HP-236 Subject: Re: Installing Libraries on the 49G+ >> libraries he always wants attached in the STARTUP variable, === Subject: HP41C manual? I found a what I thought was a cheap 41C (25 dollars), so I bought it. I never had an ancient machine like this one, and I'm looking for the manual in PDF format. Does anyone have it? Can you send it to me?. Not to this address, so if you have it please contact me. -- === Subject: Re: HP41C manual? says... > I found a what I thought was a cheap 41C (25 dollars), so I bought it. I > never had an ancient machine like this one, ANCIENT??? If HP reintroduced it today *exactly* as it was made originally, but perhaps with something other than N sized batteries, it would probably outsell every other calculator on the market. -- Randy Howard (2reply remove FOOBAR) The most amazing achievement of the computer software industry is its continuing cancellation of the steady and staggering gains made by the computer hardware industry... - Henry Petroski === Subject: Graphing Function in Loop while Plot The program MBP is supposed to show a series of logistic function bifurcation plots each using more iterations of the function. However, as the ITERS variable gets to 10 it has a very long pause before completing the graph and I've never seen it actually graph at 11--requires a reboot. I don't see anything here that should be a problem and I've written a few other programs trying to get the same result without everything here, but I can't find any specific sub-part of this that causes the problem. Am I mistaken that the calculator should have no problem running this program? -jgs %%HP: T(3)A(R)F(.); DIR MBP << 5. 30. FOR n n 'ITERS' STO ERASE PICT { # 0d # 0d } n 1. ->GROB REPL DRAW NEXT >> BIFFUNCT << -> Gl << .5 1. ITERS START 1. OVER - * 4. * Gl * NEXT 1. ITERS START 1. OVER - * 4. * Gl * Gl OVER R->C PIXON NEXT >> >> ITERS 10. EQ 'BIFFUNCT(X)' PPAR { (0.,0.) (1.,1.) X # 1d (0.,0.) FUNCTION Y } END === Subject: Re: USB to RS-232 for 49g+ Why bother with expensive, difficult and painfully programmable atmels, just take a PIC if you're smart you'll program everything in a 16f84, but in anycase there's a pic16c74 (or analogue) that has a serial comm. The programmers for theese are either homebuilt or cost 15$ (the pic costs 5$) If something ask me and i'll be happyt to help you with pic programming www.microchip.com > I've been thinking about getting an RS-232 port on the 49g+ ... I think > I can probably build a device for about $20 (estimated ... the most I > can see it is $50) or so that can do this. How many people would be > interested? > > I'll probably be using an Atmel AT43USB370 controller on a cus board. > I will have to figure out how the HP USB-serial protocol works :) I > can go looking through the Generic USB Serial code in the Linux kernel; > I can prove that works. > > Just an idea ... if I can get more than 5 people interested in buying > said item, I'll probably build it and sell it. > > (Please tell me if this has already been done.) > > -MrM > Well, folks, it seems that the AT43USB370 won't do what I need -- it > won't speak RS-232. So I'm going to have to use an AT91RM9200 (also > Atmel), which costs a whopping US$27! So it's gonna end up closer to > the $50 end of that range :) > -MrM === Subject: Re: USB to RS-232 for 49g+ > Why bother with expensive, difficult and painfully programmable > atmels, just take a PIC Wrong chip for the job - too slow to implement a USB host in software. Even a PIC18f4320 @ 40MHz is only 10 MIPs. > if you're smart you'll program everything in a 16f84 Why? The 16F84 is horribly obsolete. Its only good as a raw beginners chip because it is so simple. For a little more then same price you can get a PIC18xxx series PIC, which is much nicer to program. They include things like ADC, PWM, more timers, 3 file select registers, etc etc. Banking/paging is no longer an issue, and they run @ 40MHz. , but in anycase > there's a pic16c74 (or analogue) that has a serial comm. My favourite PIC at the moment is a PIC18f4320, but that isn't suitable for this project. cheers, Al > The programmers for theese are either homebuilt or cost 15$ > (the pic costs 5$) > If something ask me and i'll be happyt to help you with pic > programming > www.microchip.com === Subject: Re: USB to RS-232 for 49g+ > Why bother with expensive, difficult and painfully programmable > atmels, just take a PIC > if you're smart you'll program everything in a 16f84, but in anycase > there's a pic16c74 (or analogue) that has a serial comm. > The programmers for theese are either homebuilt or cost 15$ > (the pic costs 5$) > If something ask me and i'll be happyt to help you with pic > programming > www.microchip.com >I've been thinking about getting an RS-232 port on the 49g+ ... I think >I can probably build a device for about $20 (estimated ... the most I >can see it is $50) or so that can do this. How many people would be >interested? >I'll probably be using an Atmel AT43USB370 controller on a cus board. > I will have to figure out how the HP USB-serial protocol works :) I >can go looking through the Generic USB Serial code in the Linux kernel; >I can prove that works. >Just an idea ... if I can get more than 5 people interested in buying >said item, I'll probably build it and sell it. >(Please tell me if this has already been done.) >-MrM >>Well, folks, it seems that the AT43USB370 won't do what I need -- it >>won't speak RS-232. So I'm going to have to use an AT91RM9200 (also >>Atmel), which costs a whopping US$27! So it's gonna end up closer to >>the $50 end of that range :) >>-MrM The problem there: there's no PICs that act as a USB _host_. And at that, the only USB-device PICs are UV-erase and OTP. However, the chip I decided on only speaks RS-232 at 8 data bits, no parity, and one stop bit... so I'll end up using a cheap PIC with 2 UARTs to repeat this data in different formats :) So I'm probably going to end up with 2 processors on one board ... this is going to be real fun! -MrM === Subject: Re: USB to RS-232 for 49g+ X-Eric-Conspiracy: There is no conspiracy. > So I'm going to have to use an AT91RM9200 (also > Atmel), which costs a whopping US$27! So it's gonna end up closer to > the $50 end of that range :) === Subject: Re: USB to RS-232 for 49g+ >>So I'm going to have to use an AT91RM9200 (also >>Atmel), which costs a whopping US$27! So it's gonna end up closer to >>the $50 end of that range :) Yeah ... I know :) I remember now why I didn't see the Cypress bunch a while earlier ... their site was down when I looked! -MrM === Subject: Re: USB to RS-232 for 49g+ X-Eric-Conspiracy: There is no conspiracy. > think I can probably build a device for about $20 (estimated ... the > most I can see it is $50) I'd very much expect closer to $50 than $20. > I'll probably be using an Atmel AT43USB370 controller on a cus > board. I will have to figure out how the HP USB-serial protocol works Unless I'm mistaken, that requires an additional processor. Why not use a Cypress CY7C67300, and run the code on the built-in processor? Then you just need to add a serial EEPROM for it to load your firmware I've been thinking about doing this, but haven't found the time. === Subject: Re: USB to RS-232 for 49g+ >>think I can probably build a device for about $20 (estimated ... the >>most I can see it is $50) > I'd very much expect closer to $50 than $20. >>I'll probably be using an Atmel AT43USB370 controller on a cus >>board. I will have to figure out how the HP USB-serial protocol works > Unless I'm mistaken, that requires an additional processor. Why not > use a Cypress CY7C67300, and run the code on the built-in processor? > Then you just need to add a serial EEPROM for it to load your firmware > I've been thinking about doing this, but haven't found the time. Yep ... I found that less than a few hours after I suggested the Atmel two :) -MrM === Subject: Re: Debug4x 2.1 BUG recompiling suprom49.a file? > ASSEMBLE > ... > name EQU #12345 Remove the # Jean-Yves === Subject: Re: The Hydrix Calc/PDA > I just wonder if this > Scientific Digital Assistant > or > Scientific Graphing Assistant > Sigma-Gamma-Alpha (Sum, Product, First [letter]) > Are you talking about this logo?: > http://www.asturlinux.org/~jsmanrique/sda/ Nice logo Why I think now that SGA would be better? Seventh Day Adventists=SDA > will have a superior tactile keyboard ala 41C series > It would be nice! > Also I think that the Flash must be huge, since it has ~8MB embedded > Linux > and Parisse's Xiac ~2MB and other tools... > I think that Giac (the readline) version is more than 3.5MB, and much more > in its graphical aproach. Yes, I hit the ENTER too soon, must be the first time (-; > and room for updates and room for userFlash 16MB min I think... > I think that 32MB should be the min. Yes, the graphical interface to Parisse's CAS added I think you're right > I hope that it has BOTH sdio and the CFII for microdrive (I like to have > 2GB HD for Linux etc..) > Yes, great expensive tools and much battery eaters. BUT you don't have to use them, off switch for microdrive... > Just for wishes, add a 640x480 landscape screen ;-) X In HP-200LX craddle, the display could be thinner keytops should have calc functions and a separate overlay would give QWERTY === Subject: Re: The Hydrix Calc/PDA i think they will use a calculator display which has only one bit per pixel (on/off) and no pointer. therefore they will no be able to use xlib (and its derivations -> toolkits). if they were use a grayscale display (intensity) and a touchscreen then they could use xlib. i think they'll use a regular LCD by looking a the sketch (which looks like has a hard cover display) and because it draws less power. greetings > , > > i think they use a normal LCD display (like a calculator) to save > battery + i think it is not touchscren. the question would be what > library they use for the GUI (not like a toolkit for X11). do they make > one or do they buy one? > Well, with X11 (or kdrive-xserver) + Gtk, or microwindows, they don't have > to buy, just adapt it to their device. With Qt, they have to pay for it, > if they get money with it. It is a license matter. >I need to purchase a new calculator. I used a 48SX in high school >years ago, so Im familiar with the very high quality of the >case/keypad/interface of the 48 series. Looking at pics of the 49G+, >it looks more toylike, compareable to the quality of the TI calcs. >I'd like to find out if this is true, and if the operation of the 49G+ >is more comparable to the 48's or TIs. > How does the 49G+ stand up to the older 48's as far as build quality >goes? Other than the dropped character glitch how does the keypad >stand up? Is this glitch likely to be fixed in future ROM revisions? >Would the 49G+ be fairly easy to pick up and use if already familiar >with the 48 series? How does the 49G+ stand as far as calculation >speed? Is it as versital and powerful as the 48's? Would you >recommend picking up a used 48GX or 48G+ as opposed to buying the >newer 49G+? > I'm looking for pro's and con's both ways. === Subject: Quick question re: NullLams... Just a quick question regarding local variables in SysRPL: from what I Is it a significant difference (i.e. worth the trouble of converting a large program to use them)? === Subject: HP 15C sold for $ 481 !!! Unbelievable: Just sold on German ebay: HP 15C for 402 Euros! Watch this item number: 3816163518 Jens === Subject: Re: Simplify option in hp > It could be that the 49G has more powerful general CAS algorithms. There is no real algorithm for simplification, as far as I know, partly because the problem is somewhat vaguely defined and partly because of the zero-testing problem. -- === Subject: Re: Need HP PDL replacement for linux > If you already have the HP 49G, > which does not need batteries > and is included in the price > I wonder why you prefer the HP 48 models??? but there is probably not a large enough market for this. I still like the 48 for the expandability. My requirement is to move the most current backup to a spare HP48 without serial I/O if the primary 48 dies for some reason. This means a memory card is the medium to transport the backup to a new machine. So the 49 (or any G model) fails this requirement. The GX model is stand-alone and needs not support system. So expandability is necessary without compromise. -- -MMax- === Subject: Re: HP 49G+ noisy/unreliable Keys > What ROM Revision are you using? Werner === Subject: [Announce] New release of RPL/2 RPL/2 is a RPL interpreter under GPL licence. It can be used on a Posix system (Linux, BSD, Solaris...). You can find the last release (4.00pre7j) at http://www.systella.fr/~bertrand/rpl2 JKB === Subject: Re: [Announce] New release of RPL/2 Hi! This seems nice for a handheld device, doesn't it? -- J.Manrique L.97pez de la Fuente Club de Usuarios de Gij.97n Socio 1077 del HPCC Socio de AsturLiNUX e HispaLiNUX === Subject: Math Fun: Composite Repunit Heads! The Recreational Mathematics Department Presents: COMPOSITE REPUNIT HEADS! by Joe Horn A repunit is an integer whose digits are all 1's, e.g. 11, 111, 1111 and so on. If there are n digits, then it's called Rn. Thus R7 (Repunit 7) = 1111111. The mathematical definition of Rn = (10^n-1)/9. Prime repunits are rare; the only ones known are R19, R23, R317, and R1031. It is *possible* that R49081 and R86453 are prime, but nobody is sure. However, what about numbers like 781111111111111, which have a repunit tail following a non-repunit head? We could call this example 78R13, since it's a 78 followed by Repunit 13. Numbers of this form are often more prime dense than repunits. For example, if you start with a repunit head of 2, then you get the following primes: 211 = 2R2 (prime) 2111 = 2R3 (prime) 2111111111111 = 2R12 (prime) 2111111111111111111 = 2R18 (prime) 211111111111111111111111 = 2R23 (prime) 2111111111111111111111111111111111111111111111111111111111 = 2R57 (prime) etc. The set of 3Rn is even more prime dense. All of the following are prime: 3R1 (better known as 31), 3R2, 3R5, 3R10, 3R11, 3R13, 3R34, 3R47, 3R52, 3R77, 3R88, etc. 35Rn is prime for the following n: 2, 3, 8, 9, 11, 14, 17, 23, 29, 38, 50, etc. Some xRn, however, are outrageously prime sparse, such as 12Rn, which has only one prime between 12R1 and 12R845. [Excercise: find it. Hint: REPEAT 10 * 1 + UNTIL DUP ISPRIME END.] The following brainstorm is probably obvious to Rodger and friends, but was not obvious to me until I thought about it for a while: Some xRn are NEVER prime! For example, *ALL* 37Rn are composite (when n>0 of course). [Exercise: prove this. Hint: factor them.] Since xRn is always composite when x=37, we can call 37 a composite repunit head. The Complete Sequence of Composite Repunit Heads begins like this: { 37, 176, 209, 407, 814, 936, 1023 ... } EASY HOMEWORK: find the next element in this sequence. Nonhint: it's NOT in Sloane's online encyclopedia of integer sequences at http://www.research.att.com/cgi-bin/access.cgi/as/njas/sequences/eismum.cgi DIFFICULT HOMEWORK: classify these numbers so that they can be generated by a formula of some kind, rather than finding them by brute force examination of the factors of xRn. EXTRA CREDIT HOMEWORK: Concoct a real-world use for this sequence. -Joe- === Subject: Re: EQLPlus ! Transform the database into a list, then you can store it in the SD-card. You have to modify the browser lib too, but that isn't much work either. I can't remember exactly how i did it, but there are a command there who makes a rompointer out of two numbers (1009(?) and another). Modify that. You give the database a name (now a list), and let that name and the object's place in that list be the new adresses. Good luck! Emil === Subject: Is there a FAQ Please ? Just wondering. Jo === Subject: price on hp48G+ Hey, I was wondering what the average price for used HP 48g+ calculators go for (with the manuals and link). I just put one up on ebay and was curious on what they go for. http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=5701425824 Anand Sekaran === Subject: Re: Sending/receiving data to HP-48 calculator under FreeBSD > people > How to receive/send files from HP-48GX calculator under the FreeBSD OS? > I've managed to work it in Linux (Slackware 9.1), for FreeBSD's is it the same > process as in Linux? > []s Yep, it aught to be the same way. Just be sure you've got the right software installed (sx and/or kermit) -MrM === Subject: Re: Sending/receiving data to HP-48 calculator under FreeBSD :> How to receive/send files from HP-48GX calculator under the FreeBSD OS? :> I've managed to work it in Linux (Slackware 9.1), for FreeBSD's is it the :> same process as in Linux? : : Yep, it aught to be the same way. Just be sure you've got the right : software installed (sx and/or kermit) : See: http://www.columbia.edu/kermit/hp48.html - Frank === Subject: Re: HP49: Recursive EDIT? > This must be a mistake. InputLine does use the normal editor. For > example, if you have the file on the stack, you can use this as a > viewer: > :: > NULL$SWAP (prompt for input line - we use empty string) > 1 (initial cursor position - can be any bint) > ZEROZEROZERO (Flags for alpha mode etc...) > MenuDef@ MenuRow@ (Lets use the current menu - or specify other) > TRUE (attn flag) > 0 (parse flag) > ' Inputline RunInNewContext_ (run Inputline) > caseDROP (if the user exited with OK, remove string) That's *much* better! The example I cut out of RPLMAN does nothing more than duplicate the UsrRPL INPUT command. NULL$SWAP instead of NULL$ made a big difference, especially in restoring full cursor control. The reason I mentioned SCROLL is that it's a viewer with no features, just like my initial attempt. This has restarted my interest in SysRPL! === Subject: Re: Any word on ROM 1.24 for the 49G+? Where have you seen ROM 1.24 for the hp49g+ ? I haven't seen anything new in hp's site. === Subject: Re: Any word on ROM 1.24 for the 49G+? It was announced around january 04. for end Q1 O4. Instead 1.23 came much earlier fixing the battery bug, since then no news... > Where have you seen ROM 1.24 for the hp49g+ ? > I haven't seen anything new in hp's site. === Subject: Euler's Constant boundary=----=_NextPart_000_04EA_01C44B13.E7EAE7D0 --------------------------------------------------------------------- I'm trying to calculate Euler's Constant on my 49g+. In level 2 I have N .91£ 1/n-LN(N) n=1 Level 1 has N=.89Ì.9b When I try to take the limit, I get Operator not implemented (SERIES) What am I doing wrong? L === Subject: Re: Euler's Constant boundary=----=_NextPart_000_0023_01C44C1A.50E08A90 --------------------------------------------------------------------- I guess the calc is trying to give a symbolic answer and has not been loaled with the formal definition of Euler constant... Why not writing a summation loop with a stop test when two consecurive steps differ by less than something very small ? Knowing the series converges, you should not be too far from a correct value... I'm trying to calculate Euler's Constant on my 49g+. In level 2 I have N .91£ 1/n-LN(N) n=1 Level 1 has N=.89Ì.9b When I try to take the limit, I get Operator not implemented (SERIES) What am I doing wrong? L === Subject: Re: EQL+ (WR) INVALID CARD DATA... still remain > system animation but INVALID CARD DATA is RETURNED damn it !! > PLEASE remake the lastest version of this lib with correct CRC..... Why don't you do it? Install OT49, type in the library number of EQL+ on the stack, and push D<->L in the OT49 menu. It decompiles the library. Then push D<->L again and you will have a library sitting there. Then install it. I assume you have already run the command PINIT? TW === Subject: Re: EQL+ (WR) INVALID CARD DATA... still remain > system animation but INVALID CARD DATA is RETURNED damn it !! > PLEASE remake the lastest version of this lib with correct CRC..... Oby, you misunderstood JYA. The library has 37013 bytes, CRC A50B. Verify this! It's definitely not corrupted. Download only with Conn4x! Only you yourself may corrupt it with some unusual manipulation. Put the lib on the stack and purge it. Only then put it in Port1 and make a warmstart. I forgot to update the UsrBox example in EQL+.zip. This has nothing to do with the library and will be corrected in an hour. === Subject: Re: Mass-Spring-System animation in EQL+ > the exciting mass-spring-system animation from the 48 equation library > now also runs in EQL+ revision from today! A special animator has been > written for the right 49+ tempo (maybe ok also on the 49). In contrast > to the 48 animation, ours keeps running until interrupting it with > CANCEL. Hence, there is enough time for admiration I've been adminring it. . . I really don't like how I have to change my fingers to dive into the equations now. you have on F5-F6 arrows to do it. Some like that, other (such as myself) would rather have the L-R arrow keys do the same function. I think it is much quicker. Why not stick both methods in? TW 1038 === Subject: Re: OT Re: Giac/Xcas plot3d problem Please Help > You have a last hope: try the red geo configuration button > and look if you have a 3-d pipes check button. Try to uncheck it > and save the configuration. No luck... but I learned that I can save the config so the program starts with a bigger font. > If it still does not work, I'm afraid I won't be able to > fix the problem before long since I'm on holiday with a linux-only PC. > Hopefully the rest should work (there are only a few differences > between the win and linux version, gnuplot interface being one of > them). And for 3-d plotting, you still have the option to add a linux > partition to your hard drive:-). Actually, I found out I can just use GNUPlot directly to graph. Version 4.0 can interactively rotate the 3D graph. I actually did install Linux several years ago (from a CD), but I thought it was too difficult to learn and there wasn't an easy manual:-(. > I also expect to bring 3-d geometry > at the beginning of September... Great! === Subject: Re: Why the French guyz rule the HP scene? hA@comcast.com: > Coulomb 1736 > Descartes 1596 > Fresnel 1788 > Legendre 1752 > Fermat 1601 > Pascal 1662 > Biot 1774 > Ampere 1775 > Navier 1785 > Jean Picard 1620 > If only the French could bring back the 18th century > Invented the calc? > mechanical? > and the what about William Wickes? You left off Fourier and Laplace. Couldn't find out more information about them? Our modern civilization is based on the works of French mathematicans and the United States exists because of the French government. Just to point out the obvious. There's a bigot in every crowd... the name of which, comes from France. Stan === Subject: Re: Why the French guyz rule the HP scene? Fourier 1768 Laplace 1827 well at least one in the 19th century. It is typical of the French to take credit for everything. An American proved Fermats theorem not Fermat, and a Russian recently proved Poincare's conjecture. The numerals that everyone uses whether than write with a the Hindu contributions). France would love to return to it's period of Greatness when it was lead by a Corsican France did not give us our freedom. The government of France was at war with England and anything they could do to make trouble for the British was their agenda. And of course arms sales (nothing has changed). Lafayette a wealthy French nobleman was very helpful. He is the one who convinced Van deGraf to confront the British fleet off the the coast of Yorktown, preventing the retreat of Cornwallis. Van deGraf was already on the way to confront the British off the West Indies. After the revolution was over Lafayette was persona non gratis in France as he spoke of this American Democracy. Arrangements were made by the newly formed congress for him to live here in the States, then after the French revolution he was once again welcome. DeGaule's insurrection in Paris almost lead to its destruction as the Nazis had planted explosives at all the monuments in Paris. Patton who had the German Army on the run was trying to reach the Rhine before they had a chance to counter attack or blow the bridges. Eisenhower insisted that save Paris. The Americans and British liberated Paris and deGaule marched under the Arche de Triumph claiming victory. The cost in fuel halted the advance and prolonged the war. The Battle of the Bulge which involved more than a million men and in which the Americans suffered 81,000 casualties and 19,000 dead. After the war deGaule said he wished the Americans never came. They would not join Nato developed their on Nuclear force and pointed the missles at us (we will point the missles at all points of the compass). The HP community has had contributions from Australia, United States, Finland, England, Germany, Italy, Spain, Portugal, Mexico, and many other countries. One has to seriously question why they sit on the security council. India will soon be the most populous country on earth and Japan's economic prowess dwarfs France. Have a nice bottle of Chilean wine and stop singing about the French. > hA@comcast.com: > Coulomb 1736 > Descartes 1596 > Fresnel 1788 > Legendre 1752 > Fermat 1601 > Pascal 1662 > Biot 1774 > Ampere 1775 > Navier 1785 > Jean Picard 1620 > If only the French could bring back the 18th century > Invented the calc? > mechanical? > and the what about William Wickes? > You left off Fourier and Laplace. Couldn't find out more information about > them? Our modern civilization is based on the works of French > mathematicans and the United States exists because of the French > government. Just to point out the obvious. There's a bigot in every > crowd... the name of which, comes from France. > Stan === Subject: Re: Why the French guyz rule the HP scene? They invented champagne and camembert ... > Fourier 1768 > Laplace 1827 > well at least one in the 19th century. > It is typical of the French to take credit for everything. An American > proved Fermats theorem not Fermat, and a Russian recently proved Poincare's > conjecture. The numerals that everyone uses whether than write with a to > the Hindu contributions). > France would love to return to it's period of Greatness when it was lead by > a Corsican > France did not give us our freedom. The government of France was at war > with England and anything they could do to make trouble for the British was > their agenda. And of course arms sales (nothing has changed). Lafayette a > wealthy French nobleman was very helpful. He is the one who convinced Van > deGraf to confront the British fleet off the the coast of Yorktown, > preventing the retreat of Cornwallis. Van deGraf was already on the way to > confront the British off the West Indies. After the revolution was over > Lafayette was persona non gratis in France as he spoke of this American > Democracy. Arrangements were made by the newly formed congress for him to > live here in the States, then after the French revolution he was once again > welcome. > DeGaule's insurrection in Paris almost lead to its destruction as the Nazis > had planted explosives at all the monuments in Paris. Patton who had the > German Army on the run was trying to reach the Rhine before they had a > chance to counter attack or blow the bridges. Eisenhower insisted that save > Paris. The Americans and British liberated Paris and deGaule marched under > the Arche de Triumph claiming victory. The cost in fuel halted the advance > and prolonged the war. The Battle of the Bulge which involved more than a > million men and in which the Americans suffered 81,000 casualties and 19,000 > dead. After the war deGaule said he wished the Americans never came. They > would not join Nato developed their on Nuclear force and pointed the missles > at us (we will point the missles at all points of the compass). > The HP community has had contributions from Australia, United States, > Finland, England, Germany, Italy, Spain, Portugal, Mexico, and many other > countries. > One has to seriously question why they sit on the security council. India > will soon be the most populous country on earth and Japan's economic prowess > dwarfs France. > Have a nice bottle of Chilean wine and stop singing about the French. > hA@comcast.com: > Coulomb 1736 > Descartes 1596 > Fresnel 1788 > Legendre 1752 > Fermat 1601 > Pascal 1662 > Biot 1774 > Ampere 1775 > Navier 1785 > Jean Picard 1620 If only the French could bring back the 18th century > Invented the calc? > mechanical? > and the what about William Wickes? > You left off Fourier and Laplace. Couldn't find out more information > about > them? Our modern civilization is based on the works of French > mathematicans and the United States exists because of the French > government. Just to point out the obvious. There's a bigot in every > crowd... the name of which, comes from France. > Stan === Subject: Re: Why the French guyz rule the HP scene? > They invented champagne and camembert ... Yeah but they also eat truffles. Imaging using pigs to find your food! Oh well, at least Haggis (oatmeal in sheep guts!) isn't French. Lake === Subject: Re: Why the French guyz rule the HP scene? Jersey Devil, you made a wonderfull presentation, it is significant that you left out exactly those two dead scientists. there is something else that puzzles me though: How come you (inteligent, informed guy) can be offended by simple statement where i congratulate the French people on their work. I didn't say they are the only ones who contribute to the sceene, i just stated i think they are doing a great job. And it shows. (you cann see it everywhere -just look around) Let's stop worrying about politics, about who invented the wheel and staff. Because all this is not about French people as a nation, it is more like French people as enthusiastic individuals. And Stan i agree absolutely with you modern world is based on gret findings of the past, no matter if the finding was British, American, French, German Irish or whatever natinality of the reader reading this may be. One american stated though: in the early dayes of moder civilization in France (europe in general) there were cultural events, where on the american continent there was a chase for buffalos and meere survival the main issues of life. Let's be honest America is young nation and there is nothing we can do about it -the best thing is to respect it as is. Now you got me getting political -i just want to say i give up. I don't wanna get caught in useless political debate. I just said i admire people's work, and my mind is at peace. Let's break this politico-socio-natio chains that tie us to the ground and fly away to 24. century in peace. If enegry and food were free -would there be any wars ? === Subject: Re: Why the French guyz rule the HP scene? No arguements about the French.... However, from an American, this sounds alot like the pot calling the kettle black... Almost every criticism you aim at the french can be turned around onto America... Not withstanding your fearless leader's decision to try to make a scapegoad of the French when pretty much the whole world expressed unhappiness at the decision to invade another country against the UN. Can anyone think of unpleasant instances in the past when politicians who were on shaky ground used the diversionary tactic look over at those (evil) people over there so that they could do what they wanted? How many times has this tactic worked out for the good in the past? And what makes it worse in this particular case is the percentage of we were right is tiny compared to those facts turned out not to be correct. Even if the Bush administration aren't working with evil intent, they are incompetant in global politics and should gain some humility. Al... (this is the 4th time I've started to post something on this totaly off-topic thread, but I kept deleting them... sorry for those who, like me feel this is an inappropriate forum for this discussion.) > Fourier 1768 > Laplace 1827 > well at least one in the 19th century. === Subject: Re: How $ Important is the HP Calculators Market? Maybe you might find something in their annual report? > Where would I search for quantitative data on HP Calculator sales? > This question began as I was raving about my 33s to a very patient > friend of mine and he asked me, > Well, if it's absolutly the best to be found, > then How many does HP sell, and how much is it worth? > I Dunno-, I said, Lots and Lots ? I began to feel as if I may have > directed my concerns toward the tempest in a tea-cup, so to speak. > Later, I went nosing around on WWW and found: Nothing. > Where would I access quantitative data on HP Calculator sales? > Jo ( Who can never have enough information. ) === Subject: Re: HP49g+ - NUM.SLV>Solve linear System - Bug or...? >You can do this by recalling the A and B on stack >and then doing ->NUM for both of them >and finally storing them back. >Try it now - Integers are considered symbolic You're right ;-) I didn't see that all the zeros were stored as integers. Entering them as 0. and storing the matrix everything works fine after choosing them from within the NUM.SLV Lin Solver. Pascal === Subject: Re: HP49g+ - NUM.SLV>Solve linear System - Bug or...? > You can do this by recalling the A and B on stack > and then doing ->NUM for both of them > and finally storing them back. > Try it now - Integers are considered symbolic But they should be auatically coerced to approximate values, right, since all entries are numeric? Anyway, I guess that can't be changed now. -- === Subject: Re: HP49g+ - NUM.SLV>Solve linear System - Bug or...? > You can do this by recalling the A and B on stack > and then doing ->NUM for both of them > and finally storing them back. > Try it now - Integers are considered symbolic But they should be auatically coerced to approximate values, right, since all entries are numeric? Anyway, I guess that can't be changed now. -- === Subject: Re: HP49g+ - NUM.SLV>Solve linear System - Bug or...? >But they should be auatically coerced to approximate values, right, >since all entries are numeric? Anyway, I guess that can't be changed >now. Take a look at my answer on Hugo'a posting... The HP is right, I was wrong... Pascal === Subject: Re: HP49g+ - NUM.SLV>Solve linear System - Bug or...? I've also had this very annoying bug happening to me several times. And only if I choose the Matrixes with CHOOSE from my local vars. It triggered one Try To Recover Memory on my 49G+ for 3 times (until I was shure it was a bug). Even though one can resolve this the way suggested I consider this a very annoyng bug which has to be ironed out in a next ROM Update. And when is that comming anyway?!!! Hugo Rodrigues >> Hi Group... >> Today I tried to use the linear system solver with stored matrices >> (normally just using the S.SLV LINSOLVE to get exact results or the >> NUM.SLV lin. solver with manually entered matrices) and got a really >> odd reaction of my calc... >> I created the following matrices: >> [.9 -.7 0 0 ] >> [.7 .7 0 0 ] >> A: [0 .7 .5 -.9] >> [0 -.7 .5 .5 ] >> [0 ] >> [1 ] >> B: [0 ] >> [.5] > [.9 -.7 0. 0. ] > [.7 .7 0. 0. ] > A: [0. .7 .5 -.9] > [0. -.7 .5 .5 ] > [0. ] > [1. ] > B: [0. ] > [.5] > You can do this by recalling the A and B on stack > and then doing ->NUM for both of them > and finally storing them back. > Try it now - Integers are considered symbolic > >> I stored both into variables and selected the matrices using the >> choose-key. A for A and B for B to solve for X in A*X=B. >> When I set the cursor to X: and press solve my calc starts working and >> working... Something between 5 and ten minutes... After that, it shows >> 'Symbolic Matrix' as result for X... >> The odd thing is: When I enter the matrices by hand (exactly as shown >> above) using edit everything works fine... >> Is it a known problem and/or duplicateable? Or did I something really >> dumb wrong? >> Pascal === Subject: Re: HP49g+ - NUM.SLV>Solve linear System - Bug or...? >Even though one can resolve this the way suggested I consider this a >very annoyng bug which has to be ironed out in a next ROM Update. And when >is that comming anyway?!!! It's not a bug - it's more a malfunction of some synapses in the users brain ;-) Look at that: My matrices: [.9 -.7 0 0 ] [.7 .7 0 0 ] A: [0 .7 .5 -.9] [0 -.7 .5 .5 ] [0 ] [1 ] B: [0 ] [.5] Take a look on all the zeros and the 1 - all these numbers are integers because the decimal point is missing.... If enter them as: [.9 -.7 0. 0. ] [.7 .7 0. 0. ] A: [0. .7 .5 -.9] [0. -.7 .5 .5 ] [0.] [1.] B: [0.] [.5] (Or using ->NUM with the original one) All numbers are floating point values - these matrices can be choosen by the solver and solved afterwards. The solver is not able to convert matrices containing real and integer values... And if you enter the matrices using edit within the solver all values are stored as floating. Try it... Sometimes, I'm too dumb... Grmpf! That's all... Pascal === Subject: Re: HP49g+ - NUM.SLV>Solve linear System - Bug or...? > It's not a bug - it's more a malfunction of some synapses in the users > brain ;-) It is a bug alright. > The solver is not able to convert matrices containing real and integer > values... It should be able to. At the very least it shouldn't result in a TTRM. === Subject: Re: SD-card formatting: cluster size? >> I can confirm approx 248000 KB free with just formatted card: so it >> seems > it >> is the filer to uncorrectly show the amount of free disk space... >> Now that you mention it, I think I've seen other posts about the filer >> showing only 128KB free on larger cards. > Do you mean 128000 KB? > Anyway what I mean is that maybe the Filer shows > uncorrectly the frre disk space, BUT indeed I WAS WRONG! > I've used the Analyze function from Defrag built-in in Windows2000 and it > says that the cluster size is 2 KB! What a great mistery! Microsoft > clearly > saysy at > http://www.microsoft.com/resources/documentation/Windows/XP/all/reskit/en-us > /Default.asp?url=/resources/documentation/windows/xp/all/reskit/en-us/prkc_f > il_lxty.asp that for a 256MB drive the cluster size will be 4KB with > FAT-16... >> I guess someone at Microsoft is smart enough to figure out that with >> FAT16, using 2K clusters would waste about half of a 256MB drive. But >> sometimes their software gives strange results. > Is the calc formatting the SD-card in FAT32? >> I haven't heard of that, but who knows? I don't have an SD card big >> enough to make FAT32 worthwhile. Actually, I not even sure that FAT32 >> is supported for drives smaller than 512MB. >> Is your 49g+ running on ROM revision 1.23? > Can the pre-formatting of the > SD-card influence further formats? >> Yes. For example, when I was fooling around, I trashed the card's boot >> record, and experimentally found that neither the Windows 98SE format >> nor the 49g+ format would change the number of sectors recorded in the >> boot record. If I recall correctly, I temporarily changed my 128KB >> card to a 16KB card. I ended up tweaking it with the disk editor. It >> was fun though. I don't know whether a DOS format would've gotten it >> straightened out. > Nononononono ... boot record contains partition table, no actual FAT > data. All format utilities just format one partition, at whatever > size is reported in the boot record. Perhaps you mean a master boot record (or partition sector)? But as far as I can tell, my SD card has no master boot record, and thus no partition table. Some disk utilities do show me a partition table, but I'm a bit skeptical of a partition on the card being hundreds of gigabytes or even terabytes in size. I think what happens is that they look at sector 0 (the boot record) and tell me what it would mean if it really were a master boot record. The boot record (or boot sector ) certainly does specify such things as bytes per sector, sectors per cluster, number of FATs, sectors per FAT, number of root directory entries, and total sectors on the drive. If you have partitions it's at sector 0 of each partition. But since my SD card doesn't have partitions, it's at sector 0 of the card (or at least what the SD system presents as being logically sector 0). It seems to be much like a floppy disk, except a lot bigger than any floppy I've ever used. If I null out the boot record and then tell the 49g+ to format the sectors, which indeed works out to exactly 128MB (as long as we define 1MB as 2^20 bytes). But if I try reading from or writing to all of those sectors, I run into a bit of a problem. I chose 246016 sectors simply because that's the highest value that works with Norton Disk Doctor without freezing up near the end of the drive when I try a surface test. Once the card has a valid boot record, the 49g+ uses the total sectors already recorded in the card's boot record to determine the appropriate sectors per cluster and sectors per FAT and doesn't prompt me for the card size. > To un-do what you did, you'd have > to use fdisk. Oh, well, it works just fine for me without using FDISK on it. > Except DOS fdisk doesn't do USB-attached disks (AFAIK). That seems to be true. > So you'd have to use Linux fdisk =) Free FDISK seems to work pretty well. What does your Linux fdisk tell you about the SD card? But it might be interesting to put an MBR on the card. I wonder how the 49g+ would respond to multiple partitions on the card. Or what it would do with a card formatted as FAT32, for that matter. -- === Subject: Re: HP49G+ informations Well, I would knowit someone here is a little bit interrested by what I'm doing, because I have no answers but a hundred visits. I think it is spambots... :(( well, i'll continue updating my page. Hope it's useful... === Subject: Re: crashed while updating ROM - What do I do now? > Same thing happend to me, I just reset and it worked fine. >>My PC crashed as I was updating to ROM 1.23. Now my calc does not have any >>ROM anymore and the connectivity kit can't find my 49g+ anymore. >>The calc is stuck on: >>FLASH NEED UPDATE. >>1.USB >>2.CARD >>I don't have a card reader so I can't go this way. >>Anyone with an idea on how do I put a ROM back in the calc? >>Arnaud to the card, but then tried to flash and HP49+ wouldn't boot. Took me to the screen where I need ROM update and was given the choice: 1 USB or 2 Card, and I chose card, since Conn4 was completely wasted. The memory addresses were displayed continously in what seemed an endless loop. Somehow, the thing managed to reboot and miraculously upon typing version HP read Revision #1.23 which is the one a was trying to flash in. Moral: Do not rely on Conn4 or anything having to do with Wincrap. Use the card. Buy the card and make sure it works before attempting an upgrade. Plug and Pray. BTW, was the original Q ever Aswd. What support is there when the ROM gets fried? John Baranowsky === Subject: Re: SPLIT challenge > can i interest you in another challenge ? > -to write a standard SPLIT comand: > command would search the STK2 string for occurances of string in STK1, when > found string would be splited at that place leaving split-bits on the > stack. > If not found STK2 should remain intact. STK1 removed. OK, I'll try it. But I'm still learning SysRPL basics, so it might not be a great program. === Subject: 48GX to 49G+ I'm a newbie to HP graphing caculators. So before I plop down the cash for a 49G+ I was wondering if programs from my friends old 48GX can be transfered/converted to a new 49G+. === Subject: reversing bits I believe someone lately asked, whether it's possible to reverse the bits of a n-bit quantity without iteration, i.e. using a function. Yes, indeed it is. Here is a program in UserRPL, which reverses the bits for an 8 bit integer in stack level 1: %%HP: T(3)A(R)F(.); << # 202020202h * # 10884422010h AND DUP # 3FFh DUP UNROT / * - This one has it's origin in the famous PDP-10 Programimng Hacks, developed at MIT in the mid 70s. Note, that this implementation is *not* efficient on a 4 bit machine, since it involves large integer arithmetic. So for the HP49(+) it's probably just of academic interest.. On a modern 64 bit CPU however, this could be implemented very efficiently. reverse8: ; 8 bit argument in A MULU A,#0x000202020202 AND A,#0x010884422010 DIV A,#0x3FF (fictitious CPU) For reference here the C implememtation: // #define unsigned41 unsigned __int64 // #define unsigned41 unsigned long long // uncomment one for cl or gcc unsigned reverse_8bits(unsigned41 a) { return ((a * (unsigned41) 0x000202020202UL) /* 5 copies in 40 bits */ & (unsigned41) 0x010884422010UL) /* where bits coincide with reverse repeated base 2^10 */ /* PDP-10: 041(6 bits):020420420020(35 bits) */ % 1023; /* casting out 2^10 - 1's */ } -- MM 1775 === Subject: Re: HP48G+ programming Is that a HP49G+ (which is new) or a HP48G+ (quite old) you're talking about? Should work in both cases anyway. I'd use -1 WAIT (forces menu to be displayed and pauses program execution) instead of HALT. Hope this helps > Hi all, > after 15 years I switched from my old hp28s to the new HP48G+, trying to > convert my old programs for the new machine, I found a few problems. > First of all in the box HP sent me there is just the user's guide. This last > one talks about a HP 48G Series Advenced User's Reference # 00048-90136, > but HP Italy doesn't know anything about it, nor can I find it anywhere. > Besides, I'm looking in particular for instruction for data insertion, not > just for local variables but for global ones. For the old HP the > instructions are: << { STO X Y Z } MENU HALT ..... > while for the new 48G+ the command MENU doesn't work anymore. > Can anybody help me? > You can download the HP 48G Series Advanced User's Reference here: > http://www.hpcalc.org/search.php?query=advanced+reference&hp48=1 > Warning: It's a 36 mb ZIP file that unzips into a 39 mb PDF file. > -- > 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: TI89 Titanium -- opinions? > Here then it is you who don't know what you are talking. > If you have spent as much time as me on TI calculators and following > their evolution as close as i have you would know that: > *Since the TI89,TI has never tried to improve the speed of its TI68k > calculators by software optimisation You can't know this for a fact unless you are a developer or tester at > So knowing this and knowing that the TI89 Titanium is to the best a > TI89 Hw2 with more Flash ROM or to the worse a TI89 version of the > Voyage 200,i can easily conclude that the TI89 Titanic is not faster > than the TI89 HW2. No, you cannot conclude that without actually doing tests with the TI-89 Titanium. It is quite likely that you are correct in your assertion, but there is a chance that you are not. > I don't really know. Nostalgia? You haven't noticed that 90% of the '95-'00 > regulars aren't here anymore? You don't think HP is dropping on the quality > scale and have been doing for some years? You sincerely mean that the tone > of c.s.hp48 today is reminiscent of what it was 6-8 years back? > The HP quality has decreased anyone will agree with that. > Your second statement was attack toward me so what did you expect me > to do ? I don't see why you think that statement was an attack on you. I think was simply commenting on the general quality of posts. > Btw try this on your TI89: > for i,1,3000:0->#(k&string(k)):endfor > and tell me what happen. > But you should have a way to launch this other sequence: > for i,1,3000:delvar #(k&string(k)):endfor I know, you brought this up earlier too. Oddly enough, I don't see it in the bug list; I'll add it. . === Subject: Re: TI89 Titanium -- opinions? > Here then it is you who don't know what you are talking. > If you have spent as much time as me on TI calculators and following > their evolution as close as i have you would know that: > *Since the TI89,TI has never tried to improve the speed of its TI68k > calculators by software optimisation You can't know this for a fact unless you are a developer or tester at > So knowing this and knowing that the TI89 Titanium is to the best a > TI89 Hw2 with more Flash ROM or to the worse a TI89 version of the > Voyage 200,i can easily conclude that the TI89 Titanic is not faster > than the TI89 HW2. No, you cannot conclude that without actually doing tests with the TI-89 Titanium. It is quite likely that you are correct in your assertion, but there is a chance that you are not. > I don't really know. Nostalgia? You haven't noticed that 90% of the '95-'00 > regulars aren't here anymore? You don't think HP is dropping on the quality > scale and have been doing for some years? You sincerely mean that the tone > of c.s.hp48 today is reminiscent of what it was 6-8 years back? > The HP quality has decreased anyone will agree with that. > Your second statement was attack toward me so what did you expect me > to do ? I don't see why you think that statement was an attack on you. I think was simply commenting on the general quality of posts. > Btw try this on your TI89: > for i,1,3000:0->#(k&string(k)):endfor > and tell me what happen. > But you should have a way to launch this other sequence: > for i,1,3000:delvar #(k&string(k)):endfor I know, you brought this up earlier too. Oddly enough, I don't see it in the bug list; I'll add it. . === Subject: Re: HP49G+ informations Some news : determinate where is was... :))) I began understanding the memory. Same link : http://lebonpoint.chez.tiscali.fr/hp49gp/ If you have something to say about this page, or any useful informations to be added, tell me !!! a++ y. === Subject: Re: Why can I imagime this on LCD ? >> http://research.microsoft.com/users/blinn/MATHDESN.HTM > I don't know why you can imagine that on an LCD. I would prefer to > imagine it erased. In the few seconds I gave it, I found this: >> We used the letter t for the angle because some feedback indicated > that many >high schoolers might be intimidated by Greek letters. > And they spelled Archimedes wrong. Microsoft seems to be doing > research for droolers and knuckle-draggers. > Grrr - haven't had coffee yet Markwick Gates is an idiot and you don't need to apologize for pointing it out. http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000 Newsgroups ---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =--- === Subject: OT Re: Giac/Xcas plot3d problem Please Help > I downloaded and extracted this package, and it's working. But I have > no idea why when I try to plot a 3D function, eg, > plot3d(sin(x*y),x,y), or any other functions, it returns the same plot > over and over again. I've checked and it's the image in the file > session03d.png. > How can I make plotting a function in 3D work? Do I need to install You don't need latex for 3d plots but you need gnuplot, therefore the Mac version of xcas can not do 3-d graph but the situation should change when the mac port of gnuplot 4.0 will be finished. If you're not on a mac, I need more precisions: what version do you use? Then after you entered your command, did you see the gnuplot window with the right function? After that, you should get the graph in the history window, but I have a race condition that I did not solve yet (the png image computed by gnuplot is not finished when the history is redrawn), and sometimes you must enter something else in the history to update the graph correctly. === Subject: Re: OT Re: Giac/Xcas plot3d problem Please Help > You don't need latex for 3d plots but you need gnuplot, > therefore the Mac version of xcas can not do 3-d graph > but the situation should change when the mac port of gnuplot > 4.0 will be finished. > If you're not on a mac, I need more precisions: what version do you use? > Then after you entered your command, did you see the gnuplot > window with the right function? After that, you should > get the graph in the history window, but I have a race condition > that I did not solve yet (the png image computed by gnuplot > is not finished when the history is redrawn), and sometimes > you must enter something else in the history to update the > graph correctly. Here are the details that you asked: My computer is running Windows XP. My XCAS version is 0.4.0 copyright and it came with GNUPlot. Nothing appeared in the GNUPlot window when I enter my plot3d command (although plotting in 2D works - the 2D graph shows up in GEO and in history, but not in GNUPlot).... The history just returns the same graph every time when I plot3d. I ran XCAS by clicking on xcas.exe. What happens then is the GNUPlot window appears. In the window is displayed the general info about GNUPlot with a prompt at the bot. This window shows the blank space at the bot when it opened so I have to scroll up to see the info. Then the XCAS GUI window appears, which works (except for plotting 3D). The GNUPlot window just shows blank the entire time I use XCAS. Maybe I have to run from the .bat files and need to change the directory in the .bat file? I hope you can help me. Being able to plot 3D is really cool. Dave === Subject: Re: OT Re: Giac/Xcas plot3d problem Please Help > Here are the details that you asked: > My computer is running Windows XP. My XCAS version is 0.4.0 copyright > and it came with GNUPlot. > Nothing appeared in the GNUPlot window when I enter my plot3d command > (although plotting in 2D works - the 2D graph shows up in GEO and in > history, but not in GNUPlot).... The history just returns the same > graph every time when I plot3d. There are 2 modes for gnuplot under windows, either Unix-like or not interactive (the 2nd mode should work on all windows), the mode is controled by a check button. I'm afraid I did not test 3-d plots for long under windows (I program under linux and then I compile under windows but I don't make regression tests under windows) therefore you have probably found a bug. You have a last hope: try the red geo configuration button and look if you have a 3-d pipes check button. Try to uncheck it and save the configuration. If it still does not work, I'm afraid I won't be able to fix the problem before long since I'm on holiday with a linux-only PC. Hopefully the rest should work (there are only a few differences between the win and linux version, gnuplot interface being one of them). And for 3-d plotting, you still have the option to add a linux partition to your hard drive:-). I also expect to bring 3-d geometry at the beginning of September... === Subject: Re: OT Re: Giac/Xcas plot3d problem Please Help >> I downloaded and extracted this package, and it's working. But I have >> no idea why when I try to plot a 3D function, eg, >> plot3d(sin(x*y),x,y), or any other functions, it returns the same plot >> over and over again. I've checked and it's the image in the file >> session03d.png. >> How can I make plotting a function in 3D work? Do I need to install > You don't need latex for 3d plots but you need gnuplot, > therefore the Mac version of xcas can not do 3-d graph > but the situation should change when the mac port of gnuplot > 4.0 will be finished. > If you're not on a mac, I need more precisions: what version do you use? > Then after you entered your command, did you see the gnuplot > window with the right function? After that, you should > get the graph in the history window, but I have a race condition > that I did not solve yet (the png image computed by gnuplot > is not finished when the history is redrawn), and sometimes > you must enter something else in the history to update the > graph correctly. Are you still developing the Giac/Xcas? Is there a new version? What about the manual? Would it be easy to translate the program into Finnish? (UI and/or the error messages or/and help - maybe even the manual) Is there a Lambert W function on your future plans? Needed for solving ALL roots of, say 2^X=X^2 AND then Bessel, Zeta, Errf, etc... {} === Subject: Re: OT Re: Giac/Xcas plot3d problem Please Help > Are you still developing the Giac/Xcas? I work fulltime on it. > Is there a new version? Versioning is not my main preoccupation, you will find under the current 0.4.0 name the latest version which is updated frequently. I will adopt more rigorous version names when the code is more stable and clean and CVS-ed. > What about the manual? The French manual is almost complete. Only the tutorial is available in English, and it should be updated (since some simplifications in the interface have occured, especially for function plots directly inserted in the history). Everyone who want to contribute to the documentation is welcome! > Would it be easy to translate the program into Finnish? I have just added fi to ALL_LINGUAS and updated the source package. To localise in Finnish, edit giac-0.4.0/po/fi.po and translate all menus. Then make install will add finnish localisation. > (UI and/or the error messages or/and help - maybe even the manual) For the manual, I can add a doc/fi subdirectory. For on-line help, the file giac-0.4.0/doc/aide_cas should be edited, and Finnish translations added, for example using 4 as language number, then I can add a menu entry for Finnish on-line help. > Is there a Lambert W function on your future plans? It's probably not hard to add, just a question of time. As well as more HP4xG compatability... > Needed for solving ALL roots of, say 2^X=X^2 > AND then > Bessel, Zeta, Errf, etc... erf and zeta are available. Bessel will also be added of course, I just don't know when! === Subject: Re: OT Re: Giac/Xcas plot3d problem Please Help Cc: Veli-Pekka Nousiainen <@welho.com Are you still developing the Giac/Xcas? I work fulltime on it. > Is there a new version? Versioning is not my main preoccupation, you will find under the current 0.4.0 name the latest version which is updated frequently. I will adopt more rigorous version names when the code is more stable and clean and CVS-ed. > What about the manual? The French manual is almost complete. Only the tutorial is available in English, and it should be updated (since some simplifications in the interface have occured, especially for function plots directly inserted in the history). Everyone who want to contribute to the documentation is welcome! > Would it be easy to translate the program into Finnish? I have just added fi to ALL_LINGUAS and updated the source package. To localise in Finnish, edit giac-0.4.0/po/fi.po and translate all menus. Then make install will add finnish localisation. > (UI and/or the error messages or/and help - maybe even the manual) For the manual, I can add a doc/fi subdirectory. For on-line help, the file giac-0.4.0/doc/aide_cas should be edited, and Finnish translations added, for example using 4 as language number, then I can add a menu entry for Finnish on-line help. > Is there a Lambert W function on your future plans? It's probably not hard to add, just a question of time. As well as more HP4xG compatability... > Needed for solving ALL roots of, say 2^X=X^2 > AND then > Bessel, Zeta, Errf, etc... erf and zeta are available. Bessel will also be added of course, I just don't know when! === Subject: Re: OT Re: Giac/Xcas plot3d problem Please Help > Are you still developing the Giac/Xcas? > I work fulltime on it. I wonder why === Subject: Re: OT Re: Giac/Xcas plot3d problem Please Help >Are you still developing the Giac/Xcas? >>I work fulltime on it. > I wonder why There is nothing special about that, I work fulltime on giac since 4 years and 2 months now. === Subject: Re: Serial buffer overflow modifications: > ------------ > on the STK1 i'll have a large empty string, (say standard string object > about 1k or 2k in size -as large as expected data block will be) > In a very tight loop (compact and fast) i'll copy the data from the receive > buffer direcly in to the mentioned string -this will actualy be secondary > buffer, then i'll stop interrupts just for a brief moment so i can reset the > receive pointer/counter (which is somwhere in the mem near the buffer) > enable interrupts, > wait in a loop for pointer to change. > In the end (when sequence of characters is received or timeout period) > length of the string needs to be adjusted to my counter (which counted all > characters transfered). and the string should be ready to use. > I bet this will be a lot faster compared to creating, duplicating, droping, > adding objects on stack and whatever sysRPL does to the characters from > serial buffer when we SREC-v it. > I just hope it will be fast enough (it better be and i hope the issue is > not in the speed how emulator updates hardware related registers and memory > locations in the virtual Saturn. I'm probably venturing beyond my depth here, but for what it's worth: I'm not sure what you mean by an empty string but with a size of 1K or 2k. Maybe a string with that size field but no characters following? Is that possible? Maybe you could start what I mean as an empty string; one with a size of zero, and then append incoming characters to it using >T$. Maybe use a virtual stack to start with an empty stack, put the incoming characters on the stack (one per level), and then when finished receiving, put an empty string on the stack, and (in a loop) prepend the characters to it using SWAP >H$ until only the string is on the stack. Of course follow that up with saving the string, restoring the saved virtual stack, and then putting the string on the stack. Perhaps there's no need to keep a count of the characters as you're receiving, as long as it can be determined later. Clearly, the less time used dealing with incoming data while actually receiving, the better. Re-arranging it into a nice character string in level one can be done at a more leisurely pace after the receiving is finished. -- === Subject: Re: Serial buffer overflow > I'll try to do an equivalent of the reading loop in Saturn with slight > modifications: > ------------ > on the STK1 i'll have a large empty string, (say standard string object > about 1k or 2k in size -as large as expected data block will be) > In a very tight loop (compact and fast) i'll copy the data from the receive > buffer direcly in to the mentioned string -this will actualy be secondary > buffer, then i'll stop interrupts just for a brief moment so i can reset the > receive pointer/counter (which is somwhere in the mem near the buffer) > enable interrupts, > wait in a loop for pointer to change. > In the end (when sequence of characters is received or timeout period) > length of the string needs to be adjusted to my counter (which counted all > characters transfered). and the string should be ready to use. > I bet this will be a lot faster compared to creating, duplicating, droping, > adding objects on stack and whatever sysRPL does to the characters from > serial buffer when we SREC-v it. > I just hope it will be fast enough (it better be and i hope the issue is > not in the speed how emulator updates hardware related registers and memory > locations in the virtual Saturn. > I'm probably venturing beyond my depth here, but for what it's > worth: > I'm not sure what you mean by an empty string but with a size of > 1K or 2k. Maybe a string with that size field but no characters > following? Is that possible? > Maybe you could start what I mean as an empty string; one with a > size of zero, and then append incoming characters to it using >T$. > Maybe use a virtual stack to start with an empty stack, put the > incoming characters on the stack (one per level), and then when > finished receiving, put an empty string on the stack, and (in a > loop) prepend the characters to it using SWAP >H$ until only the > string is on the stack. Of course follow that up with saving the > string, restoring the saved virtual stack, and then putting the > string on the stack. > Perhaps there's no need to keep a count of the characters as > you're receiving, as long as it can be determined later. > Clearly, the less time used dealing with incoming data while > actually receiving, the better. Re-arranging it into a nice > character string in level one can be done at a more leisurely pace > after the receiving is finished. Oh, no no i love to explain my idea in detail so somone reading this migt give it a try as well :) (and spare us of hard work :) writing few lines) I agree, you are right about EMPTY is not being defined right: empty in my case would be normal proper string, produced with #1000 Blank$ this is actualy a string of 1000 spaces, but since content at that point is of no importance i simply said it was blank. I bet you have guessed this will be a way to allocate memory for secondary buffer at the same time it is on the STK1 and string type -in other words almost ready to use. Counting has 2 functions for one: it is a pointer (location) where next character will be placed in empty string once read from the serial buffer and that count divided by two added 5 is actual length of the string the actual is number that will be stored in the string prologue (object identifier). Pseudo code: So the loop is this { read the number of characters received if non zero { move the characters from the hardware buffer to secondary add chars count to mycounter reset the chars count of primary buffer } } My current reading loop worked exactly as you're sugesting: it produced portions of the received data on the stack, once finished reading it collected the pieces tougether, in all the time and slowness of this world. -that worked nicely but the main loop was too slow (still missed a character or two) i wonder if virtual stack works any faster, but then again DOSRECV doesnt work with virtual stack -right ? rpl is out of the question: ------------------------- i'm very strong about the idea of programming Saturn the way i was explaining i might even add a few options like: #5000 #1000 SSRECV <- this would mean receive buffer would be 5kletters and timeout 1000 ticks so if no additional characters are received within 1000 ticks SSRECV would stop execution and allow the hardware buffer to overflow (timeout mechanism basically) actualy in that particular case it would manage 5255 characters. 5000 would be in the string on the STK1 and 255 in the overflowed buffer how to detect overflow in the primary/secondary buffer: if hardware buffer is 255 then it's overflown in any other case it's not and chars can be srecvd and added to string. i just hope buffer won't get stuck by the time first command will be executed (sysRPL #5000 Blank$) 115200 255 / 10 / 1/X (:-) is about the time it has to do it's job while i write this i took some time to type and measure the execution time for #5000 Blank$ 5000 blanks look very useful and time: 0.0093 is that 9.3 mSec so I think this will do nicely If none of this works i will get mean and programm the ARM to burn the UART or something === Subject: Re: Serial buffer overflow If you are able to figure out what addresses to talk to the UART, you could use saturn assembly and the following scheme: http://www.exzodia.net/hp48/docs/asm.txt It basically allocates all free memory to a string object and then you shrink it at the end based on where D0 points, no counting necessary. Good luck, Aaron > Oh, no no i love to explain my idea in detail so somone reading this migt >give it a try as well :) >(and spare us of hard work :) writing few lines) >I agree, you are right about EMPTY is not being defined right: >empty in my case would be normal proper string, produced with >#1000 Blank$ this is actualy a string of 1000 spaces, but since content >at that point is of no importance i simply said it was blank. >I bet you have guessed this will be a way to allocate memory for secondary >buffer at the same time it is >on the STK1 and string type -in other words almost ready to use. >Counting has 2 functions for one: it is a pointer (location) where next >character will be placed >in empty string once read from the serial buffer and that count divided by >two added 5 is actual length of the string the actual is number that >will be stored in the string prologue (object identifier). >Pseudo code: >So the loop is this > read the number of characters received > if non zero > { > move the characters from the hardware buffer to secondary > add chars count to mycounter > reset the chars count of primary buffer > } >My current reading loop worked exactly as you're sugesting: >it produced portions of the received data on the stack, once finished >reading it collected the pieces tougether, in all the time and slowness of >this world. >-that worked nicely but the main loop was too slow (still missed a character >or two) >i wonder if virtual stack works any faster, but then again DOSRECV doesnt >work with virtual stack -right ? >rpl is out of the question: >------------------------- >i'm very strong about the idea of programming Saturn the way i was >explaining >i might even add a few options like: >#5000 #1000 SSRECV <- this would mean receive buffer would be 5kletters and >timeout 1000 ticks so if no additional characters are received within 1000 >ticks SSRECV would stop execution and allow the hardware buffer to >overflow (timeout mechanism basically) >actualy in that particular case it would manage 5255 characters. 5000 would >be in the string on the STK1 and 255 in the overflowed buffer >how to detect overflow in the primary/secondary buffer: >if hardware buffer is 255 then it's overflown >in any other case it's not and chars can be srecvd and added to string. >i just hope buffer won't get stuck by the time first command will be >executed (sysRPL #5000 Blank$) >115200 255 / 10 / 1/X (:-) is about the time it has to do it's job >while i write this i took some time to type and measure the execution time >for #5000 Blank$ >5000 blanks look very useful and time: 0.0093 is that 9.3 mSec so I >think this will do nicely >If none of this works i will get mean and programm the ARM to burn the UART >or something > === Subject: Re: Serial buffer overflow > If you are able to figure out what addresses to talk to the UART, you > could use saturn assembly and the following scheme: > http://www.exzodia.net/hp48/docs/asm.txt > It basically allocates all free memory to a string object and then you > shrink it at the end based on where D0 points, no counting necessary. > Good luck, > Aaron only i don't need whole memory allocated, 5k or 1k will suffice for me. As a matter of fact that was the idea i was talking about -don't you agree that D0 IS as a matter of fact a counter -i'd better casll it pointer then right ? I would call it a pointer but i am in fact using it as a counter as well as a pointer, because it counts characters at the same time it points to memory location where last character is, or where next will be put -depending on implementation. i downloaded the page you gave me -this are sources from what i saw at first site, the implementations of something very much like i've been describing here, this will be helpful. -althoug i have the code clear in my head, just need to write it down. There are only 2 variables (EQU-s missing) for my code to work and that is: 1. where is the SERIAL buffer in saturn memory of 49G+ 2. where is pointer/counter related to serial buffer I kinda hoped somone could thell me these addresses that apply to 49G+ series (or 49G if it's at the same place) at worst case: i'm gonna send a word or something in to serial buffer and scan trough whole memory to find that word and count related in some way to the length of the word near the buffer === Subject: Re: Serial buffer overflow Correction: -document in your post deals with how to allocate the largest possible amount of memory with/without forcing garbage collect, as i said before i value this and have it on my files, but in this case 5k will do nicely for me === Subject: Re: Serial buffer overflow New update about serial buffer: using DUMP49 (thank you NovaHP -who ever is under that alias) and the method i described in previous post i found out that base of the serial buffer should be about #80319, now i'm on the quest for pointers/counters hehe one anegdote: ------------------- at address #80419 i was still seeing characters from serial buffer so i said nice maybe the buffer is larger than 255 chars after all... then i remembered the addresses are nibbles (not bytes) :( so much for the update === Subject: UART memory buffer and counters information for 49G+ Maybe what i discovered is warm water for you, but since nobody came up with the information... maybe it is the same as on 49G i couldn't tell -can somone confirm this ? here it is: ---------- serial buffer start is at: #80319 number of characters (byte) in the buffer is @ #80519 offset (byte) is @ #8051C ---------------------- what SRECV does is: it gets number of chars characters from 80319+offset, increases offset by number of chars and reduces the number of chars by the number of chars. why reducing -not resetting ? because if in the meantime UART received char or two this count would be bigger, and after substraction exactly that number would remain. (number of fresh chars) knowing all that i will still try to use only the length counter, read entire buffer each time, and reset both counters to zero i hope virtual Saturn can be that fast. === Subject: Re: HP49g+ - NUM.SLV>Solve linear System - Bug or...? > Hi Group... > Today I tried to use the linear system solver with stored matrices > (normally just using the S.SLV LINSOLVE to get exact results or the > NUM.SLV lin. solver with manually entered matrices) and got a really > odd reaction of my calc... > I created the following matrices: > [.9 -.7 0 0 ] > [.7 .7 0 0 ] > A: [0 .7 .5 -.9] > [0 -.7 .5 .5 ] > [0 ] > [1 ] > B: [0 ] > [.5] [.9 -.7 0. 0. ] [.7 .7 0. 0. ] A: [0. .7 .5 -.9] [0. -.7 .5 .5 ] [0. ] [1. ] B: [0. ] [.5] You can do this by recalling the A and B on stack and then doing ->NUM for both of them and finally storing them back. Try it now - Integers are considered symbolic > I stored both into variables and selected the matrices using the > choose-key. A for A and B for B to solve for X in A*X=B. > When I set the cursor to X: and press solve my calc starts working and > working... Something between 5 and ten minutes... After that, it shows > 'Symbolic Matrix' as result for X... > The odd thing is: When I enter the matrices by hand (exactly as shown > above) using edit everything works fine... > Is it a known problem and/or duplicateable? Or did I something really > dumb wrong? > Pascal === Subject: Re: sysRPL: WEIRD finding about COERCE > COERCE does it's job but result is #FF instead of #2 > i like your program, but think my program is faster though Anyway, it works okay now, but I can't reproduce the #FF error. Maybe I didn't set up the UsrRPL levels in the right way. By the way, I'm using an HP49G - if you have a 49G+, the emulator might work in a slightly different way. I'm not sure about that. You're welcome! It's a lot of fun solving RPL puzzles and a great way to learn new programming. Toronto, Canada === Subject: SPLIT challenge > Anyway, it works okay now, but I can't reproduce the #FF error. Maybe > I didn't set up the UsrRPL levels in the right way. By the way, I'm > using an HP49G - if you have a 49G+, the emulator might work in a > slightly different way. I'm not sure about that. > You're welcome! It's a lot of fun solving RPL puzzles and a great way > to learn new programming. > > Toronto, Canada i'm happy you found it working I made a workaround i noticed that on that level i'm using 2 SLICE all the command, called it SLICE2 (slices first 2 characters) which found a good use on other places too :) Agreed... it IS a challenge and i do use G+ Do you know what's the greatest issue? Low confidence in the emulator ! -so each time something get's stuck i wonder if that was an emulation problem or my bad manners: not-checking-short-cuting-kick-ass programming :P can i interest you in another challenge ? -to write a standard SPLIT comand: command would search the STK2 string for occurances of string in STK1, when found string would be splited at that place leaving split-bits on the stack. If not found STK2 should remain intact. STK1 removed. example 1: 1234543556 3 SPLIT would give you: 12 454 556 example 2: 1234543556 454 SPLIT would give you: 123 3556 Thought: in my opinion there is a need for such command, but found none in userRPL nor sysRPL although SREV and similar string commands are there. === Subject: Re: Witnessed in the Laboratory. > Warning: The Following is Hearsay, and may be construed to be > Derogatory to the nation of France, and it's abilities. > However, none of the above is intended in this material. > I assume that like all the people who post anti-France nonsense, you have little or no knowledge of history, no consideration for other people, and no common sense. Try substituting nigger or Jew in your post and you'll immediately get the effect. This crap has no business on comp.sys.hp48. Toronto, Canada === Subject: Re: Witnessed in the Laboratory. > Warning: The Following is Hearsay, and may be construed to be > Derogatory to the nation of France, and it's abilities. > However, none of the above is intended in this material. > Demonstrated In The Lab ! > The Tech powers-up and indicates that I should observe the > 49G+ is Turned on, at Default, and Ready. Then I should note: > ( ON + F3 ) = Unconditional Clear/Restart > ( Cancel ) = Clear Mode > ( F6 ) = Clear > ( Delete ) = Clear Entry > ( Clear ) = Clear > ( B-Arrow ) = Clear Keystroke > ( Undo ) = Clear Operation > After pointing out these features, the Tech looks up and states: > It Must have been designed in _France_, Because at Boot, > the Keyboard offers seven different routes of escape! > ( Insert Drum-Shot here ) > ( I thought I'd pass this awful thing along . . ) What about [TOOL] [F6] = Clear [ON] + [F1] + [F6] = Clear All Memory (conditional) [Right Shift] + [ON] = Off If command line is active [TOOL] gives many different clearing options plus you could use [CUT] () === Subject: Re: Witnessed in the Laboratory. > Warning: The Following is Hearsay, and may be construed to be > Derogatory to the nation of France, and it's abilities. > However, none of the above is intended in this material. > > Demonstrated In The Lab ! > The Tech powers-up and indicates that I should observe the > 49G+ is Turned on, at Default, and Ready. Then I should note: > ( ON + F3 ) = Unconditional Clear/Restart > ( Cancel ) = Clear Mode > ( F6 ) = Clear > ( Delete ) = Clear Entry > ( Clear ) = Clear > ( B-Arrow ) = Clear Keystroke > ( Undo ) = Clear Operation > After pointing out these features, the Tech looks up and states: > It Must have been designed in _France_, Because at Boot, > the Keyboard offers seven different routes of escape! > ( Insert Drum-Shot here ) > ( I thought I'd pass this awful thing along . . ) > Jo Sorry, I didn't get the point of this story. Care to explain? === Subject: HP-49 to Hp-49G+ connectivity? USB. Can an old 49 and a 49+ be directly connected? Can I take an archive off my old 49 and upload it through the PC to the 49+? And are the keys on the 49+ more like the old 49 (stiff and crappy) or more like earlier HPs, like the 48/42/28? === Subject: Re: HP-49 to Hp-49G+ connectivity? > USB. Can an old 49 and a 49+ be directly connected? not directly > Can I take an archive off my old 49 and upload it through the PC to > the 49+? yes > And are the keys on the 49+ more like the old 49 (stiff and crappy) or > more like earlier HPs, like the 48/42/28? not stiff, but seem to miss keystrokes - usually the first on a series after a light sleep wake up (I think it might be rather the ARM slow mode) The plotting interface, CATalog, etc are now snappier and I will never go back to the old 49 === Subject: Re: Why the French guyz rule the HP scene? me again :) i guess there is some influence from Deep Forest as well :) Wouldn't you agree that music has influence on what and how we do something? keep your mind... ...open ! === Subject: Re: Why the French guyz rule the HP scene? @ls219.htnet.hr: > Because of Enigma i guess > i bet they listen to the Enigma's greatest hits when thinking or relaxing > SALUTE to FRANCH guyz > Personally, I think it has something to do with D'Alembert, Laplace, Fourier, Coulomb, Descartes, Fresnel, Legendre, Fermat, Pascal, Biot, Ampere, Navier, and Jean Picard. Stan === Subject: Re: Why the French guyz rule the HP scene? Coulomb 1736 Descartes 1596 Fresnel 1788 Legendre 1752 Fermat 1601 Pascal 1662 Biot 1774 Ampere 1775 Navier 1785 Jean Picard 1620 If only the French could bring back the 18th century Invented the calc? mechanical? and the what about William Wickes? > @ls219.htnet.hr: > Because of Enigma i guess > i bet they listen to the Enigma's greatest hits when thinking or relaxing > SALUTE to FRANCH guyz > > Personally, > I think it has something to do with D'Alembert, Laplace, Fourier, > Coulomb, Descartes, Fresnel, Legendre, Fermat, Pascal, Biot, Ampere, > Navier, and Jean Picard. > Stan === Subject: Re: Why the French guyz rule the HP scene? > Coulomb 1736 > Descartes 1596 > Fresnel 1788 > Legendre 1752 > Fermat 1601 > Pascal 1662 > Biot 1774 > Ampere 1775 > Navier 1785 > Jean Picard 1620 > If only the French could bring back the 18th century > Invented the calc? > mechanical? > and the what about William Wickes? Jersey Devil i admire your enthusiasm (looking up all the names and years) but let's not be cinical or political about it. stuff comes from franch people* I saw this right away and my thought was exatly that So, all i wanted to do is: Openly acknowlege the great work of French colleques, buddies, programmers. Not only for HP sceene -which is in question here, but i pesonaly admire them in many ways. Even if they don't support me as much why do you have a hard time living with that ? P.S. next week if french people let go of me (it this threed dies) we will be takling about Taiwanese workers in the factory which are producing this beautyfull devices as we speak === Subject: Re: Why the French guyz rule the HP scene? I would much rather hear about your music preference when you programm (or use) your favorite HP tool. why don't you share some thought (crative thinking) instead of being cynical and falling in to denial !? ...love and peace === Subject: Re: Why the French guyz rule the HP scene? > Personally, > I think it has something to do with D'Alembert, Laplace, Fourier, > Coulomb, Descartes, Fresnel, Legendre, Fermat, Pascal, Biot, Ampere, > Navier, and Jean Picard. > Stan Franch scientists on one hand and Germans technitians/engeneers right next to it, spy-ing and making all discoverys work in real life. :-) Oh, but i don't think this new people care much about older population of dead scientists (although they are all great men) Jean Luc Picard is my personal favorite === Subject: Re: Why the French guyz rule the HP scene? > Oh, but i don't think this new people care much about older population of > dead scientists > (although they are all great men) > Jean Luc Picard is my personal favorite > Me too. Seriously, a French humorist once said about the French: On n'a pas de p.8etrole, mais on a des id.8ees! (We don't have oil, but we have ideas!) I think that sums up the French people quite nicely. === Subject: Re: SD-card formatting: cluster size? > I can confirm approx 248000 KB free with just formatted card: so it > seems >> it > is the filer to uncorrectly show the amount of free disk space... > Now that you mention it, I think I've seen other posts about the filer > showing only 128KB free on larger cards. >> Do you mean 128000 KB? >> Anyway what I mean is that maybe the Filer shows >> uncorrectly the frre disk space, BUT indeed I WAS WRONG! >> I've used the Analyze function from Defrag built-in in Windows2000 and it >> says that the cluster size is 2 KB! What a great mistery! Microsoft >> clearly >> saysy at http://www.microsoft.com/resources/documentation/Windows/XP/all/reskit/en-us /Default.asp?url=/resources/documentation/windows/xp/all/reskit/en-us/prkc_f >> il_lxty.asp that for a 256MB drive the cluster size will be 4KB with >> FAT-16... > I guess someone at Microsoft is smart enough to figure out that with > FAT16, using 2K clusters would waste about half of a 256MB drive. But > sometimes their software gives strange results. >> Is the calc formatting the SD-card in FAT32? > I haven't heard of that, but who knows? I don't have an SD card big > enough to make FAT32 worthwhile. Actually, I not even sure that FAT32 > is supported for drives smaller than 512MB. > Is your 49g+ running on ROM revision 1.23? >> Can the pre-formatting of the >> SD-card influence further formats? > Yes. For example, when I was fooling around, I trashed the card's boot > record, and experimentally found that neither the Windows 98SE format > nor the 49g+ format would change the number of sectors recorded in the > boot record. If I recall correctly, I temporarily changed my 128KB > card to a 16KB card. I ended up tweaking it with the disk editor. It > was fun though. I don't know whether a DOS format would've gotten it > straightened out. Nononononono ... boot record contains partition table, no actual FAT data. All format utilities just format one partition, at whatever size is reported in the boot record. To un-do what you did, you'd have to use fdisk. Except DOS fdisk doesn't do USB-attached disks (AFAIK). So you'd have to use Linux fdisk =) -MrM === Subject: HP 48GX compatability? I am looking to take the memory out of the broken body of my HP 48GX and put it into a new body. I am having much difficulty locating a model of this type, therefore does anyone know if any other (read: newer and easier to purchase) models are compatable with said memory. 3506 === Subject: Re: Do You Find your HP Indispensable ? I had a TI 58C - great calculator but quite complex to use. The biggest problem was that the keys were unreliable (e.g. you press the 2 once and you see three 2's on the display). Furthermore, getting the right result with complex caclulations and many parenthesis is not obvious (I usually rechecked a coupled of times). I bought a HP 15C and I got a very powerful and much easier-to-use machine. With the total reliability of the keyboard and the simplicity of RPN, I always got the right result first time. It was a truly great calculator (nothing to compare with unreliable toys like the 49G+). > The last TI I used needed to change modes to go from arithmetic to > statistics, and then there was no way to go back to arithmetic without > losing the statistical data, except by writing it down and re-entering it. > I threw it out and bought an HP. > I enjoyed reading on the topics of Most, and Least favorite > calculators. Now, I'm inclined to ask you, why is that? > This HP is quite handy for dialing-in on electrical current, or > group delays, and for actually retaining memory between sessions. > It is just idle wonderment, but would you care to comment on best > suitability or favorite usages ? > Is your HP nearly indispensable in a given style or model line? > How come is that? > Have other brands assured you, and then failed your requirment? === Subject: Re: HP49G, Cus menu problems > you're my hero =D > it now works, the parenthesis appear just fine(i aded just like you > said in the 1st place). > btw, there was a shortcut to change modes, i can't remember what it > was. [MODE] [+/-] [ENTER] or | OK | eg. [ F6 ] > and another thing, since it appears this calc is made to use RPN > mainly, where can i find a RPN users manual for hte hp49, since the > manual and the AUG comes in algebraic and i don't have too much time > to fiddle with it www.hp.com go to hp 49g+ and find manual there The new manual for the plus is much better (use the large one) << >> === Subject: Re: HP49G, Cus menu problems > [MODE] [+/-] > [ENTER] or | OK | eg. [ F6 ] hmmm my calc doesn't change choose values with +/-, i can only change with F2, this is weird... another thing, how can i enter the command directly in RPN? i type {test, basis } STO CST (the STO appears as written) enter and it gives a STO bad argument type i tried putting it into stack, then writing CST sto=> and it gives the same error. Also, when i get back in ALG mode, now i have 2 things in the cus menu: test and basis, why is it creating an item for the things i put , shouldn't it be the label for the command? (i put the , as instructed, edited in RPN) === Subject: Re: HP49G, Cus menu problems >> [MODE] [+/-] >> [ENTER] or | OK | eg. [ F6 ] > hmmm my calc doesn't change choose values with +/-, i can only change > with F2, this is weird... Yes it is, unless you tried [+] and [-] instead of [+/-] key [PVN] > another thing, how can i enter the command directly in RPN? > i type {test, basis } STO CST (the STO appears as written) > enter and it gives a STO bad argument type > i tried putting it into stack, then writing CST sto=> and it gives the > same error. {test, basis } 'CST' [STO] [NVP] > Also, when i get back in ALG mode, now i have 2 things in the cus > menu: test and basis, why is it creating an item for the things i put > , shouldn't it be the label for the command? > (i put the , as instructed, edited in RPN) {name3 {NORMAL LEFT.SHIFT RIGHT.SHIFT }}} [VNP] - little mixed up today..)- === Subject: Re: good time to buy 49g+? and screen protectors > About to upgrade my HP 49G to a 49G+.. before I do, are there any good > reasons not to do so right now? I think the Qonos is a slightly different > class (and certainly not permissible in exams *grin*) so no reason for me > to wait for that.. X Are you sure a 49 series calculator with CAS is allowed? The Flash provides room for hundreds of pages of text I will certainly get a Qonos (or two...) === Subject: HP41 HP25 I want to buy old calculators as HP 25 and hp41. I want to know if someone is interested do sell and ship it to Brazil. Please answer to josergio@ig.com.br === Subject: Re: RPN Calc button locations & shapes ~ HP 33S > | > | I want to get a hold of one of the ones from the Walmart pre-release. > Will > | Pay $45 for one. > | > What is different about the pre-release calculators? There are differences in (at least one of the) the ALG mode features. We have not found any differences in RPN. - === Subject: Re: RPN Calc button locations & shapes ~ HP 33S > | > | I want to get a hold of one of the ones from the Walmart pre-release. > Will > | Pay $45 for one. > | > What is different about the pre-release calculators? They are more worthy as collectables {} === Subject: Re: How would you answer this question? > Ask him, whether he can imagine the difference between driving a Toyota and > a Mercedes Benz. > (Both get you from A to B though ...) I like Toyota better than Benz, but I vastly prefer HP over TI. Hmmmmm..... > MM > Chris W <1qazse4XYZ@cox.net.removeXYZ> schrieb im Newsbeitrag > In reference to my petition at http://hp15c.org, someone sent me this > message; > Well, I am working for TI in the calculator division and interested to > know what is special about HP calculators and why you want HP back > [rather] th[a]n using TI. > That is a loaded question, how would you answer it? > -- > Chris W > Bring Back the HP 15C > http://hp15c.org > Not getting the gifts you want? The Wish Zone can help. > http://thewishzone.com === Subject: Re: How would you answer this question? > (Both get you from A to B though ...) > I like Toyota better than Benz, but I vastly prefer HP over TI. Hmmmmm..... what about flying business class vs economy class? :) === Subject: Re: How would you answer this question? > (Both get you from A to B though ...) > I like Toyota better than Benz, but I vastly prefer HP over TI. > Hmmmmm..... > what about flying business class vs economy class? :) I flew *first class* last month on a transcontinental flight----what a great deal! Now I am totally spoiled! But my absolute favorite flight of all time was last year, in a Beechcraft twin turboprop--13 passenger plane. (They have really nice leather seats.) Why so great? Flies at 275 kts, at just the right altitude so that when you have to detour around a huge meso-scale thunderhead cluster, you are right at the perfect level to see the clouds up close---absolutely beautiful and stunning. I used my hp48 as a timer during takeoff (and flight) to track our speed. with 2 x 1300 hp, and only 4 passengers aboard, she makes a jet airplane seem slow. We were off the ground in something like 7 seconds (or was it 4!) . faaaaast. So far, I have only been told to turn off my hp48 once during a flight. Usually, they don't seem to consider them to be electronic devices . === Subject: Re: How would you answer this question? >> (Both get you from A to B though ...) >> I like Toyota better than Benz, but I vastly prefer HP over TI. > Hmmmmm..... > what about flying business class vs economy class? :) TI & HP are both economy class today anything cheaper and you will be flying cargo Qonos = Business Class (no color TV?) Well at least the fuel supply is still intercontinental eg. lasts more than a month - waiting for Xmas again like a child this year... === Subject: Re: How would you answer this question? I was only checking sonone's taste (Merc vs Toy) :o) > TI & HP are both economy class today > anything cheaper and you will be flying cargo > Qonos = Business Class (no color TV?) > Well at least the fuel supply is still intercontinental > eg. lasts more than a month > - waiting for Xmas again like a child this year... === Subject: Re: How would you answer this question? I was only checking sonone's taste (Merc vs Toy) :o) > TI & HP are both economy class today > anything cheaper and you will be flying cargo > Qonos = Business Class (no color TV?) > Well at least the fuel supply is still intercontinental > eg. lasts more than a month > - waiting for Xmas again like a child this year... === Subject: Re: Eql+ - a different approach > Hi folks, > soon available on hpcalc.org: > A flexible and easy to use browser which can be expanded by your own > equations and pictures. > TreeBrowser supplies an easy to use graphical environment (like the > filer) for collections of formulas and shows them in a tree structure. > The equations are passed to the built-in SOLVER of the HP 49G/49G+ for > solving. > So you can easily build your own favavorite collection of formulas. > As an example data set the original Equation Library, including all > pictures, from Hewlett-Packard is available for TreeBrowser. > Until it shows up on hpcalc.org it can be downloaded from my website > at > http://www.andreasmoeller.gmxhome.de/ I couldn't download it. Your website showed a page not found, error 404 when I tried. === Subject: Re: Eql+ - a different approach You have to use the Internet Explorer at the moment, the site is build with Word 2000, I try to change it asap. Anndreas > Hi folks, > > soon available on hpcalc.org: > A flexible and easy to use browser which can be expanded by your own > equations and pictures. > > TreeBrowser supplies an easy to use graphical environment (like the > filer) for collections of formulas and shows them in a tree structure. > The equations are passed to the built-in SOLVER of the HP 49G/49G+ for > solving. > So you can easily build your own favavorite collection of formulas. > > As an example data set the original Equation Library, including all > pictures, from Hewlett-Packard is available for TreeBrowser. > > Until it shows up on hpcalc.org it can be downloaded from my website > at > http://www.andreasmoeller.gmxhome.de/ > I couldn't download it. Your website showed a page not found, error > 404 when I tried. === Subject: Re: HP49G+ SD-Card: Filer6-Problem mak schrieb im Newsbeitrag > mak schrieb im Newsbeitrag > Hi out there, when I try to create a directory on a SD-Card > with Filer6 (from Prof.Rautenberg) I always get > the message XLIB 688 49 Error: Too Few Arguments > and no directory is created. Can someone explain to me what to do? > In Filer6 Press '3' to select the SD card. ( 3 w/o ticks ...) > Then press LS DEF, which pops up a menu. > Check the 'Directory' item and enter a name for your folder. > Press the OK soft key. > That's it. > Thats exactly the way I did it ... and after pressing o.k. I always > get the ERROR-Message XLIB 688 49 Error: Too Few Arguments. (I use > ROM 1.23). What could be the problem ? I can't find any prerequisites other than ROM 1.23 in the Filer6 doc. What libraries have you installed? Does it help, if you remove them all and retry? If so, add them successively back one by one, until you get this error again. A more radical way may be a cold reset with memory clearing. Of course, don't forget to backup your data before doing that For me - and I think most of the others here - Filer6 worked out of the box, without any problems. MM > MAK === Subject: Re: Manually assembling library > 45 /T/ Ok, I can't compile this with MASD - it errors at the above statement (at the slashes). I'm daft at ASM syntax === Subject: Re: Manually assembling library > 45 /T/ > Ok, I can't compile this with MASD - it errors at the above statement (at > the slashes). I'm daft at ASM syntax > Regarding my previous answer: MASD on the MetaKernel did accept the statement CON(5) &-.label where .label was somewhere in a preceding line. So.. how do I enter it in a 49? Jean-Yves? Cyrille? ; I was able to compile it on a 48GX+MK I just had to change the *.label GOIN5 .label statements to CON(5) $0 === Subject: Re: Manually assembling library > 45 /T/ > Ok, I can't compile this with MASD - it errors at the above statement (at > the slashes). I'm daft at ASM syntax > Mh there are a few errors.. /text/ is written with a '/' because the actual character to be used in MASD is char 162 (dec) and you'll need to invoke the char application to enter it. There's one instance of GOIN(5) - should be GOIN5 There's one more thing that I can't solve right away (I thought I was able to - but I can't find any source any more): the CON(5) &-.label instruction is not accepted - yet it is one instance where I cannot replace it with GOIN5 as that is equivalent to .label-& Moreover, you'll need a CRC patcher - I have that somewhere, too.. I'll keep on looking.. Werner === Subject: Re: Can emu48CE do BEEP on iPAQ? If you go to Eric Rechlin's www.hpcalc.org site and go to the PC Emulators area, you will see a reference to iPaq 49G KML 1.0 by Vincent Weber, which does two things. First, it replaces the HP49 key layout with a wide-ENTER-key HP48 key layout (which I posted back in 2001), but then fixes the top-row problem by shifting the image downward and adjusting the KML file accordingly. Give it a look. If you get the PDA to emit any sound when the emuCE BEEP function is executed, I'd love to hear about it. Jake Schwartz > Hi Jake, > since PDA's have no Speaker in the usual sense, they cannot beep. > i.e. there is no function in the Windows CE API. > Of course they do have sound hardware. > They can play waveforms. A beep can be simulated with any *.wav file, > applied to the function for Windows CE API >= Version 2.0. > I haven't looked into the emu48CE sources yet, whether there is support for > that. > If not, it shouldn't be too hard to implement. > Now a question to you, since you are using a Jornada 548, as I do. > I have a KML script named Jornada 49G by Chris Roper. > This is for me the nicest GUI for the Jornada. > Unfortunately I have a weird vertical offset, so that the topmost row of > keys and the display are cut, while there > is a sparse row at the bot of the screen. > Do you know this KML and does it work for you? > I'm absolutely clueless with KML, maybe it's a snap for someone who knows > what he/she's doing. > Any hints? > MM === Subject: Re: Can emu48CE do BEEP on iPAQ? Jake Schwartz schrieb im Newsbeitrag > If you go to Eric Rechlin's www.hpcalc.org site and go to the PC > Emulators area, you will see a reference to iPaq 49G KML 1.0 by > Vincent Weber, which does two things. First, it replaces the HP49 key > layout with a wide-ENTER-key HP48 key layout (which I posted back in > 2001), but then fixes the top-row problem by shifting the image > downward and adjusting the KML file accordingly. Give it a look. > If you get the PDA to emit any sound when the emuCE BEEP function is > executed, I'd love to hear about it. > Jake Schwartz thank you! It works perfectly on my Jornada 548. W.r.t. the BEEP function: Maybe my first statement was somewhat overenthusiastic, since I thought of a simple Message Beep or Alert Signal at first. That should be fairly simple, because it can be simulated by playing a *.wav soundfile. If that's all what you want, that shouldn't be hard. BUT THEN, I remembered the two parameters to BEEP, and . Now things might become tricky, if you want a *correct* implementation, which is capable of playing (synthetic) music via BEEP. There are two solutions I can think of, probably there are much more... (i) A set of pre-sampled (*.wav) soundfiles, according to a canonical scale, which are algorithmically altered at runtime. (ii) A complete dynamic, algorithmic, in-memory synthesis of a wave form (maybe square for example) Either could serve as an input to WIN CE's API, which is capable of playing in-memory bytestreams. I don't like (i) too much, since it seems less flexible and hardly less complex than (ii). So I think (ii) is worth being investigated, and it would definitely be a fun challenge. Unfortunately, I'm not an expert for sound synthesis and the like. But I remember a C library, which fell into my hands many years ago ... Maybe I can find and leverage this one. I'll definitely post here, if/when I find something. -- MM > Hi Jake, > since PDA's have no Speaker in the usual sense, they cannot beep. > i.e. there is no function in the Windows CE API. > Of course they do have sound hardware. > They can play waveforms. A beep can be simulated with any *.wav file, > applied to the function for Windows CE API >= Version 2.0. > I haven't looked into the emu48CE sources yet, whether there is support for > that. > If not, it shouldn't be too hard to implement. > Now a question to you, since you are using a Jornada 548, as I do. > I have a KML script named Jornada 49G by Chris Roper. > This is for me the nicest GUI for the Jornada. > Unfortunately I have a weird vertical offset, so that the topmost row of > keys and the display are cut, while there > is a sparse row at the bot of the screen. > Do you know this KML and does it work for you? > I'm absolutely clueless with KML, maybe it's a snap for someone who knows > what he/she's doing. > Any hints? > MM Wayne: The on thing the NDA applied to was about a 1 to 2 hour (as I recall) presentation and Q/A session where in HP shared some of their company marketing and product plans. All the rest of the sessions were open. John > Will there be a non-disclosure agreement to sign this year? > -- > 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 Wayne: The on thing the NDA applied to was about a 1 to 2 hour (as I recall) presentation and Q/A session where in HP shared some of their company marketing and product plans. All the rest of the sessions were open. John > Will there be a non-disclosure agreement to sign this year? > -- > 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 > Will there be a non-disclosure agreement to sign this year? They haven't provided any details as to what HP will present, so it's anybody's guess at this point. I wouldn't count on whether or not HP Jake Schwartz === Subject: Re: TI89 Titanium -- opinions? > I didn't know you could code in C directly in a TI-89 > You know you can't - there's really no need to be a wiseguy. > Yes, you can. > http://darkclem.free.fr/ > Scroll down quite far to find it. Are all these French people brilliant? > It seem like all the great software comes from them! 8-) > Lake Here's another one http://mxm.ticalc.org/ontic.html Dunno if it works as I haven't tried either of these compilers. === Subject: Re: TI89 Titanium -- opinions? > I didn't know you could code in C directly in a TI-89 You know you can't - there's really no need to be a wiseguy. > Yes, you can. > http://darkclem.free.fr/ > Scroll down quite far to find it. Are all these French people brilliant? > It seem like all the great software comes from them! 8-) > Lake > Here's another one > http://mxm.ticalc.org/ontic.html > Dunno if it works as I haven't tried either of these compilers. The only problem with that one is that you need to send the source back to a PC to be compiled with TIGCC and then send the executable back to the TI. With the other C, there's no need to use a PC at all. Lake === Subject: Re: TI89 Titanium -- opinions? > You know what i don't like about you ? > It is your superiority complex and the fact that you think that you > are smarter than everyone. You know...this post of yours did almost make me send one very angry email to you, but I counted to 100, and will try to be very calm answering it. He doesn't need to used a TI89 Titanium to know that it is slow. That's false. You cannot speak about a thing you haven't tried. You don't have the knowledge we speak of here in your genes, and it can't be telepatically transferred. What is so hard to understand about the need for first hand experience to *know* something? It's not enough to *think* you now. It's not enough for it to be plausible or even probable. How about trying being sure on for size? > I thought that your failure with your advanced handheld project would > have made you more humble or less arrogant but obviously i was wrong. See, this was what made me very angry. You, my friend, do not know what you're talking about here. Our (no, I wasn't alone in that adventure) project was not a failure. It wasn't even close. See, I still have invested a good deal of the *money* I made from the sale of the IP, and I am still co-owner (and founder) of a company here in Denmark (Realconsult). Most back then were sold of to Toshiba, whom we worked very close with designing around the XScale (try taking a peek at their PDA line-up a couple of years back, and see if anything seems familiar) I won't go into much detail as to why we stopped, but it's centered around one of my friends getting killed in a motorbike accident on N.9frburgring and me getting divorced and having to fight for my kid. > If you think that both HP and the comp.sys.hp48 are losing their > quality,what are you doing here ? I don't really know. Nostalgia? You haven't noticed that 90% of the '95-'00 regulars aren't here anymore? You don't think HP is dropping on the quality scale and have been doing for some years? You sincerely mean that the tone of c.s.hp48 today is reminiscent of what it was 6-8 years back? > And contrary to what you said i do know exactly what i am talking. So you have used the Titanic - in what areas is it slower than a HW2 TI89? 3.00). > So unless i manage to always get TI68k with problems,i do know cleary > what i am talking. I agree with you about the original TI92/HP48 business - a sad story. And the TI92 was buggy as h***. > Your statements almost prove my point that the HP49G+ is generally > faster than the TI68K. Yes, we agree on this. I don't agree with the statement that the TI89 HW2 is > For example when you say that the TI is faster than the HP49G+ for > numeric sums. > Are you talking of the built-in command or of all ways to do a sum ? Built-in summation function. > Also how the hell have you manage to get your TI being 2-4 times > faster than your HP49G+ when my HP49G was faster than my TI92+ for > that ? Is your HP49G faster than the TI92+ for integer arithmetic? I don't have a TI92+ at hand to try it, but I can provide some numbers for my H.8149G+ and my HW2 TI89; Example: TI89: :Prgm :For x,0,2000 :EndFor :EndPrgm runs in 14.789474 seconds on my TI89. :Prgm :For x,0,2000 :2528578459656*103 :EndFor :EndPrgm runs in 18.736842 seconds on my TI89, which means 507 integer multiplications pr. second. HP49G+: << #1d #1000d START NEXT >> runs in 0.4371 seconds on my HP49G+. << #1d #1000d START 2528578459656 103 * DROP NEXT >> runs in 5.6277 seconds on my HP49G+, which means 193 integer multiplications pr. second (DROP adds just a very slight overhead). Similar tests for various (very much larger) integers: (210-digit int)+(231-digit int): HP=257/s & TI=404/s (210-digit int)-(231-digit int): HP=243/s & TI=328/s (145-digit int)*(97-digit int): HP=10/s & TI=27/s (145-digit int)/(97-digit int): HP=20/s & TI=26/s (6-digit int)^(2-digit int): HP=7/s & TI=22/s Sqrt(6-digit int): HP=10/s & TI=40/s Result: the TI89 HW2 is 2-4 times faster than the HP49G+ for basic integer arithmetic. === Subject: Re: TI89 Titanium -- opinions? I won't apologize for my first statement because i really think that you have a superiority complex and that you are quite arrogant but i sincerely apologize for this: > I thought that your failure with your advanced handheld project would > have made you more humble or less arrogant but obviously i was wrong. > See, this was what made me very angry. You, my friend, do not know what > you're talking about here. Our (no, I wasn't alone in that adventure) > project was not a failure. It wasn't even close. See, I still have > invested a good deal of the *money* I made from the sale of the IP, and I am > still co-owner (and founder) of a company here in Denmark (Realconsult). > Most back then were sold of to Toshiba, whom we worked very close with > designing around the XScale (try taking a peek at their PDA line-up a couple > of years back, and see if anything seems familiar) > I won't go into much detail as to why we stopped, but it's centered around > one of my friends getting killed in a motorbike accident on N.9frburgring and > me getting divorced and having to fight for my kid. I have gone too far here and again i apologize. > He doesn't need to used a TI89 Titanium to know that it is slow. > That's false. You cannot speak about a thing you haven't tried. You don't > have the knowledge we speak of here in your genes, and it can't be > telepatically transferred. What is so hard to understand about the need for > first hand experience to *know* something? It's not enough to *think* you > now. It's not enough for it to be plausible or even probable. How about > trying being sure on for size? Here then it is you who don't know what you are talking. If you have spent as much time as me on TI calculators and following their evolution as close as i have you would know that: *Since the TI89,TI has never tried to improve the speed of its TI68k calculators by software optimisation *TI68k generally get faster by significant hardware modification such as faster C.P.U or fixing of a hardware problem such as the second mode. So knowing this and knowing that the TI89 Titanium is to the best a TI89 Hw2 with more Flash ROM or to the worse a TI89 version of the Voyage 200,i can easily conclude that the TI89 Titanic is not faster than the TI89 HW2. > If you think that both HP and the comp.sys.hp48 are losing their > quality,what are you doing here ? > I don't really know. Nostalgia? You haven't noticed that 90% of the '95-'00 > regulars aren't here anymore? You don't think HP is dropping on the quality > scale and have been doing for some years? You sincerely mean that the tone > of c.s.hp48 today is reminiscent of what it was 6-8 years back? The HP quality has decreased anyone will agree with that. Your second statement was attack toward me so what did you expect me to do ? To agree with you perhaps ? > And contrary to what you said i do know exactly what i am talking. > So you have used the Titanic - in what areas is it slower than a HW2 TI89? > 3.00). I don't and will never have a Titanic or a Voyage 200 or any supossedly new TI68k which brought nothing besides more user Flash ROM. And i don't need to get a TI89 Titanium to make accurate predictions about it. The A.M.S 3.00 is certainly there just to handle hardware modifications such as the USB,to fix some bugs and to add some new limitations or to remove some old ones,nothing more nothing. > So unless i manage to always get TI68k with problems,i do know cleary > what i am talking. > I agree with you about the original TI92/HP48 business - a sad story. And > the TI92 was buggy as h***. > Your statements almost prove my point that the HP49G+ is generally > faster than the TI68K. > Yes, we agree on this. I don't agree with the statement that the TI89 HW2 is Where did i say that the HP49G+ is always faster than the TI89 HW2 ? > For example when you say that the TI is faster than the HP49G+ for > numeric sums. > Are you talking of the built-in command or of all ways to do a sum ? > Built-in summation function. This is possible as the built-in function doesn't seem optimised. However with RPN program this is another matter. > Also how the hell have you manage to get your TI being 2-4 times > faster than your HP49G+ when my HP49G was faster than my TI92+ for > that ? > Is your HP49G faster than the TI92+ for integer arithmetic? Yes it was but unfortunately i can't do test anymore as my HP49G has been stolen a several years ago. I still have a TI92+ HW1 though. >I don't have a > TI92+ at hand to try it, but I can provide some numbers for my H.8149G+ and my > HW2 TI89; > Example: > TI89: > :Prgm > :For x,0,2000 > :EndFor > :EndPrgm > runs in 14.789474 seconds on my TI89. > :Prgm > :For x,0,2000 > :2528578459656*103 > :EndFor > :EndPrgm > runs in 18.736842 seconds on my TI89, which means 507 integer > multiplications pr. second. > HP49G+: > << #1d #1000d START NEXT > runs in 0.4371 seconds on my HP49G+. > << #1d #1000d START 2528578459656 103 * DROP NEXT > runs in 5.6277 seconds on my HP49G+, which means 193 integer multiplications > pr. second (DROP adds just a very slight overhead). Very curious but if what you say is true than the HP49G+ is more disapointing than i thought. > Similar tests for various (very much larger) integers: > (210-digit int)+(231-digit int): HP=257/s & TI=404/s > (210-digit int)-(231-digit int): HP=243/s & TI=328/s > (145-digit int)*(97-digit int): HP=10/s & TI=27/s > (145-digit int)/(97-digit int): HP=20/s & TI=26/s > (6-digit int)^(2-digit int): HP=7/s & TI=22/s > Sqrt(6-digit int): HP=10/s & TI=40/s > Result: the TI89 HW2 is 2-4 times faster than the HP49G+ for basic integer > arithmetic. Your first comparaison give you a point but all your other comparaisons give a clear advantage to the HP. So what is exactly your point ? I still don't see how you manage to conclude that the TI is faster than the HP when almost all your tests show the opposite. Btw try this on your TI89: for i,1,3000:0->#(k&string(k)):endfor and tell me what happen. But you should have a way to launch this other sequence: for i,1,3000:delvar #(k&string(k)):endfor > === Subject: Re: TI89 Titanium -- opinions? > I have gone too far here and again i apologize. I appreciate it. You unfortunately hit a sore spot there, so I reacted a good deal more defensive than was necessary - I apologize as well. > *Since the TI89,TI has never tried to improve the speed of its TI68k > calculators by software optimisation > *TI68k generally get faster by significant hardware modification such > as faster C.P.U or fixing of a hardware problem such as the second > mode. predictions are correct though. > So knowing this and knowing that the TI89 Titanium is to the best a > TI89 Hw2 with more Flash ROM or to the worse a TI89 version of the > Voyage 200,i can easily conclude that the TI89 Titanic is not faster > than the TI89 HW2. I never said I expected the Titanic to be faster than a HW2 TI89, but I expect it to be very close performance wise, if not identical with it. > Your second statement was attack toward me so what did you expect me > to do ? It wasn't meant as an attack on you personally, so I'm sorry if you got stung. > Yes, we agree on this. I don't agree with the statement that the TI89 HW2 is > Where did i say that the HP49G+ is always faster than the TI89 HW2 ? It wasn't your statement I was referring to. You just seemed to back Veli-Pekkas initial statement (The Titanic is f*ing slow), which I extrapolated into meaning the Titatinic was slower than the HP49G+ without exception. With the above sentence I tried to explain why I entered this discussion in the first place - namely because I think it's too simple an answer to give a guy who asks sincerely. But, what the heck - this is an HP newsgroup, what should he expect when asking something like that here :-) > This is possible as the built-in function doesn't seem optimised. > However with RPN program this is another matter. Yes, for numeric sums there is possibly some performance to gain. > Similar tests for various (very much larger) integers: > (210-digit int)+(231-digit int): HP=257/s & TI=404/s > (210-digit int)-(231-digit int): HP=243/s & TI=328/s > (145-digit int)*(97-digit int): HP=10/s & TI=27/s > (145-digit int)/(97-digit int): HP=20/s & TI=26/s > (6-digit int)^(2-digit int): HP=7/s & TI=22/s > Sqrt(6-digit int): HP=10/s & TI=40/s > Result: the TI89 HW2 is 2-4 times faster than the HP49G+ for basic integer > arithmetic. > Your first comparaison give you a point but all your other > comparaisons give a clear advantage to the HP. You misunderstand my sorry attempt at providing units - in the above higher numbers are better, as the unit is s^-1 (10/s -> 10 per second -> 10 Hz if you will). The TI is faster in *all* the above tests. For comparison, I have just completed the same set of tests on my HP49G: (210-digit int)+(231-digit int): HP49G=128/s (210-digit int)-(231-digit int): HP49G=128/s (145-digit int)*(97-digit int): HP49G=4.5/s (145-digit int)/(97-digit int): HP49G=8/s (6-digit int)^(2-digit int): HP49G=4/s Sqrt(6-digit int): HP49G=3.5/s The above means (for this particular test) that the HP49G+ is 2-3 times faster than the HP49G for integer arithmetic, and consequently that the TI89 is 3-11 times faster than the HP49G. > Btw try this on your TI89: > for i,1,3000:0->#(k&string(k)):endfor You mean for i,1,3000:0->#(k&string(i)):endfor right? > and tell me what happen. I got a memory error after a while (and one sluggish TI89. === Subject: Re: TI89 Titanium -- opinions? >> If you think that both HP and the comp.sys.hp48 are losing their >> quality,what are you doing here ? > I don't really know. Nostalgia? You haven't noticed that 90% of the '95-'00 > regulars aren't here anymore? You don't think HP is dropping on the quality > scale and have been doing for some years? You sincerely mean that the tone > of c.s.hp48 today is reminiscent of what it was 6-8 years back? In my experience, all of Usenet has been going downhill for more than 6-8 years. Though uncorrelated, HP's quality is also headed downwards. A bientot Paul -- Paul Floyd http://paulf.free.fr (for what it's worth) Surgery: ennobled Gerald. === Subject: Re: TI89 Titanium -- opinions? I've been gone from this group for awhile... if the pre 21st century regulars are gone, where did they go??? ICQ? hpcalc.org is still updated currently - does everybody just check around there? Just curious.. >> If you think that both HP and the comp.sys.hp48 are losing their >> quality,what are you doing here ? > I don't really know. Nostalgia? You haven't noticed that 90% of the > '95-'00 regulars aren't here anymore? You don't think HP is dropping on > the quality scale and have been doing for some years? You sincerely > mean that the tone of c.s.hp48 today is reminiscent of what it was 6-8 > years back? > In my experience, all of Usenet has been going downhill for more than 6-8 > years. Though uncorrelated, HP's quality is also headed downwards. > A bientot > Paul -- Shawn Legrand Carpe Noctem! Carpe Diem! Hieros Gamos! splegrand@yahoo.com === Subject: Re: TI89 Titanium -- opinions? I double Timit? Hassan Based on your posts here I'd be more than happy not to see you posting again. You've got problems Without any respect: Reth > You know what i don't like about you ? > It is your superiority complex and the fact that you think that you > are smarter than everyone. > You know...this post of yours did almost make me send one very angry email > to you, but I counted to 100, and will try to be very calm answering it. > He doesn't need to used a TI89 Titanium to know that it is slow. > That's false. You cannot speak about a thing you haven't tried. You don't > have the knowledge we speak of here in your genes, and it can't be > telepatically transferred. What is so hard to understand about the need for > first hand experience to *know* something? It's not enough to *think* you > now. It's not enough for it to be plausible or even probable. How about > trying being sure on for size? > I thought that your failure with your advanced handheld project would > have made you more humble or less arrogant but obviously i was wrong. > See, this was what made me very angry. You, my friend, do not know what > you're talking about here. Our (no, I wasn't alone in that adventure) > project was not a failure. It wasn't even close. See, I still have > invested a good deal of the *money* I made from the sale of the IP, and I am > still co-owner (and founder) of a company here in Denmark (Realconsult). > Most back then were sold of to Toshiba, whom we worked very close with > designing around the XScale (try taking a peek at their PDA line-up a couple > of years back, and see if anything seems familiar) > I won't go into much detail as to why we stopped, but it's centered around > one of my friends getting killed in a motorbike accident on N.9frburgring and > me getting divorced and having to fight for my kid. > If you think that both HP and the comp.sys.hp48 are losing their > quality,what are you doing here ? > I don't really know. Nostalgia? You haven't noticed that 90% of the '95-'00 > regulars aren't here anymore? You don't think HP is dropping on the quality > scale and have been doing for some years? You sincerely mean that the tone > of c.s.hp48 today is reminiscent of what it was 6-8 years back? > And contrary to what you said i do know exactly what i am talking. > So you have used the Titanic - in what areas is it slower than a HW2 TI89? > 3.00). > So unless i manage to always get TI68k with problems,i do know cleary > what i am talking. > I agree with you about the original TI92/HP48 business - a sad story. And > the TI92 was buggy as h***. > Your statements almost prove my point that the HP49G+ is generally > faster than the TI68K. > Yes, we agree on this. I don't agree with the statement that the TI89 HW2 is > For example when you say that the TI is faster than the HP49G+ for > numeric sums. > Are you talking of the built-in command or of all ways to do a sum ? > Built-in summation function. > Also how the hell have you manage to get your TI being 2-4 times > faster than your HP49G+ when my HP49G was faster than my TI92+ for > that ? > Is your HP49G faster than the TI92+ for integer arithmetic? I don't have a > TI92+ at hand to try it, but I can provide some numbers for my H.8149G+ and my > HW2 TI89; > Example: > TI89: > :Prgm > :For x,0,2000 > :EndFor > :EndPrgm > runs in 14.789474 seconds on my TI89. > :Prgm > :For x,0,2000 > :2528578459656*103 > :EndFor > :EndPrgm > runs in 18.736842 seconds on my TI89, which means 507 integer > multiplications pr. second. > HP49G+: > << #1d #1000d START NEXT > runs in 0.4371 seconds on my HP49G+. > << #1d #1000d START 2528578459656 103 * DROP NEXT > runs in 5.6277 seconds on my HP49G+, which means 193 integer multiplications > pr. second (DROP adds just a very slight overhead). > Similar tests for various (very much larger) integers: > (210-digit int)+(231-digit int): HP=257/s & TI=404/s > (210-digit int)-(231-digit int): HP=243/s & TI=328/s > (145-digit int)*(97-digit int): HP=10/s & TI=27/s > (145-digit int)/(97-digit int): HP=20/s & TI=26/s > (6-digit int)^(2-digit int): HP=7/s & TI=22/s > Sqrt(6-digit int): HP=10/s & TI=40/s > Result: the TI89 HW2 is 2-4 times faster than the HP49G+ for basic integer > arithmetic. > === Subject: Re: TI89 Titanium -- opinions? > Based on your posts here I'd be more than happy not to see you posting > again. Non-constructive bull. Explain your statement please. === Subject: Re: TI89 Titanium -- opinions? >> Based on your posts here I'd be more than happy not to see you posting >> again. > Non-constructive bull. Explain your statement please. You really have changed more to the depressive/aggressive or should I say chynical side Just try to lighten up and ignore hostality, remain cool Try to be polite even when others are not (Don't use you HP for counting to 100, nor a TI :) The humor and happiness will come back when the bad times are long forgotten Keep on writing opinions, ! {} PS: Be aware - I'm going to e-mail flames to your personal mailbox (-; === Subject: Re: TI89 Titanium -- opinions? > The new Titanic is so called on the TI newsgroup > because there is _NO_ _NEW_ CPU It got the nickname titanic by the tigcc developers out of annoyance. === Subject: Re: Day of Week built-in function? > I want a