymond === Subject: Re: SUDOKU solver for HP48 Will you include a commented source code? That would make it easier to suggest improvements Arnaud === Subject: Re: SUDOKU solver for HP48 > Hi all, > here's a small but nice update. > My HP-48 port of the SuDoKu Solver algo > from Valentin runs fine (and fast) now. > Timings on a real HP-48G : > [Test case : seconds] > #01 : 18.81 > #13 : 86.93 > #14 : 334.90 > Test cases were published by Valentin Albillo in > HPCC Datafile V24 N3 P29 ff. > Test case #14 is kinda 'diabolical example', > which runs about quarter of an hour on a > real HP-71B, and now in about 5.5 minutes > on a normal HP-48G;-) > Again, these values are in seconds, > and made with a real HP-48G, not Emu48 ! > On the HP-48G series, you'll just > have the input matrix on stack level 1, > and then edit it or key the values in when the > program prompts you with the MatrixWriter. > Then you'll be presented with a custom input form, > where you can accept or adjust the usual > parameter settings, like width, depth, and verbose. > Since the HP-48 S series (S,SX) doesn't > have an input form engine built-in, > you'll be prompted for the parameter settings > if using an HP-48 S series machine. > The program will be on www.hpcalc.org > within the next few days. > If I find the time, then maybe there'll be > an even faster version in the near future. > Currently the program size is just below 3kb . > The program uses almost only supported entry points, > and the code is clean, so it should be easy > (relatively) to convert it to run on the newer machines, > like the rubber duck, or the broken gold tooth. > With a suitable TOL it should also run > on the 38G/39G(+) type machines > Before you ask: I won't do the conversion, > since the newer machines are not of much interest for me. X Who will do the conversion then? It would nice to know the speed of the code in a hp 49g+ Is it under magical 100 seconds in test case #14 ?? VPN === Subject: Re: SUDOKU solver for HP48 Hi Veli, now test case 13 runs in 49 seconds, andtest case 14 consumes only 169 seconds on a real 48G;-) I don't think it can be further sped up by normal means. The next step would be to rewrite the whole thing in ML. If I find the time, I'll put the current HP-48 version up to www.hpcalc.org tomorrow. As I didn't promise a rubber duck version, I can't guarantee that it will happen... The package will consist of the binary, a small doc file, a description of the principal program mechanics, and some goodies, like routines to convert matrices to streams and back, and a fast accesss routine for the streams. Raymond Veli-Pekka Nousiainen schrieb im Newsbeitrag >> Hi all, >> here's a small but nice update. >> My HP-48 port of the SuDoKu Solver algo >> from Valentin runs fine (and fast) now. >> Timings on a real HP-48G : >> [Test case : seconds] >> #01 : 18.81 >> #13 : 86.93 >> #14 : 334.90 >> Test cases were published by Valentin Albillo in >> HPCC Datafile V24 N3 P29 ff. >> Test case #14 is kinda 'diabolical example', >> which runs about quarter of an hour on a >> real HP-71B, and now in about 5.5 minutes >> on a normal HP-48G;-) >> Again, these values are in seconds, >> and made with a real HP-48G, not Emu48 ! >> On the HP-48G series, you'll just >> have the input matrix on stack level 1, >> and then edit it or key the values in when the >> program prompts you with the MatrixWriter. >> Then you'll be presented with a custom input form, >> where you can accept or adjust the usual >> parameter settings, like width, depth, and verbose. >> Since the HP-48 S series (S,SX) doesn't >> have an input form engine built-in, >> you'll be prompted for the parameter settings >> if using an HP-48 S series machine. >> The program will be on www.hpcalc.org >> within the next few days. >> If I find the time, then maybe there'll be >> an even faster version in the near future. >> Currently the program size is just below 3kb . >> The program uses almost only supported entry points, >> and the code is clean, so it should be easy >> (relatively) to convert it to run on the newer machines, >> like the rubber duck, or the broken gold tooth. >> With a suitable TOL it should also run >> on the 38G/39G(+) type machines >> Before you ask: I won't do the conversion, >> since the newer machines are not of much interest for me. > Who will do the conversion then? > It would nice to know the speed of the code in a hp 49g+ > Is it under magical 100 seconds in test case #14 ?? > VPN === Subject: Re: SUDOKU solver for HP48 -=[ Sun, 14.8.05 3:58 p.m. +1200 (NZT) ]=- Hello Raymond Del Tondo, in message ID : > Hi Veli, > now test case 13 runs in 49 seconds, > andtest case 14 consumes only 169 seconds on a real 48G;-) Very impressive! [...] > If I find the time, I'll put the current HP-48 version > up to www.hpcalc.org tomorrow. Great! > As I didn't promise a rubber duck version, > I can't guarantee that it will happen... [...] > The program uses almost only supported entry points, > and the code is clean, so it should be easy > (relatively) to convert it to run on the newer machines, > like the rubber duck, or the broken gold tooth. LOL - I have the dental one :-) It would nice to know the speed of the code in a hp 49g+ Is it under magical 100 seconds in test case #14 ?? VPN Here I cheated - with help - and used hpgcc to compile Nihilist's sudoku.c for the 49g+. It does #14 in under 2 seconds. It takes only the puzzle as input (no recursive limits are required). -- Tony === Subject: Re: SUDOKU solver for HP48 Tony Hutchins schrieb im Newsbeitrag > -=[ Sun, 14.8.05 3:58 p.m. +1200 (NZT) ]=- > Hello Raymond Del Tondo, > in message ID : >> Hi Veli, >> now test case 13 runs in 49 seconds, >> andtest case 14 consumes only 169 seconds on a real 48G;-) > Very impressive! Test case 14 now runs in about 108 seconds:-) A few days ago I uploaded a version of the package to hpcalc.org. That version takes about 139 seconds for test case 14. > [...] > It would nice to know the speed of the code in a hp 49g+ > Is it under magical 100 seconds in test case #14 ?? > VPN > Here I cheated - with help - and used hpgcc to compile > Nihilist's sudoku.c for the 49g+. > It does #14 in under 2 seconds. > It takes only the puzzle as input (no recursive limits are > required). I can't compete with that, of course. But (near to) 100 seconds for a SysRPL program on a normal HP-48G isn't that bad, too;-) It should run in at least a quarter of the time on a broken gold tooth, a rubber duck version (if one existed) should run slightly faster than the HP-48 original. Of course these values are only hypothetical, since there only exists an HP-48 version... Raymond === Subject: Re: SUDOKU solver for HP48 -=[ Sun, 21.8.05 08:39 a.m. +1200 (NZT) ]=- Hello Raymond Del Tondo, in message ID : [...] > Test case 14 now runs in about 108 seconds:-) A very humanised speed :-) It's good to see a measurable variation in time - gives one an idea of difficulty. Here everything happens in flash :-( [...] Here I cheated - with help - and used hpgcc to compile Nihilist's sudoku.c for the 49g+. It does #14 in under 2 seconds. It takes only the puzzle as input (no recursive limits are required). > I can't compete with that, of course. > But (near to) 100 seconds for a SysRPL program > on a normal HP-48G isn't that bad, too;-) Simply amazing!! I have cheated in more ways than one too - hehe. I haven't upgraded the Nihilist algorithm to resist recursion for cases like #13 where recursion is not really needed. Your program will do that one *much* more efficiently (relatively speaking Nihilist might be say 100X faster on #14 but only say 5X faster on #13, but on all other cases Nihilist is pretty much immesurably fast, on the gold tooth). > It should run in at least a quarter of the time on a broken gold tooth, > a rubber duck version (if one existed) should run > slightly faster than the HP-48 original. > Of course these values are only hypothetical, > since there only exists an HP-48 version... Next step - generalise it to handle other than 3 by 3 blocks :-) -- Tony === Subject: Re: SUDOKU solver for HP48 X > It would nice to know the speed of the code in a hp 49g+ > Is it under magical 100 seconds in test case #14 ?? > VPN > Here I cheated - with help - and used hpgcc to compile > Nihilist's sudoku.c for the 49g+. > It does #14 in under 2 seconds. > It takes only the puzzle as input (no recursive limits are > required). Where can I download the commented source and the binary VPN === Subject: Re: SUDOKU solver for HP48 <4FJNe.10661$ys3.8262@reader1.news.jippii.net> -=[ Sun, 21.8.05 08:23 a.m. +1200 (NZT) ]=- Hello Veli-Pekka Nousiainen, in message ID <4FJNe.10661$ys3.8262@reader1.news.jippii.net> : > Where can I download the commented source and the binary http://homepages.paradise.net.nz/th/sud49gp.zip should work - it's the top link on http://homepages.paradise.ne.nz/th just under Felix the cat :-) As uploaded last Sunday - yup it's still there, a 16Kb file including the C source (99.99% as per Nihilist but renamed as sudhpgcc.c, a 26K file), a little readme.txt and SUDHP, the binary (just a complete directory copied to SD card then zipped). SUDHP is 15K. I never tested how much memory is needed to run it even with the maximum 49 levels of recursion - probably about 30K free should do it. I will generalise it to handle any block-size - eg a 3 by 4 or a 4 by 3 or 2 by 6 or 6 by 2 could be used for a 12 by 12 Sudoku... but that means 2 more parameters. -- Tony === Subject: Re: SUDOKU solver for HP48 <10583619ROBOTLX@news.clear.net.nz> -=[ Sun, 21.8.05 3:08 p.m. +1200 (NZT) ]=- Hello Tony Hutchins, in message ID <10583619ROBOTLX@news.clear.net.nz> : > I never tested how much memory is needed > to run it even with the maximum 49 levels of recursion - > probably about 30K free should do it. No, for 49 levels we need more like 100K free. So, if it is set running on a 9 by 9 array of zeros it probably needs the 100K - if it only has 50 will also need a paper clip reset (not tested). For real puzzles with one solution 50K should be fine - even less. > I will generalise it to handle any block-size - eg a 3 by > 4 or a 4 by 3 or 2 by 6 or 6 by 2 could be used for a 12 > by 12 Sudoku... but that means 2 more parameters. Done - easy in theory too, but it gets really hungry for memory, and if it runs out it needs a paper clip. Fun though! Can do Latin squares as a byproduct - 1 by n blocks :-) Too dangerous to upload :-) I'm not quite sure if I can check memory availability before each recursive call. Or maybe pre-calculate the maximum recursive depth allowed. Something for later :-) -- Tony === Subject: Re: SUDOKU solver for HP48 <11012642ROBOTLX@news.clear.net.nz> -=[ Wed, 24.8.05 9:24 p.m. +1200 (NZT) ]=- Hello Tony Hutchins, 3 days 05h57m ago, on Sun, in message ID <11012642ROBOTLX@news.clear.net.nz> : [...] I will generalise it to handle any block-size - eg a 3 by 4 or a 4 by 3 or 2 by 6 or 6 by 2 could be used for a 12 by 12 Sudoku... but that means 2 more parameters. The default block size is {3 3}. > Done - easy in theory too, but it gets really hungry for > memory, and if it runs out it needs a paper clip. Fun though! > Can do Latin squares as a byproduct - 1 by n blocks :-) Too > dangerous to upload :-) It was uploaded - with a global MAX_SIZE of 9 in the source. Dynamic arrays in C are way beyond me, but the source can simply be re-compiled with any MAX_SIZE. Now for example you can solve an 8 by 8 puzzle with blocks of {4 2} or {2 4} or {1 8} etc..., including the {3 3} :-) -- Tony #281 Many people would sooner die than think. In fact they do. Bertrand Russell === Subject: Re: SUDOKU solver for HP48 Hi Veli, Veli-Pekka Nousiainen schrieb im Newsbeitrag >> Hi all, >> here's a small but nice update. >> My HP-48 port of the SuDoKu Solver algo >> from Valentin runs fine (and fast) now. >> Timings on a real HP-48G : >> [Test case : seconds] >> #01 : 18.81 >> #13 : 86.93 >> #14 : 334.90 >> Test cases were published by Valentin Albillo in >> HPCC Datafile V24 N3 P29 ff. >>[..] >> Before you ask: I won't do the conversion, >> since the newer machines are not of much interest for me. > Who will do the conversion then? Dunno. Maybe one of the people who actually like those machines;-) Okok, maybe I can provide a FHB (rubber duck) version, too. But that's not guaranteed... > It would nice to know the speed of the code in a hp 49g+ > Is it under magical 100 seconds in test case #14 ?? The current version is the first approach, but not too bad one IMHO;-) Maybe I can speed up things a bit in the near future. There are some ideas for changes inside the virtual drawer... Raymond (HPCC #1120) === Subject: Re: SUDOKU solver for HP48 Hi Veli, again a small update: Raymond Del Tondo schrieb im Newsbeitrag > Hi Veli, > Veli-Pekka Nousiainen schrieb im Newsbeitrag > Hi all, > here's a small but nice update. > My HP-48 port of the SuDoKu Solver algo > from Valentin runs fine (and fast) now. > Timings on a real HP-48G : > [Test case : seconds] > #01 : 18.81 > #13 : 86.93 > #14 : 334.90 > [..] >> It would nice to know the speed of the code in a hp 49g+ >> Is it under magical 100 seconds in test case #14 ?? Test case #14 now takes 244.5 seconds on a normal HP-48G, that's just some seconds above 4 minutes :-) The other cases are faster than before, too. So if there will be a 49g+ version at some point (again: not from me), that should run below the 'magical 100 secs' even with the current algo. The speed gain is due to some more optimizations, like replacing some arrays by nibble streams. That means less memory to be moved, and optimized data access routines. Let's see if I can squeeze out some more seconds;-) Raymond (HPCC #1120) === Subject: Re: SUDOKU solver for HP48 Hi all, my HP-48 port of the SUDOKU V2 from Valentin Albillo now works! There are still some things to optimize, but at least on Emu48 in 'real speed' mode, my SDK48 solves test case #01 from Datafile in about 50 seconds. Not tried all of the test matrices, but I estimate the speed to be at least comparable to the HP-71B version. Raymond Raymond Del Tondo schrieb im Newsbeitrag > stefanmo_hp48 schrieb im Newsbeitrag >> Avatar_e schreef: > I dont find anything. Sodoku becomes popular in this year, maybe is too > soon. :D > But with a few knows about userRpl is very easy do a solver :) >> How would you do it? Start with a LIST containing 81 LISTs with the >> numbers 1 to 9. Or would you use a matrix? > Lists are too memory consuming and too slow to access. > I use matrices (in the still not finished 1st version) so far, > but the final version will use flat streams of bytes for the data items. > This reduces memory usage significantly. > A real 9x9 matrix consumes about 663 bytes, > whereas a corresponding byte stream will consume > about 81 bytes in the byte-word case, > or 40.5 bytes in the nibble-word case, > each stream also has a small header of a few nibs. > It'll take some time to finish the 2nd version, > since I don't have much time currently:-( > Raymond === Subject: Re: SUDOKU solver for HP48 posting-account=i7SfpAwAAACLYtWZL1afsMAUuHZzafp0 This is an interesting discussion. I'm not very expert to solve sudoku schemes, but I'd like to understand a bit more about this game. That is: - why a scheme is more difficult than other? in general I think a scheme is difficult if the number of objects (numbers or letters)already in the squares is little and such as, for this number, there is only a solution (the most little number of objects such as there is only a solution). so take a complete scheme and take away some numbers. well, if you take away two numbers you surely have a scheme completely determined but very very simple because you must add only two numbers. but if you go on to take away the numbers you can obtain different schemes from a initial complete scheme. to obtain the most difficult scheme (difficult for the high number of line of reasoning) you probably must take away the highest number of numbers from scheme but until this scheme is determined by two or more solutions. To do this starting from a complete scheme is not very easy and I think the software do this with attempts. is it right? - I want to calculate the number of possible complete schemes of sudoku considering the rules (every line must have all the number 1-9, every column....). Starting from the top at left of a scheme (first square) I'd like to discover how many numbers I must write until I obtain a scheme with only a solution (completely determined). then I'd calculate the total number of possible combinations (for the first line the num of combinations is 362880 and to go on it is necessary consider the sud rules and some trick). maybe I'll obtain a wrong result but if one doesn't try? - Do all the sudoku schemes that you find in newspaper, for example, have the minimum number of numbers in squares such as there is only a solution or some easy scheme has a num of numbers a bit bigger than the minimum to ensure the existence of only a solution? if what I've writed is clear, does someone discuss about these points? Matteo === Subject: Re: SUDOKU solver for HP48 -=[ Sat, 6.8.05 08:09 a.m. +1200 (NZT) ]=- Hello Matteo, [...] > if what I've writed is clear, does someone discuss about http://it.wikipedia.org/wiki/Sudoku is a great starting point, with links to discussion forums (eg www.sudoku.com/forums). -- Tony === Subject: Re: SUDOKU solver for HP48 <10545912ROBOTLX@news.clear.net.nz> posting-account=i7SfpAwAAACLYtWZL1afsMAUuHZzafp0 Tony Hutchins ha scritto: > -=[ Sat, 6.8.05 08:09 a.m. +1200 (NZT) ]=- > Hello Matteo, > [...] if what I've writed is clear, does someone discuss about > http://it.wikipedia.org/wiki/Sudoku > is a great starting point, with links to discussion forums (eg > www.sudoku.com/forums). > -- > Tony I've writed in this forum (hp48) because I consider the people that write here open mind and strong mind. and I'm an hp user :-). But you are right, otherwise this forum would go out from the theme that is hp48/49. best wishes matteo === Subject: Re: SUDOKU solver for HP48 >I dont find anything. Sodoku becomes popular in this year, maybe is too > soon. :D > But with a few knows about userRpl is very easy do a solver :) I believe Raymond is working on it... Arnaud === Subject: HP48GX Sudden Memory Loss I was using my 48GX last night when I suddenly got a memory loss and the dreaded Recover Memory screen. I was doing some programming at the time and cannot figure out what I might have done to cause that. I selected 'Yes' and everything came back. I promptly stored a copy of the program I was working on, onto the Ram Card (port 1). About an hour later, same thing, except this time, I ended up with a blank machine...*including the ram card*. Does anyone have any idea how/why this could happen? I was just key pushing away, in edit mode with only a small subroutine on the stack being edited. Can overly fast key presses confound the hardware and cause some sort of IRQ conflict in the system? And how/why would that *also* blow away the ram card memory? If I had inadvertantly jiggled the batteries, I could understand some sort of effect, but not the entire loss of memory in the system *and* in the ram-card... which does have its own battery. (Yes, I possibly should have changed the card to read-only...) Geoff === Subject: Re: HP48GX Sudden Memory Loss maybe there was some critical 3rd party software installed on your unit? Or did you use some SYSEVAL addresses in your code? Raymond R. G. Newbury schrieb im Newsbeitrag >I was using my 48GX last night when I suddenly got a memory loss and > the dreaded Recover Memory screen. > I was doing some programming at the time and cannot figure out what I > might have done to cause that. > I selected 'Yes' and everything came back. I promptly stored a copy of > the program I was working on, onto the Ram Card (port 1). > About an hour later, same thing, except this time, I ended up with a > blank machine...*including the ram card*. > Does anyone have any idea how/why this could happen? I was just key > pushing away, in edit mode with only a small subroutine on the stack > being edited. Can overly fast key presses confound the hardware and > cause some sort of IRQ conflict in the system? > And how/why would that *also* blow away the ram card memory? If I had > inadvertantly jiggled the batteries, I could understand some sort of > effect, but not the entire loss of memory in the system *and* in the > ram-card... which does have its own battery. > (Yes, I possibly should have changed the card to read-only...) > Geoff === Subject: Re: HP48GX Sudden Memory Loss On Thu, 21 Jul 2005 09:17:47 UTC, Raymond Del Tondo > maybe there was some critical 3rd party > software installed on your unit? No. Only home-grown user rpl. And I was in program mode: editing a program. > Or did you use some SYSEVAL addresses in your code? No. None. As noted I was editing. That's what makes this so weird. If I had pulled the batteries I could expect a minute of memory retention. And the ram card should be unaffected even so. But *everything* was blown away, instantly. G. > Raymond > R. G. Newbury schrieb im Newsbeitrag I was using my 48GX last night when I suddenly got a memory loss and the dreaded Recover Memory screen. I was doing some programming at the time and cannot figure out what I might have done to cause that. I selected 'Yes' and everything came back. I promptly stored a copy of the program I was working on, onto the Ram Card (port 1). About an hour later, same thing, except this time, I ended up with a blank machine...*including the ram card*. Does anyone have any idea how/why this could happen? I was just key pushing away, in edit mode with only a small subroutine on the stack being edited. Can overly fast key presses confound the hardware and cause some sort of IRQ conflict in the system? And how/why would that *also* blow away the ram card memory? If I had inadvertantly jiggled the batteries, I could understand some sort of effect, but not the entire loss of memory in the system *and* in the ram-card... which does have its own battery. (Yes, I possibly should have changed the card to read-only...) Geoff === Subject: Re: HP48GX Sudden Memory Loss posting-account=9wLUeQ0AAACd9G5v3T6VPoZctF9pMIKW The same happened to me some time ago, but I haven't found a decent explanation yet. By the way, I had (and still have) Metakernel installed on my 48 - was the same for you? Giancarlo (Italy) === Subject: how to update ROM posting-account=O6knhQ0AAAD8zE21TyMVwtCasYnRdHOm i installed the connectivity software ( win 98 ) and downloaded the latest ROM from hp site. Now i click ROM -> download calculator rom. i choose the file 49Gp_50.bin. then the program tells me to do this: 1. Hold + and - (3 secs.) while pressing RESET 2. Press 1 to load OS/ROM 3. Press 1 for USB 4. Unplug then plug USB cable 5. Press OK below. ok which RESET is this ? Is that the hole in the back of the calculator? so i have to manage to press two keys ( plus and minus together ) then use a paper click to press that button ? im kinda awkward so that seems really hard. === Subject: Re: how to update ROM posting-account=O6knhQ0AAAD8zE21TyMVwtCasYnRdHOm ok i've just uploaded ( or downloaded from calculators point of view ). thankfully it has all worked. === Subject: Re: how to update ROM posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE That's it, the small hole... ;) But you can also simply doing this: - 1) press [ON]-[D], it shows the version - 2) [ON]-[C] and just after pressing [ON]-[C], press simultaneously [+][-]. With this procedure, no need to press with a key with your finger while pressing the remote control with the tong while pushing the oven's button with your left foot, all simultaneously ;) I strongly advise you to flash the new ROM from an SD card: copy the ROM on an SD card and proceed to the OS replacement. I have already had problems with flashing the ROM with Conn4x: it erased the ROM and after, nothing happened. I couldn't even switch off the calculator because it has no more OS. Be carefull by doing this. Fred. === Subject: Re: how to update ROM > I strongly advise you to flash the new ROM from an SD card: copy the > ROM on an SD card and proceed to the OS replacement. > I have already had problems with flashing the ROM with Conn4x: While this is certainly possible, I think you are needlessly frightening the users! I have done the updates with Conn4x many, many times (I have 3 49G+ calcs to keep up-to-date) and I have never experienced a problem. Follow the directions carefully - Don't forget to read all the proompt and UNPLUG and plug-in the USB cable after doing the + - RESET thing. -- - - - - - - - - - - - - - - - - Bill Graves RKBA! bgraves@ix.netcom.com === Subject: Re: how to update ROM posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Hi William, What's up? I don't blame anybody and I don't want to frighten anybody here. I just report what happened to my calc when I tried to update my calc to ROM 2.0 with Conn4x. I tried many times to update the ROM with Conn4x and it always fails after the erase part... I think I already mentionned this problem, but maybe it's happened only to me and I don't do the correct process (this is also a possibility) :p Since this time, I have an SD card with an SD card reader/writter which is more convinient :) Fred. === Subject: Re: how to update ROM > I tried many times to update the ROM with Conn4x and it always fails > after the erase part... I think I already mentionned this problem, but > maybe it's happened only to me and I don't do the correct process (this > is also a possibility) :p It is easy to overlook a step (the final one about plug/unplug the USB cable is very important)! > Since this time, I have an SD card with an SD card reader/writter which > is more convinient :) I have many SD cards but I always use Conn4x. It takes a lot longer with the SD cards for the update. This does not even include the copying of files to the SD cards. So maybe the difference might be a few minutes!!!! -- - - - - - - - - - - - - - - - - Bill Graves RKBA! bgraves@ix.netcom.com === Subject: Re: how to update ROM > It is easy to overlook a step (the final one about plug/unplug the USB cable > is very important)! Funny.. I've never done it, and it always worked well for my calculator. JY === Subject: Re: how to update ROM >> It is easy to overlook a step (the final one about plug/unplug the USB >> cable is very important)! > Funny.. I've never done it, and it always worked well for my calculator. Funny..I always do it and it STILL works ;-) VPN === Subject: Re: how to update ROM >> It is easy to overlook a step (the final one about plug/unplug the USB >> cable is very important)! > Funny.. I've never done it, and it always worked well for my calculator. > JY I think I have done some updates without it also (but I am not sure). What I do know is that the USB driver sometimes gets lost during the RESET. This can lead to a failure to download. -- - - - - - - - - - - - - - - - - Bill Graves RKBA! bgraves@ix.netcom.com === Subject: Re: how to update ROM posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Hi William, What's up? I don't blame anybody and I don't want to frighten anybody here. I just report what happened to my calc when I tried to update my calc to ROM 2.0 with Conn4x. I tried many times to update the ROM with Conn4x and it always fails after the erase part... I think I already mentionned this problem, but maybe it's happened only to me and I don't do the correct process (this is also a possibility) :p Since this time, I have an SD card with an SD card reader/writter which is more convinient :) Fred. === Subject: Re: how to update ROM posting-account=O6knhQ0AAAD8zE21TyMVwtCasYnRdHOm i dont have SD card. i guess i will just postpone the ROM upgrade. === Subject: HPGCC: How to create a surface? (gglsurface) posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Hello people, Using HPGCC's ggl library, I'd like to know how I can correctly initialize the members of a gglsurface structure that will be 80 pixels in width and 80 pixels in height (for example). Fred. === Subject: Re: HPGCC: How to create a surface? (gglsurface) posting-account=FCAFHAwAAADaPCjuTWXm3ER18SIzldtO > Hello people, > Using HPGCC's ggl library, I'd like to know how I can correctly > initialize the members of a gglsurface structure that will be 80 pixels > in width and 80 pixels in height (for example). > Fred. gglsurface mysurface; mysurface.x=mysurface.y=0; mysurface.width=80; mysurface.addr=(int *)malloc(80*80/2); // width*height/2 (2 pixels per byte) And that's it. Claudio === Subject: Re: HPGCC: How to create a surface? (gglsurface) posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Hi Claudio, In that case, I have a situation then... Take a look at your mail. Fred. === Subject: I revived my 49g+! posting-account=cDjNzQ0AAABYi8nSTqQBPwRMH3jHIbL8 A few weeks ago I posted about a dead 49 that wouldn't respond to anything. Well today just for kicks, I pulled it out and replaced all the batteries. Nothing happened, RESET didn't work, and I figured it was toast. Now the strange part. About 5 MINUTES later, I heard a beep. I looked in the drawer where it went, and to my suprise I saw a TTRM. I answered no, and the calculator was working again. I still had the two missing flash banks. I then loaded up the old 2MB rom 1.23 hoping that would reset the flash banks. Sure enough, I looked in the filer and it showed 788kb in port 2. Woohoo! Then I updated to the latest ROM and still had 700kb in my port 2. Everything is working now. So if it dies on you, don't give up I guess. =) TW === Subject: Re: I revived my 49g+! > A few weeks ago I posted about a dead 49 that wouldn't respond to > anything. Well today just for kicks, I pulled it out and replaced all > the batteries. Nothing happened, RESET didn't work, and I figured it > was toast. Now the strange part. About 5 MINUTES later, I heard a > beep. I looked in the drawer where it went, and to my suprise I saw a > TTRM. > I answered no, and the calculator was working again. I still had the > two missing flash banks. I then loaded up the old 2MB rom 1.23 hoping > that would reset the flash banks. Sure enough, I looked in the filer > and it showed 788kb in port 2. Woohoo! Then I updated to the latest > ROM and still had 700kb in my port 2. Everything is working now. I had a similar problem last week. When I emailed HP they suggested taking the batteries out for 48 hrs because the calc might have crashed due to a stray static charge(??!!?). Anyway since I didn't have anything to lose I tried it. When I put the batteries back in nothing happened so I thought it was BS. But after about 3-minutes the screen came up with the usual try to recover message. Back in business, YAY. === Subject: Re: I revived my 49g+! posting-account=U2_nsg0AAACMehSjdUxQpFmS8kMQCYl9 Just one question, did you take out all the batteries including the backup battery? === Subject: Re: I revived my 49g+! > Just one question, did you take out all the batteries including the > backup battery? Yes. === Subject: Re: I revived my 49g+! posting-account=L8x4YQ0AAADtpf_klHFqoPGjVS6clfnr It may be dirt in the inside face of the keyboar. Not easy to clean, It just happen to me a few weeks ago, and after the recovering it is working fine. Daniel === Subject: Re: I revived my 49g+! > A few weeks ago I posted about a dead 49 that wouldn't respond to > anything. Well today just for kicks, I pulled it out and replaced all > the batteries. Nothing happened, RESET didn't work, and I figured it > was toast. Now the strange part. About 5 MINUTES later, I heard a > beep. I looked in the drawer where it went, and to my suprise I saw a > TTRM. > I answered no, and the calculator was working again. I still had the > two missing flash banks. I then loaded up the old 2MB rom 1.23 hoping > that would reset the flash banks. Sure enough, I looked in the filer > and it showed 788kb in port 2. Woohoo! Then I updated to the latest > ROM and still had 700kb in my port 2. Everything is working now. > So if it dies on you, don't give up I guess. =) > TW Have you called Ripley's yet?! Mine's started to freeze for no reason, other than perhaps too rapid a rate of key entry. Scott Chapin === Subject: Re: I revived my 49g+! posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Ghost in the machine... isn't it an x-file? Call the FBI. Maybe they can help you to solve the spooky problem you have on your calculator... A new proof that the first basic instinct developped by any intelligent behaviour is: survival ;) fred. === Subject: Codebar generation, library wizard, app translator :D posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J I do a few programs in my hp49g+, in pure UserRPL: 1. Codebar generator and codebar database, can draw the supermarket codebar directly in your calculator, supports EAN8 and EAN13 with auto generation of check digit. The main app is in english. Works with 131x64 screen and 131x80 too. Preview: http://avatarintheedge.iespana.es/hp49gp/codebar.gif Package: http://avatarintheedge.iespana.es/hp49gp/codebar.zip 2. Library wizard can create libraries with your programs using a step by step wizard. Uses personalized chooseboxes created in pure UserRPL (a bit slow) with source of choosebox 'SBOX'. The main app is in spanish. Works with 131x64 screen, but can run in 131x80 (uses a blank space). Preview: http://avatarintheedge.iespana.es/hp49gp/easylib.gif Package: http://avatarintheedge.iespana.es/hp49gp/easylib.zip 3. Application translator, can extract strings of vars and inject them again to any var that support ->STR and OBJ->. Status messages of app are in spanish. Package: http://avatarintheedge.iespana.es/hp49gp/apptrans.rar Example: 1: << hello world MSGBOX >> TEXTR 2: << hello world MSGBOX >> 1: {{hello world hello world}} (Modify the list: {{original translated} {...}}) 2: << hello world MSGBOX >> 1: {{hello world hola mundo}} TINJE Result: 1: << hola mundo MSGBOX >> === Subject: Re: Codebar generation, library wizard, app translator :D I have just tested your codebar generator in a supermarket, and it worked fine even though I have had fears in a first time that the calc's resolution wasn't good enough to do such a task. Great job, and thank's a lot. It can be very useful in different situations. Christophe LALUC >I do a few programs in my hp49g+, in pure UserRPL: > 1. Codebar generator and codebar database, can draw the supermarket > codebar directly in your calculator, supports EAN8 and EAN13 with auto > generation of check digit. The main app is in english. > Works with 131x64 screen and 131x80 too. > Preview: http://avatarintheedge.iespana.es/hp49gp/codebar.gif > Package: http://avatarintheedge.iespana.es/hp49gp/codebar.zip > 2. Library wizard can create libraries with your programs using a step > by step wizard. Uses personalized chooseboxes created in pure UserRPL > (a bit slow) with source of choosebox 'SBOX'. The main app is in > spanish. > Works with 131x64 screen, but can run in 131x80 (uses a blank space). > Preview: http://avatarintheedge.iespana.es/hp49gp/easylib.gif > Package: http://avatarintheedge.iespana.es/hp49gp/easylib.zip > 3. Application translator, can extract strings of vars and inject them > again to any var that support ->STR and OBJ->. Status messages of app > are in spanish. > Package: http://avatarintheedge.iespana.es/hp49gp/apptrans.rar > Example: > 1: << hello world MSGBOX >> > TEXTR > 2: << hello world MSGBOX >> > 1: {{hello world hello world}} > (Modify the list: {{original translated} {...}}) > 2: << hello world MSGBOX >> > 1: {{hello world hola mundo}} > TINJE > Result: > 1: << hola mundo MSGBOX >> === Subject: Re: Codebar generation, library wizard, app translator :D posting-account=lAaz6QwAAABR-WEJBEnTuXipZ8zCV0L0 I played with the codebar one but could not figure out how to correctly enter the numbers. And it removes the clock but should put it back if needed when quiting the program RCLF STOF. Looks like it will be quite good very soon. A good option would be to be able to send to infrared printer. Arnaud === Subject: Re: Codebar generation, library wizard, app translator :D posting-account=FjwNTA0AAACkoOZvbPG9Ru9bQtWL4f1J Arnaud ha escrito: > I played with the codebar one but could not figure out how to correctly > enter the numbers. And it removes the clock but should put it back if > needed when quiting the program RCLF STOF. > Looks like it will be quite good very soon. > A good option would be to be able to send to infrared printer. > Arnaud You can send the codebar with ON+UP, maybe setting IR transmission. === Subject: Memory Card posting-account=UZpa1g0AAAAGK-b1uuwdq2SDcb2fyNNB I have 2MB of ram on my 48GX. I see all of the port numbers. When I try to store info to one of these, I get a beep and error message. Do you know what may or what I'm doing so wrong! === Subject: Re: Memory Card PINIT > I have 2MB of ram on my 48GX. I see all of the port numbers. When I > try to store info to one of these, I get a beep and error message. Do > you know what may or what I'm doing so wrong! === Subject: 49g+ library issue Hello all, I am certain that someone has already asked this question, but I was unable to find it. I recently updated the ROM of my 49G+ to the new version (2.0) build 50. Everything seems fine with the exception of the equation library. When I check on the library menu (right shift, 2) I only see the ports 0, 1, and 2. when I check each port I see two libraries (226, and 227). I have tried to move them from port to port, turn off the calculator wait a min or so, then turn the calculator back on without sucess. Am I doing something wrong or am I not supposed to be doing that? By the way these two libraries came with the ROM upgrade. I have searched the web without success either to include the HP web site. Any ideas? Angel === Subject: Re: 49g+ library issue Press APPS, then down-arrow until you see the end of the list. Equation Library should be there (maybe in 12th position or so). If not, hold the ON button then press and release F3, release ON. Try the APPS button again. Once you see the Equation Library entry, select it and you will get a menu for the library above the function keys. -- - - - - - - - - - - - - - - - - Bill Graves RKBA! bgraves@ix.netcom.com > Hello all, > I am certain that someone has already asked this question, but I was > unable to find it. > I recently updated the ROM of my 49G+ to the new version (2.0) build > 50. Everything seems fine with the exception of the equation library. > When I check on the library menu (right shift, 2) I only see the ports > 0, 1, and 2. when I check each port I see two libraries (226, and > 227). I have tried to move them from port to port, turn off the > calculator wait a min or so, then turn the calculator back on without > sucess. Am I doing something wrong or am I not supposed to be doing > that? By the way these two libraries came with the ROM upgrade. I > have searched the web without success either to include the HP web > site. > Any ideas? > Angel === Subject: Re: 49g+ library issue >Press APPS, then down-arrow until you see the end of the list. Equation >Library should be there (maybe in 12th position or so). >If not, hold the ON button then press and release F3, release ON. Try the >APPS button again. >Once you see the Equation Library entry, select it and you will get a menu >for the library above the function keys. In addition, you can type 114 MENU to access the equation library menu, or EQNLIB to go directly to the equation library. Bob === Subject: Re: 49g+ library issue <25d066b0220f11f85e1d5b304af439ee@localhost.talkaboutcomputing.com> posting-account=n6emfQ0AAAAdk-bm_28xT-EPOBKN7Tkw I was expecting the library to be in the right_shift+LIB. Now that I think about it since the library was integrated it is where it supposed to be. Therefore the APPS button worked!!! Angel === Subject: Re: 49g+ library issue > Hello all, > I am certain that someone has already asked this question, but I was > unable to find it. > I recently updated the ROM of my 49G+ to the new version (2.0) build > 50. Everything seems fine with the exception of the equation library. > When I check on the library menu (right shift, 2) I only see the ports > 0, 1, and 2. when I check each port I see two libraries (226, and > 227). I have tried to move them from port to port, turn off the > calculator wait a min or so, then turn the calculator back on without > sucess. Am I doing something wrong or am I not supposed to be doing > that? By the way these two libraries came with the ROM upgrade. I > have searched the web without success either to include the HP web > site. Libraries 226 and 227 are for the equation library that comes with ROM revision 2.00 and they're normally placed in port 2 (flash memory). I don't know for certain that they'd work in port 0 or 1, but why would anyone want to move them? Still, if one did move them, turning the calculator off and then on wouldn't help; you'd need to do a warmstart (hold down ON while pressing C) and possibly ATTACH them as well. > Any ideas? What is the question? What problem do you see? You're welcome in advance. -- James === Subject: Re: 49g+ library issue posting-account=n6emfQ0AAAAdk-bm_28xT-EPOBKN7Tkw Well so far I have not been able to use the libraries and you mentioned that perhaps I need to ATTACH them. How do I do that? Angel === Subject: hp49g+ -- Problem with bad libs in second port posting-account=7JSjhQwAAACHPoWhnBsy0SqBF4HYgr3E I was experimenting with a UserRPL program for automatic construction of libraries, unfortunately with a bug: it has assigned the lib number 1 to custom library of mine and has copied it to the second (flash) port. Actually, there are two copies of the same library there. The problem is that I cannot delete any of both libraries. I tried a hard reset and 2 PINIT without any success. I am a bit afraid that the calc is not going to work properly any longer because the assigned lib number (1). Alexis === Subject: Re: hp49g+ -- Problem with bad libs in second port > I was experimenting with a UserRPL program for automatic construction > of libraries, unfortunately with a bug: it has assigned the lib number > 1 to custom library of mine and has copied it to the second (flash) > port. Actually, there are two copies of the same library there. > The problem is that I cannot delete any of both libraries. I tried a > hard reset and 2 PINIT without any success. I am a bit afraid that > the calc is not going to work properly any longer because the assigned > lib number (1). FILER === Subject: Re: hp49g+ -- Problem with bad libs in second port <7Q1Ee.7084$UG3.939@reader1.news.jippii.net> posting-account=7JSjhQwAAACHPoWhnBsy0SqBF4HYgr3E yes, of course I tried already with FILER, but it does not remove the libraries, even though I do not get any warning or error message, the libs just remain there. Any other hints? Alexis I was experimenting with a UserRPL program for automatic construction of libraries, unfortunately with a bug: it has assigned the lib number 1 to custom library of mine and has copied it to the second (flash) port. Actually, there are two copies of the same library there. The problem is that I cannot delete any of both libraries. I tried a hard reset and 2 PINIT without any success. I am a bit afraid that the calc is not going to work properly any longer because the assigned lib number (1). > FILER === Subject: Re: hp49g+ -- Problem with bad libs in second port > yes, of course I tried already with FILER, but it does not remove the > libraries, even though I do not get any warning or error message, the > libs just remain there. > Any other hints? Keep backspace pressed when system boots (be quick) and no extra libraries will be attached then use FILER === Subject: Re: hp49g+ -- Problem with bad libs in second port <7Q1Ee.7084$UG3.939@reader1.news.jippii.net> posting-account=7JSjhQwAAACHPoWhnBsy0SqBF4HYgr3E I managed to avoid libs being attached while rebooting (I checked the LIB menu), but still I cannot (re)move the two libs with ID 1 from the second port. 1 DETACH gives the error Bad argument value. I guess the system prevents playing around with low lib IDs, since they seem to be system libs. There should a way to get rid of those libs. I could not believe that a simple UserRPL program causes irreversible damages to the calc (CRLIB should do the necessary checks). 2 PINIT just returns the value 2 to stack after a while, but the port is not initialized. In fact PINIT seems not to work at all, for I it behaves the same way. Any more Hints? Alexis === Subject: Re: hp49g+ -- Problem with bad libs in second port > There should a way to get rid of those libs. I could not believe that a > simple UserRPL program causes irreversible damages to the calc (CRLIB > should do the necessary checks). 2 PINIT just returns the value 2 to > stack after a while, but the port is not initialized. In fact PINIT > seems not to work at all, for I it behaves the same way. PINIT will only fix port with objects having an invalid CRC, the structure of your port 2 isn't corrupted so PINIT has nothing to do there. Jean-Yves === Subject: Re: hp49g+ -- Problem with bad libs in second port <7Q1Ee.7084$UG3.939@reader1.news.jippii.net> <3kc4n0Fl6kvvU2@individual.net> posting-account=7JSjhQwAAACHPoWhnBsy0SqBF4HYgr3E some port problems, but I did not know exactly which, that is why I reported that PINIT was not working in my case. I did what you adviced ( 1 4046 + 1 ->TAG DUP DETACH PURGE ), but unfortunately it has not worked. It did remove the contents of the second port EXCEPT for the two libs I want to get rid of. The contents of port 2 according to FILER are now: [icon] TEST L1 953 [icon] TEST L1 953 I am getting a bit scared. It seems to be more difficult than I thought. But I still hope someone had the same or a similar problem before. Or perharps is my calc defective? Alexis === Subject: Re: hp49g+ -- Problem with bad libs in second port > some port problems, but I did not know exactly which, that is why I > reported that PINIT was not working in my case. > I did what you adviced ( 1 4046 + 1 ->TAG DUP DETACH PURGE ), but Sorry it should be: 1 4096 + 2 ->TAG DUP DETACH PURGE In fact this should work for any library with a number < 256: [lib_number] 4096 + [port_number] ->TAG DUP DETACH PURGE JY === Subject: Re: hp49g+ -- Problem with bad libs in second port <7Q1Ee.7084$UG3.939@reader1.news.jippii.net> <3kc4n0Fl6kvvU2@individual.net> <3kdfs1FtqetrU1@individual.net> posting-account=7JSjhQwAAACHPoWhnBsy0SqBF4HYgr3E keep this small program for the future. Alexis some port problems, but I did not know exactly which, that is why I reported that PINIT was not working in my case. I did what you adviced ( 1 4046 + 1 ->TAG DUP DETACH PURGE ), but > Sorry it should be: > 1 4096 + 2 ->TAG DUP DETACH PURGE > In fact this should work for any library with a number < 256: > [lib_number] 4096 + [port_number] ->TAG DUP DETACH PURGE > JY === Subject: Re: hp49g+ -- Problem with bad libs in second port >some port problems, but I did not know exactly which, that is why I >reported that PINIT was not working in my case. >I did what you adviced ( 1 4046 + 1 ->TAG DUP DETACH PURGE ), but >unfortunately it has not worked. It did remove the contents of the >second port EXCEPT for the two libs I want to get rid of. >The contents of port 2 according to FILER are now: >[icon] TEST L1 953 >[icon] TEST L1 953 >I am getting a bit scared. It seems to be more difficult than I >thought. >But I still hope someone had the same or a similar problem before. Or >perharps is my calc defective? >Alexis It seems very strange that the calc isn't letting you delete them. That happened to me once, but it was because of a bad CRC and PINIT fixed it. The only way to delete the libraries may be to completely wipe out the flash banks that are not used by the OS. Someone here probably knows how to do that. If not, you could contact Christoph Giesselink and ask him to give you the program described in Flash ROM Information (see http://www.hpcalc.org/details.php?id=3019 ). Bob === Subject: Re: hp49g+ -- Problem with bad libs in second port <8907b51a354f879b00d6837ca1448973@localhost.talkaboutcomputing.com> posting-account=7JSjhQwAAACHPoWhnBsy0SqBF4HYgr3E already given me a solution (see his post in this thread). Alexis === Subject: Re: hp49g+ -- Problem with bad libs in second port > I managed to avoid libs being attached while rebooting (I checked the > LIB menu), but still I cannot (re)move the two libs with ID 1 from the > second port. 1 DETACH gives the error Bad argument value. I guess Do: 1 4046 + 1 ->TAG DUP DETACH PURGE Should take care of it. JY === Subject: Re: hp49g+ -- Problem with bad libs in second port <7Q1Ee.7084$UG3.939@reader1.news.jippii.net> posting-account=7JSjhQwAAACHPoWhnBsy0SqBF4HYgr3E One fact I forgot and that might be important: both bad libs in the second port do have both the same library ID (1) and the same title (TEST). Alexis === Subject: HP33s First impressions posting-account=voIivw0AAACoCSOxk25Vk_c9Urhrm8f4 After I have sent my 49G+ (the 6th one) back to HP, I bought myself a 33S. And what should I say? A great piece. Good keys (Kinpo CAN do it better). A solid body (brushed aluminium) A good solver RPN of course I think will love it, and my replaced 49G+ will be put in the drawer... Christian === Subject: Re: HP33s First impressions posting-account=IGUX2Q0AAAAhLOj1vOfIVSpMCs089TPn If the display radix . and , are fixed where did you buy it? Tom > After I have sent my 49G+ (the 6th one) back to HP, I bought myself a > 33S. > And what should I say? > A great piece. Good keys (Kinpo CAN do it better). > A solid body (brushed aluminium) > A good solver > RPN of course > I think will love it, and my replaced 49G+ will be put in the drawer... > Christian === Subject: Re: HP33s First impressions posting-account=voIivw0AAACoCSOxk25Vk_c9Urhrm8f4 It doesn't seem to fixed. Although I prefer the comma as decimal point. The point is really small... The s/n is CNA417xxxxx, I bought it in Germany, near Berlin. Tomorrow I will take some pictures of it... Christian > If the display radix . and , are fixed where did you buy it? > Tom After I have sent my 49G+ (the 6th one) back to HP, I bought myself a 33S. And what should I say? A great piece. Good keys (Kinpo CAN do it better). A solid body (brushed aluminium) A good solver RPN of course I think will love it, and my replaced 49G+ will be put in the drawer... Christian === Subject: Re: HP33s First impressions posting-account=cEKa6Q0AAAC3YlI6nbEywfIpggMw2WOF Would you please compare and contrast the solvers between the two units? I have heard that the 33s has a very nice solver (implying it is better than the 49G+), and am wondering how it works. === Subject: Re: HP33s First impressions posting-account=voIivw0AAACoCSOxk25Vk_c9Urhrm8f4 > Would you please compare and contrast the solvers between the two > units? I have heard that the 33s has a very nice solver (implying it > is better than the 49G+), and am wondering how it works. Its very simple. In the 33s you can store all youre equations in a list (=EQN List). First you got to choose the eqaution from the list. Then press the solve button. The calc ascs you for the variable you want to solve. All the other values are being asked. After you typed them all in (old values arr safed, just confirm with key R/S if they didnt change), the solver tries to solve. I he is not able he iterates. Compared to 49G+: - Much slower (less than 1s 49G+; about 5s the 33s), depends on the EQN - only a two line display (one line for the var name, one for the value) - var names may only have one character (A...Z), so A1, var1 is not possible See more in the 33s manual... Christian === Subject: Re: HP33s First impressions I would like to add to that that 49g+ can in addition use units in a numerical solver 49g+ has in addition to the above addition also symbolic solver 49g+ has also a differential solver and limits and symbolic integrator and symbolic derivative and so on... It really adds a whole lot more compared to the 33s The simple 33s solver might be easier at first, though. VPN >> Would you please compare and contrast the solvers between the two >> units? I have heard that the 33s has a very nice solver (implying it >> is better than the 49G+), and am wondering how it works. > Its very simple. In the 33s you can store all youre equations in a list > (=EQN List). First you got to choose the eqaution from the list. Then > press the solve button. The calc ascs you for the variable you want to > solve. All the other values are being asked. > After you typed them all in (old values arr safed, just confirm with > key R/S if they didnt change), the solver tries to solve. I he is not > able he iterates. > Compared to 49G+: > - Much slower (less than 1s 49G+; about 5s the 33s), depends on the EQN > - only a two line display (one line for the var name, one for the > value) > - var names may only have one character (A...Z), so A1, var1 is not > possible > See more in the 33s manual... > Christian === Subject: Re: HP33s First impressions I grant you the 49+ does a lot more, but at least the 33 isn't in transit half of its life for problems. Just a thought. Don On Sat, 23 Jul 2005 12:37:24 +0300, Veli-Pekka Nousiainen >I would like to add to that that >49g+ can in addition use units in a numerical solver >49g+ has in addition to the above addition also symbolic solver >49g+ has also a differential solver and limits and symbolic integrator >and symbolic derivative and so on... >It really adds a whole lot more compared to the 33s >The simple 33s solver might be easier at first, though. >VPN > Would you please compare and contrast the solvers between the two > units? I have heard that the 33s has a very nice solver (implying it > is better than the 49G+), and am wondering how it works. >> Its very simple. In the 33s you can store all youre equations in a list >> (=EQN List). First you got to choose the eqaution from the list. Then >> press the solve button. The calc ascs you for the variable you want to >> solve. All the other values are being asked. >> After you typed them all in (old values arr safed, just confirm with >> key R/S if they didnt change), the solver tries to solve. I he is not >> able he iterates. >> Compared to 49G+: >> - Much slower (less than 1s 49G+; about 5s the 33s), depends on the EQN >> - only a two line display (one line for the var name, one for the >> value) >> - var names may only have one character (A...Z), so A1, var1 is not >> possible >> See more in the 33s manual... >> Christian === Subject: Re: HP33s First impressions > 49g+ can in addition use units in a numerical solver > 49g+ has in addition to the above addition also symbolic solver > 49g+ has also a differential solver and limits and symbolic integrator > and symbolic derivative and so on... > It really adds a whole lot more compared to the 33s > The simple 33s solver might be easier at first, though. > VPN Not usually particularly relevant if you're looking for a specific answer. If I'm problems of any consequence I'm not generally using limits, or symbolic(s). If you were looking for symbolics you can get an HP49 or better yet use dedicated math software if a laptop is available. === Subject: Re: HP33s First impressions posting-account=boybFQ0AAACcJkbp4Mpsq2-jhOTZrn5e 33s is the ugliest thing I've ever seen. who designed that weird thing? I don't care what it can do, I'd better do it by hand. this is no HP as I know it reth === Subject: Re: HP33s First impressions posting-account=Ab4Erg0AAADnPJuZhQ3v71vPuxtJ--tX Did HP fix the radix? If yes, what's the serial #? Gil === Subject: Re: HP33s First impressions I don't understand how you can prefer a HP 33S to an HP49G+. That 's surprising. Programming is much more confortable with RPL than with RPN, capabilities of the HP 49G+ is really huge when the HP33S is smaller.... HP 33S is fantastic but impossible to compare with the HP 48 or HP 49... To get a very efficient solver (The best one!) please try the HP 17BII or the HP19BII, I don't like the HP 33S or HP32S solvers: - values are reduce to letter A..Z, - you should remember how to operate: how to access the list and how to choose the var you want to solve. Good luke jajacom > Did HP fix the radix? If yes, what's the serial #? > Gil === Subject: Re: HP33s First impressions And now if they could only make it *look* like a calculator instead of something from the 24th century :( CC > After I have sent my 49G+ (the 6th one) back to HP, I bought myself a > 33S. > And what should I say? > A great piece. Good keys (Kinpo CAN do it better). > A solid body (brushed aluminium) > A good solver > RPN of course > I think will love it, and my replaced 49G+ will be put in the drawer... > Christian === Subject: RPN Emu with XYZT ?? posting-account=voIivw0AAACoCSOxk25Vk_c9Urhrm8f4 Does anyone know a small RPN Emu with XYZT Stack. I want to see all registers of the stack. I just want the register to be visible, to see the behaviour of a classical RPN stack. A lot of my collegues have never heard about RPN and I want to teach them. Christian === Subject: Re: RPN Emu with XYZT ?? posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo > Does anyone know a small RPN Emu with XYZT Stack. I want to see all > registers of the stack. > I just want the register to be visible, to see the behaviour of a > classical RPN stack. > A lot of my collegues have never heard about RPN and I want to teach > them. > Christian __________________________________________________ Pls inform them that the stack is a low level thing . It is only used when working with Forth primatives . The next level up uses named constants and variables and not the stack . === Subject: Re: RPN Emu with XYZT ?? > Does anyone know a small RPN Emu with XYZT Stack. > I want to see all registers of the stack. > I just want the register to be visible, > to see the behaviour of a classical RPN stack. > A lot of my collegues have never heard about RPN > and I want to teach them. If you mean to emulate a classic 4-level HP RPN stack on any HP48 or HP49, then see: [r->][OFF] === Subject: Re: RPN Emu with XYZT ?? >> Does anyone know a small RPN Emu with XYZT Stack. >> I want to see all registers of the stack. >> I just want the register to be visible, >> to see the behaviour of a classical RPN stack. >> A lot of my collegues have never heard about RPN >> and I want to teach them. > If you mean to emulate a classic 4-level HP RPN stack > on any HP48 or HP49, then see: And try use Free42 if demostrate 'classic' 4-level HP RPN stack calculator without signing and register on HP:s teatcher corner to allow download software emulator... (idiotic philosofy from hp-side IMHO...) Free42 is easy enough to demostrate 'easy' calculate whithout digging into menues and advanced functions - and working seamless and automatic with complex number. and working now under Windows, linux and palm-T. --- tip. Is possible to make complex storing register with moment: First, setting rectangular mode: '2nd', '+/-' (modes), select 'rect', go back with 'exit' after this: '0' 'enter' '0' 'complex' 'sto' '+' '00' and after this is possibly to store complex numbers in register 00-25 for later calculate. IMHO miss from HP not make this as default... /TE === Subject: Re: RPN Emu with XYZT ?? format=flowed; reply-type=original ChristianB. escreveu na mensagem > Does anyone know a small RPN Emu with XYZT Stack. I want to see all > registers of the stack. > I just want the register to be visible, to see the behaviour of a > classical RPN stack. Emu48 (Windows) or X48 (Linux) emulating an HP-48 or HP-49 on a PC? Or XCALC (despite the name, which is similar to Xcalc on Linux, it is for Windows): http://www.tordivel.no/xcalc/ []s === Subject: Re: RPN Emu with XYZT ?? The xcalc was definitely the right one... THX !!!! Wishmaster schrieb im Newsbeitrag > ChristianB. escreveu na mensagem >> Does anyone know a small RPN Emu with XYZT Stack. I want to see all >> registers of the stack. >> I just want the register to be visible, to see the behaviour of a >> classical RPN stack. > Emu48 (Windows) or X48 (Linux) emulating an HP-48 or HP-49 on a PC? > Or XCALC (despite the name, which is similar to Xcalc on Linux, it is for > Windows): http://www.tordivel.no/xcalc/ > []s === Subject: Re: RPN Emu with XYZT ?? posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Hi Christian, Take a look at this site: http://nonpareil.brouhaha.com/ maybe it can help you. Fred. === Subject: opening programs posting-account=nkXj9A0AAAAho_VEJ5VEe1NMx9Jt08I1 After downloading new programs from the internet my 49G+ will not run the programs even though they show that they are in the calculator. I made a new folder for them and inserted them into this folder but can not open them to run on the 49G === Subject: Re: opening programs posting-account=nkXj9A0AAAAho_VEJ5VEe1NMx9Jt08I1 === Subject: Re: opening programs posting-account=7JSjhQwAAACHPoWhnBsy0SqBF4HYgr3E Are you sure that they are not libraries? If so, copy or move them to some port (a good choice is the second one, flash) and not to the home directory. You may prefer to do it with the FILER (left_shift+APPS), it is easier. Then reboot: ON+F3. The new library should have been attached. You chan check it visiting the LIB menu (right_shift+2). Alexis === Subject: test test === Subject: To HPGCC team: a bug has been isolated posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE Hello people, I think I have isolated a bug in HPGCC. This happen when I statically instanciate a new structure in my program. Some initializations are not executed correctly, and when I comment the declaration of the instatiation, the initializations are donne correctly. It is a similar thing I already mentionned a few weeks ago. I suppose it is an alignment problem. Send me a mail then I'll send you the project where this problem occurs. Fred. === Subject: Re: To HPGCC team: a bug has been isolated > I think I have isolated a bug in HPGCC. <...> > Send me a mail then I'll send you the project where this problem > occurs. If you think you found a bug, post a precise description including a meaningful code fragment, that allows a reproduction, to the project's bug tracker: http://sourceforge.net/tracker/?func=add&group_id=115846&atid=672861 ibl === Subject: Re: To HPGCC team: a bug has been isolated <42e18d00_1@news.arcor-ip.de> posting-account=tJEQEw0AAAAD1Rviuyd5QqxLNODse0hE it has been sent. fred. === Subject: hpgcc is for games posting-account=O6knhQ0AAAD8zE21TyMVwtCasYnRdHOm After reading many things ( ok so im slow ) i think i have a bigger understanding of the whole of C programs for hp49g+. These programs will be aimed basically to build games or other softwares that need great control of the hardware. So the fact that there is a calculator software ( CAS or whatever that is ) is basically ignored since one cant interact very well with it ( only maybe using the hpgcc Stack library for example ). If one needs to write a program that makes use of, for example, the built-in ROOT command, that can't be done in C ( right ? ). So im writting this to get comments and , if i got it right it, just to make it clear to other newbies that it would be great to have a simple C program to call an API function that would return the integral of an equation or whatever, but, thats simply not the case. === Subject: Re: hpgcc is for games posting-account=FCAFHAwAAADaPCjuTWXm3ER18SIzldtO > After reading many things ( ok so im slow ) i think i have a bigger > understanding of the whole of C programs for hp49g+. These programs > will be aimed basically to build games or other softwares that need > great control of the hardware. So the fact that there is a calculator > software ( CAS or whatever that is ) is basically ignored since one > cant interact very well with it ( only maybe using the hpgcc Stack > library for example ). C programs don't have access to the CAS, that's a fact (at least for now). But you seem to be ignoring the fact that C programs have access to all C mathematical functions with double precision (which is more accurate than calculator reals), and at speeds that can't be compared to the Saturn processor. This makes it very well suited for numeric applications (numeric solvers, large matrices, FFT, and any other numerical method you can think of). If you want to integrate a symbolic, then all you need is to press the right keys, why would you write a C program to do what the calculator already does and very well? Unless you need to integrate 5000 functions per second, there is no point in doing it at ARM speed. > If one needs to write a program that makes use of, for example, the > built-in ROOT command, that can't be done in C ( right ? ). > So im writting this to get comments and , if i got it right it, just to > make it clear to other newbies that it would be great to have a > simple C program to call an API function that would return the > integral of an equation or whatever, but, thats simply not the case. I don't understand what's so great in having a simple program that integrates a function written in C. Using the CAS from C would elimintate the speed advantage, since the emulator would handle the hard work, so why would you do that in C? Each language has it's specific scope and application, you should choose the language for your programs according to the problem you need to solve and not the other way (same way you wouldn't use a hammer to drive screws, or a screwdriver with nails). * UserRPL: For simple, repetitive, everyday tasks. Can be written quickly but it's execution may be slow. * SysRPL: For more complex tasks, which may require a user interface (inform box, etc.) and are more critical in speed. * Saturn assembler: Speed critical routines that need to use calculator resources. Can be replaced with C in many cases. * ARM assembler: Small and simple routines that require access to ARM resources. Can be replaced with C in most cases. * C: Medium to complex applications with intensive access to hardware resources and speed critical routines. C came to cover a gap in the programming capabilities of the new calcs in areas like grayscale, SD card access, etc. where other languages were running short. So, my opinion is that you should choose the proper tool for each problem, as opposed of trying to find a universal tool to solve everything. Claudio === Subject: Re: hpgcc is for games -=[ Sun, 24.7.05 09:40 a.m. +1200 (NZT) ]=- Hello Claudio Lapilli, > * C: Medium to complex applications with intensive access > to hardware resources and speed critical routines. A generalised sudoku solver defintely needs speed and at http://www.users.waitrose.com/~nihilist/sudoku.html there is C source for an excellent solver... and this compiles on the 16 bit 200LX using PowerC (by MIX)... and runs - even the most diabolical example (16 levels of recursion) takes only 10 minutes. I wonder if someone can compile this in hpgcc? It takes an ascii disk file as input - maybe change this to be a matrix from the stack? Tony === Subject: Re: hpgcc is for games > A generalised sudoku solver defintely needs speed and at > http://www.users.waitrose.com/~nihilist/sudoku.html > there is C source for an excellent solver... and this compiles > on the 16 bit 200LX using PowerC (by MIX)... and runs - even > the most diabolical example (16 levels of recursion) takes > only 10 minutes. Maybe you could try the file at httrp://alpage.ath.cx/sudoku.zip ? Currently it takes a raw text file, SUDOKU.TXT from the SD card. Copy sudoku.txt to the SD card, and run SUD to see the solution. The program runs at 12MHz to save batteries but I can speed it up if its too slow. By the way, ten minutes is how long it took to port it to the HP :-) Al === Subject: Re: hpgcc is for games -=[ Sun, 31.7.05 3:38 p.m. +1200 (NZT) ]=- Hello Al Borowski, 6 days 02h48m ago, on Mon, in message ID <42e434fa$0$29413$5a62ac22@per-qv1-newsreader-01.iinet.net.au> : A generalised sudoku solver defintely needs speed and at http://www.users.waitrose.com/~nihilist/sudoku.html [...] > Maybe you could try the file at http://alpage.ath.cx/sudoku.zip ? [...] http://homepages.paradise.net.nz/th/sud49gp.zip has Nihilist/Al's Sudoku with a stack interface. SD card required. -- Tony #76 All true wisdom is discovered in tag lines. === Subject: Re: hpgcc is for games -=[ Sat, 30.7.05 2:38 p.m. +1200 (NZT) ]=- Hello Al Borowski, 5 days 01h48m ago, on Mon, in message ID <42e434fa$0$29413$5a62ac22@per-qv1-newsreader-01.iinet.net.au> : A generalised sudoku solver defintely needs speed and at http://www.users.waitrose.com/~nihilist/sudoku.html there is C source for an excellent solver... and this compiles on the 16 bit 200LX using PowerC (by MIX)... and runs - even the most diabolical example (16 levels of recursion) takes only 10 minutes. > Maybe you could try the file at httrp://alpage.ath.cx/sudoku.zip ? > Currently it takes a raw text file, SUDOKU.TXT from the SD card. Copy > sudoku.txt to the SD card, and run SUD to see the solution. The program > runs at 12MHz to save batteries but I can speed it up if its too slow. > By the way, ten minutes is how long it took to port it to the HP :-) Al, I just now managed to buy an SD card and reader. The long and the short is - the performance of your SUD is phenomenal!!! It appears to solve sudoku.txt *instantly*. The diabolical sample9.txt takes only 7 seconds. I'm new to SD cards and got the cheapest - a local Dick Smith brand. It worked first up in the calculator but got stuck due to the label being munged inside the calc - probably caught when I tried to eject it. The label did protrude - the card face lacks an indent surround. Not to worry it works great now with no label, and most of the glue removed :-) Now, I must see if I can make a little User RPL interface to put puzzles on the SD card... hmm not today as the 49g+ insists on adding the HP49 string header - and SUD thinks the 49 is a bad egg as it exceeds 9 :-) -- Tony === Subject: Re: hpgcc is for games -=[ Sat, 30.7.05 8:01 p.m. +1200 (NZT) ]=- Hello Al, Me again - I recompiled sudoku to make it ignore the HP49.. header in a sudoku.txt saved from the 49g+. I run it using PrRUN from Lib275. Your executable didn't even need Lib275 - how did you do that? digitString[i]=0 rather than the [i+1], and after the use of betrays the RPL origin: if ((xx=49) && (ii == 0) && (jj == 0)) { } The speed is really sensational. Two funny things - I couldn't get Lib275 to attach on my 49g+, but it works fine on my old 49g+ which had been sitting idle for about 6 months. When I took it out of the case the bottom 2 rubber feet were almost detached and their glue had run over the calc - haven't seen active glue before. Maybe the calc only needs 2 feet :-) -------- in message ID <11023632ROBOTLX@news.clear.net.nz> : > Now, I must see if I can make a little User RPL interface to > put puzzles on the SD card... hmm not today as the 49g+ > insists on adding the HP49 string header - and SUD thinks > the 49 is a bad egg as it exceeds 9 :-) -- Tony === Subject: Re: hpgcc is for games