B30 Can you tell me how can i do to enable communication, file transfer between a PC and my HP48SX via INFRA-RED. I only want to know if there is software available to do this, or the way to configure the hyperterminal. ---------- I have the same problem with my G and my IMac. I should program a terminal an my mac but I can't access to the Infra red Port( a serial port), witch seems to be locked... ---------- In the hp48 the metakernel, erable, and possibly jazz, all these applications need to be in port one. Erable is the absolute address version. Anyway due to their use of absolute addresses they can not all be stored simultaneously, this is beside the fact that there would be no room in port one for all three, but in the hp49g will we have similar problems with an application not being able to be installed because it was designed to use a specific memory block that another application is already using? Kind of like conflicts in which both applications use the same memory addresses. Also I maintain the bean series of programs (checking account library). So as I understand it, I would disassemble the library with the metakernel, and transfer it to the hp49g and reassemble it there and convert the directory to a library, and it should work just fine? please forgive me if I misunderstand something. ---------- [good points snipped] If by layout you mean the landscape format then hear, hear. By far the most distinctive characteristic... ---------- If the derivative-command would only look in the current directory ... But it searches the whole directory-tree upwards. E.g. [HOME] X=1.6 [FST_DIR] [SND_DIR] ... take the deriv. here won't work, 1.6 is still substituted for X. Searching the whole directory tree upwards for X and storing (tempor.) all occurences ...? Of course, it's possible to do it this way, but I would like to have something more straightforward. ---------- It's much better to deal with government people... They are well known for there intelligence over the teachers.. :) ---------- Ho ho ho. When he said it was customisable I guess he _really_ meant customisable. ---------- And abandon the very nicely optimized ROM? Sounds like replacing the windshield on your 80's car by purchasing a 90's car with better gas mileage. Why not just salvage someone's oldunused SSX? They're less expensive, have the LCD, and probably have an outdated ROM. ------- (flames It depends of what you call powerful. True that the TI89 has a more powerful C.P.U but do you know that the TI89 is slower than a 6 Mhz Z80 TI86 slowdowned by RAM bankswitching for number crunching and basic speed ? I have some benchmarks if you are interested. For example For k,1,1000:Endfor runs in 17 s on a TI89 in approx mode and in 5 s on a TI86. Perhaps that if TI completely rewrite and strongly optimise the TI-OS it will be able to produce a TI89 capable to blow away the HP49 but why TI will do that as its main target is high school students ? Unless someone else completely rewrite the TI-OS you will never see a TI89 capable to blow away the HP49. And about this utopic project let me just underline that TI plan to release a SDK and ASM support in January 2000. the expect sure a very the small desk. must The question is not that HP has not enough money the question is why HP will spend a lot of money on a no-profitable sector ? I am sure that HP was ready to screw up its calculator departement. I am also sure that HP has let a last chance to its calculator departement to show that they can produce a calc which can make HP earn some bunks there. Calculator like the HP48GX are very nice but they are very bad for the business. Do you know that even the bulky TI92 has almost killed the HP48 market in France ? Now about the C.P.U: How many time will it take to devellop a full new OS on a new C.P.U ? It has taken TI almost 5 years to release the TI-OS whom parts have been released by third parties companies. It has taken 5 years to TI to produce a craspy OS ! How many time would the A.C.O have needed to produce a quality OS on a whole new C.P.U very different from the SATURN 3,4,5 years ? Do you think that HP would have accepted to finance such a project ? Of course not especially when it would do no profits but perhaps lose money during this time. About speeding up the SATURN: Why for ? The goal is to reach the best performances possibles with it and it is what the A.C.O has reached with the HP49. About the impact of the actual work: Yes the OS will have to be rewritten for a new C.P.U except if this C.P.U is an evolution of the SATURN like the 68020 for the 68010. Also they will already something to begin with even if it is a completely different C.P.U. Because you think that all these attacks agaisnt the A.C.O and the new HP were nice ? No they were crasppy. They are a few ones who have done clever and constructive critics on this new HP but for the majority it was mindless bashing. Because of that i was very angry when i have posted this message so perhaps i have been a bit too far. ---------- You have excellent English, especially when compared to many Americans! Be proud of that fact, and ignore a few mistakes. I promise that my French (or any other non-English language) is far worse than your English. BTW, I think that the verb you were looking for was has released. :: o I went to hpcalc.org and read the spec's . If what they say is true I think I will definitely buy one. They've addressed almost all of my concerns with the -48 here. As for the loss of the IR port: so what! I always thought that was gimmicky anyway! All I used it for was to make a $150 remote for my Sony Trinitron, yippee-skip! As long as I can put the thing into RPN mode and compile my favorite -48 programs for it, it will be a worthy upgrade for me. I look at this new HP product as simply that: an upgrade! like win 3.1 to win 95, not perfect, HARDLY!!! But an improvement! I work in the engineering field and I recognize how hard it is to get a good product to market. ---------- What is the total amount of integrals that you have tabled?? It would be fun to compare it with my mathematical handbook :) An even more interesting feature would be Int48 working together with Erable to speed up the integration in a way like Maple does. Maple uses a fancy technique, first it looks for patterns in different tables and the remaining part of the expression that couldn't be solved by tables is submitted to the Risch algorithm (please correct me someone if I'm wrong here). With this technique Maple is able to solve almost any type of integral. I know this requires a large amount of programming (sorry I'm lousy on programming) but it would be interesting to test it. ---------- The reason that you see the SEE command and it not do anything is that it is not implemented yet. I am trying to get Matias to implement each category. For instance, if you go to the Irrational forms tables, complicated now but in the future I will implement the HOME command to be accessed from anywhere. The SEE command will allow the user to see an input formula that is too big to fit on the screen.---------- hpcalc.org yesterday and install it. The only menu I see is the SEE, SOLVE, CANCEL and OK. Even inside the Irrational Forms, I only see these menus. Do I do something wrong? or the version I download is old one? ---------- input tables. So, you have to scroll down to the very bottom of the command a little easier to work with in the next version of INT48. ---------- I think a lot of the reaction to the 49G is pretty understandable. The new design is a fairly radical change. It removes some of the main features that we use to put the TI crowd in their place. I think, as we first looked at the pictures of the prototype, that a lot of us heard the TI crowd's laughter in our ears. Another thing happened, I think. There has been an ongoing rivalry on this newgroup between French and Americans. For the most part (for the best part) it's been a friendly rivalry and not a bad thing at all. But the new other french ideas are now part of the system. Ok. That's life and it might not be a bad choice at all. But it does hurt a little. In short, the reaction to the 49G was an ego thing. We who didn't attend the introduction didn't know enough to react in any other way. But egos are important and some were deflated. I think there's another reason for this reaction, too. For a long time, HP has seemed like our own private calculator designers. Now it's beginning to look more like HP is a company trying to make money. Our child has grown up and gone off to school. How we will miss himherit! What happened is pretty understandable, I think. I suspect that it will turn out pretty well and in a few months we'll all be discussing the wonders of our new machine. But we have some hurt to get over first. I see the light now. The HP94G (sic) is actually the long awaited replacement for the HP-94DEF Handheld Industrial Computer. The 94 series had a rubber membrane keyboard so the 94G (sic) has to have a rubber keyboard. The 94 series has the Enter key at the lower right side of the keyboard so the 94G (sic) has the Enter key at the lower right. It is uncanny! Look for yourself at: http:www.hpmuseum.orghp94.htm and http:www.hpmuseum.org94f.jpg ---------- cordic algorithms which converge more rapidly than Taylor series, so it must use some other means for computing complex functions... you could define the principal root as the one with the least non-negative argument (angle), but this definition is arbitrary (from a purely mathematical point of view). what?!!! <:-O seems to be a step backwards! the HP guys should set this as a general solution. for example, like (-8)^(23) = '2 * (cos(n1*2*pi3) + i*sin(n1*2*pi3))' where n1 is to be substituted with 0, 1, 2 Q: do the HP designers engage math-trained guys to check on the details? ---------- The TI83 give the right answer... -8^(23) = 4 The only problem here, but after asking many teachers it appears that it's the way they teach it too, -8^(2^4) is considered as -(8^(23)) This has puzzeld me for a long time... After searching in many books about math priority etc... The unary sign is always described as one with the lowest priority order. I'm still not happy with this solution, because when I type -8 , I want to work with the number -8.. -8 is a whole by itself and therefore the unary sign shouldn't be considered as an operator but as a part of the number itself.. the 4*e^(2.08...) ---------- This is one place where the APLJ family of languages distinguish `-' (a binary or unary operation) from `_' which can only begin a number. This took me some getting used to, but now I prefer it, as it seems to keep my thoughts---and my programs---simpler... ---------- Could it be they teach it this way since this is the answer given by their calculator? Sigh. My daughter is in an AP math class that relies on a book with instructions written specifically for TI calculators, including programming instructions and examples in the text which give answers that are occasionally difficult to find anywhere except on TI calculators. For example, 2 chapters at least in my recollection is devoted to the study of box and whisker plots, whose usefulness I otherwise find somewhat dubious. Dennis ---------- Well, in standard order of operations, a^b^c, or c b a is defined as a^(b^c). This is slightly off-topic, but I'm working on a parser to parse algebraic expressions (in a string) in C++ into something more usable (a tree, or an RPL object list) - I'm currently reading each object into a queue and then dequeueing each one and then pushing it to one of two stacks, depending on if it's an operator or number or parenthesis, and when I push an operator, I check to see if what's below it has higher precedence and if so, I move it to the other stack - Anyway, hopefully someone is familiar with this method, but what modification would I have to do in order to make it correctly implement a^b^c? ---------- HP parses expressions using 2 precedence values for binary operators, one with respect to the left and one for the right. In the case of ^ they should differ by by so that the right hand side has more priority, that way you get a^(b^c) automatically. Other software which does the same include Java, EQSTK, Jazz etc. Note that you use the priority values to decide whether parenthesis should be displayed or not. ---------- whether it is correct or incorrect depends on how you define it. a^(b^c) is a more _useful_ (not correct or more correct) definition because (a^b)^c is just a^(b*c). why should anyone bother to write a^b^c if a^(b*c) was meant, unless it is a math-test trick question? anyway, thanks for the posting the amendments. what about 4 arguments, 5 arguments ... --- Enter to win at http:www.gennow.com ---------- : How to tranfer doc files ( files from Microsoft Word ) to my HP48 GX ? 1. Use the PC(or Mac)-to-HP48 cable. If you don't have one, buyborrow one. Use Xmodem or Kermit to transfer the files. However, I do not believe HP48 reads MS Word files (if that is your intended purpose). 2. If the document is text, save as .txt file, and transfer to HP48 via cable. 3. Manually re-key the document contents into HP48. ---------- You can transfer it to your calc if you have that much memory available but it won't do any good because you won't see anything but junk. My suggestion is convert it to text only and edit out the stuff you don't need and then download it. ---------- it's been a while since I used userrpl.. but I think what you are looking for is CLLCD message 1 DISP.. If I remeber correctly, this will display message on the top row of text on a blank screen. great suggestion! afterall, this is still comp.sys.hp48... ---------- It looks as if I'm going to keep using UserRPL until I'm the last one left :) [actually, that was George Burns' remark about how long he was going to continue in show business :] Yes, on one row only (or on multiple rows if the text already contains linefeeds); MSGBOX also word wraps any text, which is something else. ------- Provided the tekst strings fit on one line each, I suggest CHOOSE with a start value of 0 is view - only but not full screen You could have a look at FCHOOSE by Matt Willis - FCS33 MATT WILLIS this gives a fast (CHOOSE is also OK for one-line tekst strings) full screen viewer. Unfunately as there is no view-only mode, you would have to drop some things from stack afterwards, eg: Ç HP - DIN { :HP140[Times]7: 69 :HP140[Times]8: 84 :HP140[Times]9: 90 :HP280[Times]13: 597 } 0 FCHOOSE IF 1 == THEN DROP END È A simple fast full-screen viewerchoose box for one-line tekst strings would be really nice. FCHOOSE is ok, but takes quite a while for start-up if you have a long list of strings. Start-up of CHOOSE is actually faster, but not full screen, and scrolling down by keeping down arrow pressed seems designed to give you a heart-attack, while using left-shift arrow down is slightly better. ---------- HP is such a bunch of Bozos that it cannot even manage a product announcement. However, it is full of such Savants that it starts to create bogus web identities a couple of weeks before the buffonery hits the fan, and then instigates a systematic disinformation campaign. TinyWanda, have you been listening to Art Bell replays between the regular shows? --:: o there...see...!!!...it's not so hard to understand, you've got it EXACTLY...!!! the trouble with Art Bell is that whenever he starts to get too close to the truth, he backs off...he's not nearly provocative enough for me...!!! and it's not the sunny weather that wearing away at me...it's that the whole damn world is going to hell in a handbasket...!!! if there was one single thing in the world that was genuinely admirable, maybe that would be enough for me...but there isn't, and now one of the last flickers of dignity and quality have thrown it their hat and said... Mediocrity...here we come...!!! How on earth did you get it to work? I've tried on my Dell latitude with Windows 2000 Pro Beta 3, but I'm having trouble getting Windows to recognize the IrDa device as a com port. ---------- There are several clock adjustment programs in the Time section at www.hpcalc.org. I dont know if any of them do extra compensation for libraries.. ---------- But no one uses IR. There were post here about the printer and IR. Nobody uses the printer, either. Well at least if you ask about it here, few people respond. Now when the IR is gone everyone is squealing like a stuck pig. ---------- But no one uses IR. There were post here about the printer and IR. Nobody uses the printer, either. Well at least if you ask about it here, few people respond. Now when the IR is gone everyone is squealing like a stuck pig. -:: o perhaps the reason that there aren't MORE people ( and french ) complaining about the loss of the IR is that if you have a computer and post to this newsgroup, you'd use your computer and computer ( inkjet or laser ) printer to make hard copy docuements from the data generated by your 48... but WHAT IF you don't have a computer and consequently DON'T post to this newsgroup...? before i got this big box, i used the Thermal Printer ALOT...!!! since it was the only way that i had to save programs, and data from the 48...??? === people to Interesting point, not sure if its significant. Hard to communicate with the non-PC users (unless your clairvoyant). I would guess the percentage of 48-PC vs. 48-non-PC is fairly small and shrinking. IR is a victim of design optimization. It's *nice* to have, but for most non-essential. ---------- If you want to set the x and y ranges for a plot then use PUTXMIN, PUTXMAX, PUTYMIN, and PUTYMAX. Each takes a single real number from the stack. On the other hand if you want to move the window coordinates of the currently displayed grob use #x and #y being two internal binary integers of course. Note that GROB! does no range checking so if the source grob falls outside the bounds of the target grob you will corrupt memory. A safer way to paste a grob into the currently displayed one would be to use XYGROBDISP which automatically enlarges the currently displayed Here's a snippet of code to paste a grob into the graphics grob: .... TOGDISP * switch to graphics display LAM x LAM y * recall x and y coordinates to display grob at LAM grob * recall grob to be pasted XYGROBDISP ---------- If your gripe is with the design decisions and what was expected (this is my complaint too), so be it. -----:: o well...i would be will conceed that the actual 'design' of the 49 came from some unknown source, presumably they're hiding under a table somewhere... but Jean-Y has said in no uncertain terms that hethe-French had worked on the layout of the Keyboard, which is the definitive feature that is making it most closely resemble a TI generation type machine... the blame would seem to lie there...??? (!) ---------- You should change the title for: comments on the new HP49 after an hour of alpha testing... Type version : You will get A-9 Couple of months old.. Also... There is something you have to understand... This emulator emulates a HP49 at 85Mhz on my Pentium II 450Mhz ... It means that the HP49 is running 20 times faster that what it should be ! The HP49G is fast (surely faster than the TI89 for factorial for example) but not that fast... In order to know the speed of your emulated calculator (just before you delete the emulator): Press ON-D then press 2 you will get the speed of your Saturn. This emulator is faster when the screen is ON, which is quite strange... ---------- I had Emu48 loaded on my 420 (when I still had it), and unfortunately the 320x240 screen doesn't cut it to represent a whole bunch of keys, plus a display. The BIGGER problem is the terrible performance of the 420 (imo) - due exclusively to the SH3 slow-down feature - fyi the emu speed on it is a slower than an actual 48GX (have the emu speed figures somewhere). Sold my 420 after a month of use. Also tried Emu48 on the newer Jornada 680, much better - but my current Emu use is on a mobilon pro PV-5000 with a 630x480 graphics - and it really works VERY well and FAST! (IR transfer to and from my 48GX, no problems). they become available, and of course the Casio E-105 when my order is shipped. Nice thing about Emulators is that IF you don't like the position of any key (hmmm ... where's the any key ;-) ), it's fairly easy to redesign the graphics and remap any key. BUT - back to real use, battery life is THE problem, and worse still the lack of any physical feedback of a real keyboard vs. a touch screen (or at least until force-fields (sci-fi) are a reality). I have a nasty habit of running my fingers idly around my 48GX keys as I ponder the problem at hand - anyone else does this, or is it a fetish of sorts? ;-) ---------- One thing that greatly affects algebraic entry usability is the availability of the ( and ) keys. If you look at the pic on hpcalc.org, you will see to get the parenthesis, you have to use the shift keys. If this was a true algebraic layout, the parenthesis would be main keys, not shifted. One other question, on the new calc will the menus be like the HP48 menus, or like dialog boxes that i saw on my short time with the emulator, or will it be customizable ---------- Actually, I feel much safer in the middle of the Tenderloin in SF than I do in Seattle (or Centralia). Californians are okay if they don't have cars or attitude. Granted, that can sometimes be difficult to find... Ever notice that WA, ID, and OR as a group tend to have higher than average numbers of sociopaths (Bundy, for example?) ---------- I've never had that complaint on my 48, and I wear glasses (in theory). The key assignments aren't nearly as random as you lead on. Trigs are together, powers are together, the menus are together. The enter key will require a little readjustment, but that happened when the first SX came out, too. You realize all this can be done to *your* personal satisfaction using the USER mode? Frankly, I don't want pop-up menus. My calculator's a calculator, not Bill Gate's marital aid. Eeew.. it's not a gameboy, either... This was discussed elsewhere already.. .. that would require wading through all your pop-up menus just to execute a simple plot command. (Adding the Plot menu does seem to be one of the first USER assignments for me, judging from its absence on the keyboard) I'd wager your irrational tirades embarass more people on this newsgroup alone than any perceived lacks in the as-yet-unreleased 49. You certainly make me feel ashamed to be from the same state as you. I *knew* there was a reason I moved to San Francisco. ---------- Just as a minor correction, INT48 is a table of integrals for use with JAVA. the INT.LIB that comes packaged with Alg48 is what Aaron was referring to here. ---------- Somebody else replied: Why is every other country in the world allowed to decide what to call itself, but we Americans are not? Or don't a sovereign people have the right to decide their own name? Does anybody complain that Germans call themselves Dutch even thought they're not? Almost everybody calls themselves something which made sense once but really doesn't any more. That's why we study not only geography, but also history. We call ourselves Americans not for geographical reasons, but for historical ones. A knowledge of and a respect for history. Since its invention and awkward beginnings, however, ideas and code from all over the world have made it what it is today. It's too bad that this cooperative history does not engender a cooperative attitude in its users. Does anyone know a short program for the HP48SX to produce a vector with the logarithms of the elements of a former vector? I have asked John H. Meyers who sent me this program: but this does not work. Fourth step gives a bad argument type error, as the calculator does not support list processing and cannot take the after him in the chapter 29 of the manual failed also: there is no such program in my manual. Other people suggested me to use the commands DOLIST, DOSUB or STREAM, to be found in the chapter 17, but these commands lack in my manual, both in the chapter 17 and in the index. I would like to avoid iterative procedures such as FOR...NEXT or WHILE...DO. I would be very grateful if besides posting your suggestion, ---------- Where does this put it in relationship to the TI-89? It might annihilate the TI. 'Course many teachers will still feel threatened by it. ---------- I just aquired a BRAND NEW SX ROM VERSION J ( Still sonicly sealed in inner blister pack) . I got it in trade for an extra Commodore 64 I had in my possession as well as some duplicate Commodore 64 software I got from an old friend. I will probably keep it in my desk drawer at work as a spare and lug my GX back and forth at night.. Could someone recommend a SX program that gives it the List Manipulation Capabilities of the GX ie I have a list of 20 angles that I want to take the SIN of. or I have a list of numbers that I want to subtract from the average of the list of numbers etc. I have all of the goodies disk, so If anyone knows off hand what the program name is and what goodie disk it is on I would be most appreciative. ---------- I think that if we want our dream machine we should do a wish list and send it to HP instead of bashing the HP49. So what are your suggestions about:Hardware and software of the HP5*GX ? Me i just want: -A very powerful C.P.U 64 bits or even 128 bits optimised for BCD computations and low power consumption -A 320*240 touch senitive screen with 256 colors or grayscale levels depending of the autonomy -8 MB of RAM -A mini hardrive disk if possible if not 64 MB of Flash ROM or FRAM -Fantastic number crunching -A C.A.S which can compete in performances with Maple and Mathematica,etc... What about you ? ---------- It's probably already too late...I think the new calc is already in the queue, but one could try it anyway. And: It has already been done. Do you remember the 'design your dream calculator' contest? Powerful yes, but I would prefer a widely used 'off the shelf' CPU, just because such a chip is being developed further - with or without HP. I think there are at least a few valuable candidates on the market. Probably too big. I would like to have a true handheld device. No, please not a HD. Spinning parts in a calc? A calculator has to be 'solid state'. And what about power consumption and spin-up time? No, I don't like this idea. Hmmm...l think some kind of (limited) multitasking is a good idea, but a a calc. But I could be wrong here. I remeber a project from DEC 'Itsy'. Does anybody know what happend to this project? ---------- At first thanks to be one of 2 who have had a constructive attitude in this discussion. and the remember Perahps it is too late,perhaps not. just I agree. by the way HP is not forced to designed a whole new C.P.U. It can choose one of the availlable C.P.U and design a specific co-processor to perform fast BCD computations and other mathematic operations. You know 320*240 is the half of the size of the screen of the current handheld PC. It is possible to get a screen with this resolution and with almost the dimensions of the HP49 screen(a bit larger though). Have you heard of the IBM mini harddrive disk ? It is very small,it can have until 680 MB,it is designed for handheld tool and it is said to have lower consumption than Flash memory. a However if the ACO is able to release an OS as good and as optimised as the HP49 OS for the future platform i will be very happy. I just want blinding numeric computations especially on integrations,matrix computations and other operations often slow on a calculator. I think that an optimised C.A.S on a handheld platform has all the chance to compete with Maple as Maple used interpreted librairies(only the kernel is written in C and compiled). For Mathematica it is much more difficult as the whole software is compiled. However it shouldn't be impossible with very specific C.P.U and excellent work. A lot of people if it is affordable. One more time i am talking of tool which has to be released in a few years. Who can say what will happen in 3 or 5 years ? Perhaps that there will be 1 Ghz C.P.U for handheld PC. Perhaps even that my suggestions are too realistic. ---------- Actually, what he meant (and posted) was that the HP49 could rotate a graph of a 14x14 matrix at 6 frames per second. ---------- While being fresh to ML, I ran into a similar problem. : the BSRB is dangerous. You need somehow to specify that you're working with field A BSRB.f A With BSRB youre shifting all 64 bit to right. Since the left bits of B may not be ZERO, you may run into problems, you will divide field A by 2 but also add something from field M I would also suggest that when converting to lowercase, instead of reading the the wholde byte, simply specify that the nibble for uppercase (value 3, is changed to 5) and i think using register C instead of A would be faster (i dont remember of P should be 0 or 1, adjust D0 and reset p=0 afterwards if necessary) P=0 LA(1) 5 P BGN DAT0=C P D0=D0+ 2 B=B-1 A ?B#0 A GOYES BGN P=0 this should work for both upper and lower cases ---------- reading (value 3, faster afterwards simply writing a '5' won't work: A = #41, a = #61 Z = #5A, z = #7A As to just changing a single nibble instead of a byte - you're quite right there, save that the A=DAT0 B instruction is both shorter and faster than A=DAT0 P (and likewise for DAT0=A) ---------- I can see one error: BSRB should be BSRB.F A, otherwise you may introduce a '1' bit in the most significant bit of B.A. This may also explain the difference between JAZZ debug and straight EVAL. Oh the GOSBVL =GETPTRLOOP ought to be GOVLNG - but it will work regardless. Here's how I'd do it: :: CK1NoBlame ( check 1 object on stack ) CK&DISPATCH1 ( type dispatch ) str ( if string ) :: CKREF ( if object referenced, make a copy ) CODE GOSBVL =SAVPTR A=DAT1 A D0=A D0=D0+ 5 A=DAT0 A ( string length field ) B=A A B=B-CON A,5 BSRB.F A ( # of characters ) D0=D0+ 5 ( points to first char ) LCHEX 20 ( we need only the B field ) GONC + ( Branch Every Time - makes it work for ) - A=DAT0 B ( better than A=DAT0 2 ) A=A+C B DAT0=A B D0=D0+ 2 + B=B-1 A GONC - GOVLNG =GETPTRLOOP ENDCODE ; ; For very long strings, it may be worthwile to have the main loop do: A=DAT0 W A=A+C W DAT0=C W D0=D0+ 16 with C.W containing 2020202020202020 and have a clean-up loop afterwards to do the rest, character by character. ---------- I think it will work couple of times and then, when the CPU stack is full, it may crash doe to overwriting the reserved for interrupt stack level... ---------- No, RPL does not use any of the saved stack levels, so it will just push on onto the bottom and the one on the top will be lost. And none will ever be used so it will not matter. dan ---------- I was talking about ML stack which has limited number of levels. When program ends with GOSBVL =GETPTRLOOP then the address of next ML command to be executed is pushed into ML stack. Executing such program repeatedly may saturate ML stack and cause crash. ---------- CON stands for constant, I believe. It is equivalent to B=B-5 A if you prefer the other syntax. You take 5 out of B (field A). -- Sam. ---------- I had originally not used a SAVPTR and I was just going to save all the RPL pointers (is this right?) manually, but the program got a little more complex than I expected, so I just used a SAVPTR instead, but I forgot to change the AD0EX back. oh... ok. I didn't know the CON was a constant.. this helps.. I had considered doing 5 B=B-1 A instructions.. this looks much better though. is there a difference between this an LCHEX #20? ---------- . Unfortunately, if speed is an issue on empty string, my version is faster (even if the intend of my message was just to correct the bug in the routine, not to provide the routine itself... Let's see... Where ++ will test if the job is finish or not ... B=B-CON A,7 GOC EXT BSRB.f A (NB: .5 cycles mean +1 cycle if executed at an odd address) So if the String is empty, it's a GOC (4.5 cycles or 12.5 cycles is the jump is due...) While, with your version it will be: CSRB.F A 13.5 cycles D1=D1+3 : 8.5 cycles GONC (Branch Every Time: 12.5 cycles) ++ D=C A ( 8 cycles) LC(2) #61 ( 6 cycles ) B=C B ( 6 cycles ) LC(2) #7A (6 cycles again ) lop D=D-1 A ( 8 cycles again ) GOC ( another 12.5 cycles ) My version: 12.5 cycles Yours: 87.5 so if there's nothing to do (I mean really nothing to do): I'm seven times faster than you :) If there's nothing to do (which means: no need to convert the character), since you have 2 comparisons verified : it needs 42 cycles While in Werner or my version, because there's no test at all it will need: A=A-C.B : 6 cycles DAT1=A.B: 19 cycles... So in this case, it's faster as well.... This is FALSE advertising, you are misleading the customer ! Non mais ! Who has some time to waste today.... It to but ---------- I had the same problem. You have to change the character translation scheme. To do this on the HP48G series, hit right-shift 1 ( io ), then select the XLAT field and hit the menu key for CHOOSE. Now select Chr 128-255 and press the menu key for OK or ENTER. Finally, hit NXT and hit the menu key for OK or just press ENTER. Now you're finished! It should work fine now. ---------- Perhaps we need a comp.sys.hp49 (or a comp.sys.hp[49]flames* for all the my country is better than yours! people out there) -----:: o what an excellent idea, i'll keep checking back to see if they're up yet...?? ....... * why would anyone want to flame anything HP, except the 49...? ---------- What is the difference between FLASH memory and RAM memory? Can you store your programs in the flash memory? ------ To the ACO: Will there be new software available (for easier on-PC programming, swapping files, text strings, etc...) to connect your PC to the HP49 to go with the new cable? ---------- to use Microsoft products, could you? :) Here is my suggestion, while to Windows. Or at least make it easier for other people to write software, for example by posting specs and stuff on HP's web site somewhere. ---------- One quick thing: Change your if 0 = then with if then. It should be a bit quicker, and takes less space. If i ever forget to capitalize a proper noun, forgive me. i have ee cummings in my ancestry. Have a good day, and good luck in your endeavors! ---------- The trouble with the posted solutions is that logs are prone to roundoff error for large inputs. For example, anything close to a large power of 2 will get the same result as the power of 2. Here's a completely different approach which would make any good mathematician cringe, but it's what Douglas Hofstadter might call typographical math. It converts the binary number into a string and then just uses POS to count the number of leading zeros: BYTES: #E706h 46.0 Be sure to set the desired word size first. It fails on an input of #0, but seems to work for all other cases, even large inputs close to powers of 2. If #0 is a desired input, you could tack ELSE DROP RCWS END onto the end; then it'll work for all inputs. ---------- As I recall, wasn't the 28 capable of this? ---------- I don't think that it is a bad idea to share programs. It was my first contact with the 'public domain' idea when we started sharing programs for the HP-41 in our class back in the early eighties. I remember the time when we had a HUGE library on magnetic cards. I still have most of them. So, while it might have been easier to learn about the characteristics of a LCR filter by listening to the teacher, it was MUCH more fun to write programs for the problem in the meantime. And we've learned a lot. Writing a program to design LCR filters is at least as educational as listening to the teacher, for those who write it AND for those who are testing it. Aaaaah....remember the times when the 41 displayed 'RDY 09 of 11' and after giving it the card, the next display was 'CARD ERR'... :-) Synthetic Programming on the HP-41C by Bill Wickes was very popular in our class and therefore many of us had to frequently remove the battery from the calc and wait for a while to bring it back alive :-) Sorry, I'm getting a little bit sentimental here.... I think the solution with the cable for the '49 is OK, but it's not as elegant as the IR port. One more thing: Today, I'm not developing analog filters. I'm writing programs for computers... ---------- This will only work for libraries written in userRPL. Most were written in systemRPL or machine code, which as has been said in an earlier posting, will need recompiling. I have an equation library card for the 48SX but I don't see much chance of HP recompiling it for me. Perhaps a Jazz expert can point me at some documentation that will tell me how to decompile it and then recompile it for the 49. Perhaps someone who does or doesn't officially speak for HP will tell me whether they would be prepared to turn a blind eye if I were to do this. ---------- And also, someone has to correct all of these unique tests. ( I can seei it Even w2-3, these classes are so large, I've seen big version one clusters formed my students moving around. < Don't take this as meaning I don't want the IR, just reminising about ---------- In some of our bigger classes we use what is called a SCAN TRON. The student uses an answer sheet similar in form to GRE,SAT etc tests, so the question of grading multiple tests is moot. Each test version has a place on the Scan Tron sheet marked so as to indicated which version on the test is being graded by the machine. The type of test also depends on the particular professor you have. One I know gives more tests but fewer questions and the student must show all work and gets partial credit. Another uses the scan tron to grade multiple choice tests ( Yes even in Physics Courses). ( 100 and 200 level Freshman and Sophomore level courses only). We also use similar methods for testing in General Science and Astronomy lectures.; even when a numerical answer is required. ---- ---------- If the object is not referenced at all (e.g. not referenced on the Stack, local variables etc...), then the object will be deleted normally when a garbage collecting occurs.. HOWEVER, I do not recommend at all to proceed this way, if the Garbage Collecting process change, your program will crash... and anyway, it's dirty. On the HP48: If you want to reduce a string you can use the entry =Shrink$ that will reduce your string, but be careful, the object MUST be the last object created in memory On the HP49: Use the entry =Shrink$Any to reduce the size of your string (to extend it, use the entry =Stretch$Any), your string can be anywhere in the TEMPOB area. The argument of this system ML entry point is: D0=New end of simply ---------- Don't know if it was discussed before but, will the screen of the final release of the hp0.49beta will still be 131x64? I mean, if you want compatibility with ancient programs you can make the screen twice the resolution and add a `magnify' mode. Sorry for bad english. -----(french version, please no new english_is_better-jihad on the NG) Je m'exprime mieux en francais, d216sol216. Est ce que la version finale de la hp0.49beta aura toujours un ecran de 131x64 pixels ? Ce serait dommage de rester 210 une r216solution aussi basse pour maintenir la compatibilit216 avec les anciens programmes alors qu'il suffirait de doubler la r216solution et d'ajouter un mode loupe (avec des pixels virtuels de 2x2 pixels r216els) ---------- bolts of the expansion port and its protocol? Best Wishes. Vic Lopez. ---------- The specs for the HP49 are in HP literature published for the That's official enough for me. I don't need a faster CPU, but the OS is a bit slow. I don't *really* like taking 20 minutes to plot the galactic mass distribution, which I've done several times -- it's a fun integral -- and having an HP run much faster would be nice. Based on what's been said, the speed question has been answered in that regard. However, faster CPU would run the ML faster, just as the ML runs the OS faster. I suppose it's a matter of how much impatience the user has. IMHO, if color's really necessary, you can always use a PC. No beef there. I'm perfectly satisfied to set a few flags on a reset. I do it, anyway, based on the 48 defaults (set to single-[USER], USER mode on, etc). ---------- Pacific Tech sell an excellent product called Graphing Calculator (download a windows demo from http:www.pacifict.com - Macintosh users you already have a version that comes free with MacOS) that makes an excellent companion to the 48. Anyone who does a lot of graphs on a 48 just to visualise a function or to get a feel for its behaviour should try out this product. I'll think I'll suggest to them that they extend their product to support the HP49 (i.e. upload equations from the calc and auto graph them etc.) Perhaps HP might also like to persuade them that this is a good idea. ---------- I have written a preliminary HP49 FAQ based on questions and answers from this newsgroup. Hopefully this will reduce the number of questions asked. http:www.hpcalc.orghp49faq.html If you have any ideas for additions to the FAQ, let me know and I will consider adding them. Any corrections are also appreciated. ww.hpcalc.org --------- i have a 48gx, and i am wondering how can i work with rational numbers in the matrix writer instead of decimals, ie 12 instead of .5, i have done this in a TI but i would like to do it in my hp, would a program like alg48 allow me to do this? ---------- Yes, but for a smaller program, try: ------- I was just wondering, has integration been sped up on the HP49? On the 48 it is dreadfully slow, and I was wondering about it. ... what about INTEGRATE( 0, 1, 1sqrt(x), dx ) anybody tried this on a calculator? ---------- You forgot comp.sys.hp49.fhb ;^) ---------- Motorola 68000 does not have FPU either (altough 68020 and later does). Using BCD for 64 register renders 1844 less accurate answer than using binary for the same register. Not true. Only when you use exponent of power of 2 in binary floating point (which is preferred) the accuracy is different than BCD. There is nothing that prevents you to use exponent 10 (still coded in binary) to actually get exactly the same accuracy (actually due to the better coding accuracy will be much higher for the binary than BCD). 64 bit BCD is incapable to accurately represent the following: 9999,9999,9999,9999 + 2 where binary has no problem using the same register. ---------- congratulations on your well thought suggestions :-) I hope you don't mind me adding my thoughts here. Great idea I think. The right-arrow serves as a non-shifted SWAP in RPN mode. I don't know it it has a meaning in algebraic mode. At the moment I can't see how I will work with global vars like we were used to on the 48. I often need the ' (tick mark) and would want that to be a one key press function, but I'm not sure if the memory managment (not in the filer, but for the use with the softmenu's) has anothernicer solution on the HP49. I can't see it at the moment. If the ' tick mark really needs to be accessed as often as on an 48 I would prefer to have ' as a one press key. While you're right that a CAT function will need several keypresses anyway and one additional might be o.k. anyway. I also think that the keyboard is very well thought. (the free arrow keys help to save many keystrokes while editing :-) But it seems a little bit like the keyboard is very much improved for newbies. This is really good in one way. But without labeling the RightArrow-Key as SWAP I guess that very few who will be new to HP will find out how great RPN is in daily use. Sure it's not possible (or good!) to clutter up the keyboard with too many words. So it's not possible to have *everything* I also wonder why a precious one key press is used for the MODES. I personally don't change the modes too often. For those special settings I change very often I have asssigned tiny program *switches* to a user key. I miss the CST-key and I'm not sure if some of the new keys (APPS, TOOL, SYMB) will serve for this purpose. I consider a one keypress CST-key (or something similar) as a very important key. BTW, when I had my HP48 for a while and didn't knew what this needless CST-key was for I just looked up in the manual and began to love the CST-menus soon :-) A very very minor note is that I think it would be natural to have the TRIG functions on the 3-key and the matrices on the 4-key, BASE on the 2-key, but that doesn't make a real difference. After thinking about the new place a little bit I think we will get used to it very soon. BTW, it's near the SPACE-key which can be nice also. That would defintely a plus. The HP32SII has a very clever way to handle fractions! [+ in lists like other commands] this would help not to confuse newbies, but would confuse *old* programs :- I think that doesn't make a real difference. Nice ideas :-) Do you plan to join the ACO? ;-) As far as I know this is only possible with the help of KEYOB or similar programs which fetch the meaning of keys. I'd like to add that it would be great if the OS would allow to make mode sensitive key-assignments without the help of additional programs. Didn't knew that the HP15 used this way. Nice idea on a small calc, but I personally wouldn't use this option on a calc like the HP4849. But if you want to have this function it shouldn't be difficult to write a program for this, which you can assign to the STO key if you want to. If you don't know how to do this, just ask in the NG for help. would be nice if possible I'd like to know that too! As far as I have understand TOOL will serve as a mode sensitive key and will present appropriate commands. This could become a real time saver soon! This would be a plus. I already have wondered if the HP49 will have an equation manager? I'm sure that many of us store really many (maybe hundreds) of equations into the calc. There are some programs out which help with this task, but I use none of them, as every of them has a major disadvantage IMO. It would be a cool application and a great help, when one could set up a directory with all the formulas in it. This directory could have as many subdirs as neccessary. The equations could be stored as single objects, or if one prefers one could add a GROB, a string or other equations (for MES). The objects could just be collected in a single list (with no special order) and this list could be stored in a global var. Then one should be able to browse through the EQ's and just to take a look on them, or to load them into the SOLVR, or to view the accompanying text or GROB. BTW, the library from Matt Willis (when I remember right) is almost like this, but doesn't offer the use of subdirs :-( Ah, I forgot to mention: J-Y told that ANS will serve as a LASTARG key in RPN mode :-) _______________________________ Do you know the great Frequently Asked Questions? http:www.engr.uvic.ca~aschoorlfaq and the superb HP48 Software Archive? http:www.hpcalc.org to look for *old* HP48 postings see http:www.deja.com ---------- Yes. You've put the finger on the problem. You may consider it to be exactly this. Let's consider a number, so big that the mantissa is too short and 1 + won't alter the internal representation of the number. Then what I claim is that the MOD function loses exact meaning on this number, at least Mathematical example: if f(n)=(n! modulo 6) f(0)=1 f(1)=1 f(2)=2 HP48 version: << 1 50 FOR A A ! 6 MOD For 3