A42 ==== There s a doc about what you want on www.HP-Network.com ==== RATS!!!270)(/Û#)(/)(/&%Û# It's ALL your fault, Wolfgang!!! I'm not even close to 10,000 points AND my 49G keyboard is now so soft that I get key bounce no matter what the value is that I feed to ->KEYTIME ==== can be found at http://education.ti.com/product/tech/30xa/faqs/faq83086.html. The description likely applies to HP products as well. I would show the time in ASM on the HP49G, HOW ? ==== I am making my way through Gilbert Urroz's 2 volume masterpiece on the 49G. It's very good, although very long. One thing that is a bit frustrating is that when I compare results on my calc with the ones listed in his book, the result IS the same, but displayed in a different way. I know the 49G allows many formats for display, which is great, but if you want to give someone an example result, it would be convenient to say make sure your mode value is xxxx, where xxxx, say, is something like a bitmask of ALL modes: FLAGS, CAS and DISPlay. This would allow completely reproduceable results, provided you have the same ROM, which *can* be checked using VER. Anybody have an answer? Mark. ==== Use CASCFG first If of no help then you have to downgrade to ROM 1.16 which was in the marked at the time Urroz 49G. displayed anyway to do it? ==== I was experimenting in Debug2 something about the new 49 CHOOSE engine. As reported in Carsten Dominik's tutorial (Using HP49 CHOOSE for more than just selection), this is done by the unsupported entry something like FPTR 2 72 but Debug2 compiler gave me an error in that line. I tried to change FPTR in FLASHPTR, but this won't solve the problem. So I will be grateful to the one that explains me about rompointers abd flashpointers calls in debug2 and MASD sintax (I guess they are different). /-----------------------/ Marco Tinarelli - Bologna - Italy tittiXXYYZZ_tt@tin.it (togliere XXYYZZ dall'indirizzo per rispondere) ==== I tried and it works: :: a b c 3 tyt 0 ' DROPFALSE FLASHPTR 002 072 ; Piotr Kowalewski. ==== Well, FLASHPTR 2 72 workes fine for me using debug2. Greetings AM ==== I want a smoth interpolating curv of values in the X-axes relatet to values in the Y-axes. The nearest funktion I found is the scatter funktion. But I want a curv instead of dots. ==== After SCATRPLOT press |STATL| which draws the least square model that you have chosen Note: The Plot changes to Function now! I hope this is what you wanted... values in the Y-axes. instead of dots. curve? ==== How can you solve an Initial Value Problem eg.: dy/dx=exp(-3*x) where y(0)=5/3. I'm in RPN but can't find the syntax. Also, how about a symbolic solution of an equation with a domain restriction, eg. Integral(1/(1-t)) ; (t>1). David z. ==== A) You have to give them in a matrix [ d1Y(X)=EXP(.3*X) Y(0)=5/3] Y(X) DESOLVE B) 't' STOVX 't>1' ASSUME but I don't know if REALASSUME list applies to Integrals ==== Hey. I've been using these two libs for a long time, tried the 3.7b4 version of java too. Great programs. My question is ... does it appear to you that java is significantly slower at displaying algebraics (in pretty print, of course) than EQstk is? something is screwey here... 'cause EQstk's AGROB needs .226s to process 'Integrate(a,b,Sqrt(c*x^n-1/d),x)' (small font, just a random expression) while java's AGROB command (which is the same, right?!) takes .43s, Twice as EQstk! And java 36a takes .45s :( with large font it is more or less the same situation. Accidentaly i found out that java's AGROB on a *grob* takes .29s! longer than EQSTK processes the *algebraic*!! why does this happen? ...if only java was as fast... now i have to switch back and forth... :(( ==== moin, does someone know, where to buy a leather pouch/case for a HP 49G Calculator? I'm thinking of a pouch like the original accessoire for the HP 48... Anyone know where to buy it in Europe (or better Germany)?? Or does anyone sell his own? I tried at ebay and many shops, but didn't found anything... Matthias ==== Don't use it! It's based on an earlier 1.19-6 and only the CAS is better but you'll miss for example pretty-print e^x JYA & BP have promised to make a united 1.19-7 later on - just don't hold your breath (I guess a good release date would be 1.4.2002 :) ROM?? X X Every calculator in the history will someday vanish the production stopped and the official support will disappear, so that cannot be a reason to not to buy a calc. AND there is no mention officially from the HP that the HP 49G (nor the 49GX ) is not produced anymore. ACO is dissolved, yes. But still JYA & BP (at least) from the original developers of the v'ger team continue to support the 49G. And what about the active user group?! his goodies, while totally pissed off by ACO's death. (not unlike myself and many others) The TI 92 has new members in the family the latest which of has a whopping 3MB of Flash. Still I would not hesitate to buy any of the TI models if they would fit the bill. Or yet better: buy a TI 89 AND a HP 49G and get the benefits of the both worlds!!! It is not dead until the Hex Lady Swings (or something...;) ==== PS: I'm IN for open source or even restricted open source X snipped a long ranting about 49G is not dead (yet) ==== I just checked the source, and there might be some conflicts, depending on how people are using the library. QVSOP (the command) currently outputs only planets 5-8, all at once. If a library is using *this* command rather than the planet-specific version, it will break, unless other changes are made. Certainly, the preferred command is QV_D.Y.n>LBR. A new command should be added to indicate which planets are supported by a given QVSOP library, in the order that they appear in the output of the QVSOP command.. Or, another option: set up a QVSOP library with a different romid, dedicated to the inner planets. This allows for some flexibility in memory usage, at the expense of using up another romid. So, Georg and Dennis -- what're your thoughts since you two know your own libraries? ---------- Keith J. Farmer kfarmer@thuban.org http://www.thuban.org ==== any way to solve quad, equacao do 2¡ grau(it has that name in brazil)? solver does invalid data. ==== I'm thoroughly enjoying the Stat49Pro program. I'm having a little trouble with the one-way ANOVA, though. I tried using the data that came with the program and my own data, but the HP-49 keeps giving me error messages like undefined result or invalid dimension. I suppose I'm entering the data into the calculator incorrectly. Any advice/suggestions/tips? ==== The Euro symbol is used by MASD to enter special characters.. I guess the decompiler doesn't recognize it ==== The minus sign in the argument of the KEYEVAL ensures that the original key functionality is assigned. Using a positive number (no minus sign) indicates that you'd prefer to use a possibly *assigned* key (in User Mode _only_ - naturally) to be assigned to the designated key location. . ==== it to my 49G. Could anybody give me some pointers?? ==== I hope you do not mean my STARTEQW programs?! The documentation to that is of current HP standard that is: try it out! The programs are supposed to reside in the Home level and in CASPAR plus my PARDIR AND especially in the STARTEQW which is run every time you press [LS] [MODE] e.g.. [CUSTOM] while in EQW then you can select from a CHOOSE list a sublist (in my STARTEQW case) and then the final selection of an operation that you would like to perform on the selected subexpression. == On my HP48 i've ever seen pictures with 16 levels of gray scale. On HP49 32 levels seem to be possible. Can i have a picture with 32 levels of gray scale on my hp48 ? It's impossible or it's possible but that will scintillate ? ==== What about this one: { 'cos(&n*pi)' '(-1)^n' } vMATCH DROP ? too bad it won't work on cos(pi*n)! Marco ==== I used my HP48 as a remote control. I am only using the XMIT command followed by any character to transmit. The problem is that it is a little slow because of the 2400 baud rate and my use of extra transmit repetition to decrease the probability of errors. I wonder if there is any way to increase the baud rate in the IR repetition to 9600 baud rate. Best , Tal ==== i don't have a 48. wouldn't this work? 9600 baud m ==== Not for IR. You can boost the IR up to 7500 bps, above that it will fail I've looked through the archives, but remain confused. My Mac has both USB ports and Mac-type serial ports. The hp49 seems to have come with a cable that will connect it to an included adaptor that ends in a PC-type serial connection. Can I just connect this to a PC-type serial-to-USB connector and connect the USB to my Mac? Do I need any particular software? I don't have Virtual PC, and don't care to invest in it and in Windows, especially just for this one purpose. But I certainly would like to find a way to be able to download stuff for the hp. ==== On an older Mac, I connect through the phone port, which is a serial port, and it works. ==== SOLVESYS give me the two answers posted in this thread, running on my 48GX I was interested in solving it symbolically: having some spare time I learned a little MuPAD and tried the proposed equation. I wonder whether Mathematica can get hold of it or not...? Marco ==== ****DON'T**** try this unless you back up all your memory. all will be gone :) i was trying to solve this equation: so i typed: '5^A+7^x^2=15' x solve BOOOOMMMM revision 1.19-6 verified in my two calculators. i would be pissed if this happened in a test! ==== ... '5^A+7^(X^2)=15' X SOLVE gives: 'X=SQR LN(-(EXP(A*LN(5))-15))*SQR LN(7)/LN(7)' 'X=-(SQR LN(-(EXP(A*LN(5))-15))*SQR LN(7))/LN(7)' 48GX-Erable3.2 and then, with ASIM (from ALG48): 'X=SQR((PI*i+LN(5^A-15))/LN(7))' 'X=-SQR((PI*i+LN(5^A-15))/LN(7))' ====. Did some experimentations... '5^A+Y^X^2=15' SOLVEVX crashes when Y=2,3,5,7,10,11,12,13,14,15 (especially bad when Y=2,14) but crashes not when Y=4,6,8,9 :-? Marco ==== Can I ask what's possibly a silly question? The original message said: so why have most of the postings since then talked about 5^A+7^X^2=15 instead? Is it just a typo and you're all actually using the original? PS. My HP39G seems to have no trouble finding numal solutions to the original equation although it does take a very long time (~30sec) to find them. The CAS system on my 40G (using SOLVEVX) reports EXP Error: Not reducible to a rational expression. ====================================== Applications in Mathematics http://members.iinet.net.au/~ccroft/ ====================================== I know that there are up to 4 levels of gray scale on the HP48 calculator. I generally use CPicEdit program to create my GROB in 1 level gray. I wondered how to add another GROB with another gray scale. ==== maybe http://www.xt-ce.com/ ==== Maybe, but again it is a emulation solution.. I would like a real CAS using all the features of this kind of machine.. and best solutions seems to be linux+giac ==== I also wondered about this. I'd really like to have MK, the latest Erable and Jazz coexist on my 48GX. But MK takes port1, Erable is a hefty 80+k in port 0, now where do I put Jazz, not to speak of my own stuff? I have a 128k card in port 1 and a 1m card in port 2. Is there an alternate setup which gives me some noticeable free user ram?t -- http://www.musikwissenschaft.uni-mainz.de/~ag ==== Well, I suggest that you just check the 49G and make up your mind yourself. Some advantages of the 48GX: - nicer keys (at least for those who prefer the old style HP keyboards) - IR port, printer interface - expandability (but see comments on the 49G memory below) - more user-written software (but 49G is catching up) Some advantages of the 49G: besides a lot of optimizations and other goodies like the builtin ML/SysRPL development system - better keyboard layout (well, the ENTER key is in the wrong place, but you'll get used to that) - much more builtin memory, better memory management - flash memory (ROM upgradability, protection against data losses) - nifty new design (despite some flaws with the keyboards and screencovers of the early models) Of course these lists are not exhaustive, and are based on MHO. I recently bought a used 48GX (like new) as a companion for my 49G and immediately upgraded it with two RAM cards. It's a really nice device and I like it very much, but all in all I agree with Wolfgang Rautenberg. With the 49G you get the most bang for the buck. Of course, if you already have a 48GX with RAM cards and have to shell out extra money ... But just imagine: HP stops making calcs and you never got the last hero. ;- ==== I said merged card 1 with MK but i want to said my port 1 is full with the MK, which required a merged port during the installation But it's truth it is not merged, because port 1 appeared. Sorry for my mistake. For Jazz the only solution is to install 384Ko in port 0, but you can destroy your beautiful hp during the operation :/ Boring are all these prog which required port 0 to run..... I prefer the HP48 keyboard, but... ====.. I can only say you how I've tried to avoid the short memory problem: Like the yours, my 48GX has a 128K and 1M cards. As you said, MK takes port 1. Nothing to do. I don't use Jazz: I work only a few in SysRPL (as making little modifications in some programs) and I do it with MK's ASM, ->S2, etc Erable3.2 takes the most of port 0 and about this, there are many things we can do: * My own programs are stored as Libs, in 2nd RAM card * The MK's configuration programs (STARTUP, STARTED...) are ONLY A CALL to the actual programs, that are stored in covered port, as a Lib. * The only objects I've in HOME, are system (hp and MK) variables: PPAR, SYSTEM, MATRIX, EQ, IOPAR... Also a directory, called TEMP, in which I store something temporaly. * The bigger object in my HOME is the userkeys information in the * Many other libraries are installed in 2nd RAM card, without problem when runnig. This is my way since I installed the free MK in year 2000. And this let me >32000 bytes free for working and I don't get the Insufficient Memory while calculating. (Algebra, Calculus and Statistics), but sometimes I need some extra memory for a transfer or so... I purge Erable and when I want, I restore it from a backup of port 2. Having all my programs as libs, is not the most comfortable if I want editing them, but the adventages of MK+Erable are so many... If you want/need some extra information or ideas, I'll try help. and Merry Xmas ==== In Erable, the kernel.lib need to be in port 0, but the erableg.lib can be in port 2. It will just be a little slowly. I have also the metakernel and erable compressed with bz in port 5 and 6, for reinstalling mk when i crash my hp (asm :o) ====.. ...but Erable 3.2 (the last, fastest and more estable version) has a library of 85305 bytes that must be installed in port 0, if you have MK in port 1. This is the reason of all I said in my last post. ==== be nice, but I guess the easiest way around this is to use my 49G for doing symbolic calculations, then I can keep the 48's port 0 for my programs and Jazz. Albert -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag ==== I am looking for a HP48gx in order to replace my 12 years old HP48sx, but it seems that only the 48G+ is available in the shop i asked. I need the GX because i have some a 128kb card (HP original) and some applications on ROM card. Anybody knows where it is possible to find a 48GX in Italy (Milan or Rome)or in Italian Switzerland (Bellinzona)? Marco Polo -- Marco Polo marco.polo@katai.cina ==== See at http://www.srssoftware.com/PgVenCal.htm -- ================================================ BELLANTONE MAURIZIO ================================================ Marco Polo ha scritto nel messaggio ==== My HP48 started missing keypresses. I have a 2 Meg RAM card and a merged 128k card, and have had for a long time before this started. Maybe I have something interfering, or data garbage buildup? Would a ==== Have you seen HP site about next Invent 2002 worldwide developers conference: Intelligent Devices: HP's Latest Smart Hand Held Device: Pioneer a new Jornada Application Development for Digital Music Lovers Realizing HP's Internet Vision: Chai technology compared with Sun & Microsoft .. And more.. Some of these points sounds familiar to me.. We'll wait and see.. J.Manrique CdU de la ETSIG ==== http://www.e-ventcentral.com/events/invent2002/index.php ==== When i type in the limit of n/(n+1) as n approaches infinity, in the equation writer, i get the result n/(n+1).....when i use x instead of n i get the correct answer of 1...........when i use the n's on the home screen i get the answer of 1.......why wont n work as the variable in the equation writer????? ==== screen It will, you just have to enter it correctly. When entering the limit n->infinity, you'll have to use the equal sign. Here the UNDO/inversion bug is very clear. ==== i dont have to type x=infinity, i just type in the infinity symbol and eval and get 1. but using n i do need to press n=infinity.......maybe ill try some other variables ==== If you don't supply a variable, the current variable (VX) will be used. ==== I'm amazed that you understood that at all =) When I do A, I get the correct answer of B -- but when I do A, I get the incorrect answer of B. What's wrong? I don't require the equal sign -- and trying to enter it in the EQW only inserts a -> for me. What's the UNDO/inversion bug? ==== Carly's calcs ;-) ==== Carly's Playland Inc. NoMoCalcs Inc. PrintersAreUs (or in Watts, CA - WeBePrinters) The ZeroValueStock Company :-) Hewlett Packard Compaq. A not for profit company. ==== Well, given Carly's work toward 'HP' over 'Hewlett-Packard', I bet the *real* name will be 'HP-Compaq', unless they do like DEC and name themselves any ol' idiotic pseudo-existing-term. ==== I just ran across a previous argument along these very same lines from John: http://groups.google.com/groups?hl=en&frame=right&th=e3a7b154a66a659e&seekm= 3 69B0AF8.27B2%40mum.edu#link1 ==== I wish you all a very Merry Christmas and a happy 2002. May God introduce Satan to Usama Bin Laden and all his sect very soon. , Bye. PS: Suposo que deu ser pecat desitjar la mort lenta i violenta de tota una colla de fills de puta. Per230, qu217 cony! 203s Nadal, l'217poca de les bones intencions ;-) ==== The movie is great! (but three hours, it should have had an intermission) It's unbelievable to see these stories brought to life on screen, it faithfully captures the spirit of the stories. ==== Is there any way of solving a linear system that is in the way A*x-b=0 instead of A*x=b? That is, I don't know A and b separately but I know A*x-b; For example if I have a list of the equations: { ' x+2*y ' 'y+2*x+3' } So, the solutions should be x=-2 and y=1 What can I do without computing A and b to solve it? ==== I think this should be a FAQ, anyway (assuming RPN mode): First build a vector containing your equations: LS [MTRW] [EQW] X + 2 * Y [ENTER] [EQW] Y + 2 * X + 3 [ENTER] [ENTER] Alternatively, enter ['X+3*Y' 'Y+2*X+3'] - this is what level 1 should now contain. Now you need another vector containing the variables to solve for: ['X' 'Y'] Then SOLVE, and you get the solution 1: { ['X=-2' 'Y=1'] } Note that this way of solving only works on the newer ROMs. If you don't have it yet, you should install 1.19-6 anyway. Check with VERSION. Thomas PS: If you already have the list of equations on your stack, use AXL to make it a vector. -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ==== The answer depends on your calculator: * 49: 2: list of equations 1: list of variables LINSOLVE * 48 with Erable: 1: equations AND a list of varibles, in the same list SYST For the 48 w/o Erable, your method is the only one... I think ==== Which can be produced by running LNAME over the system :) Marco ==== Just download the latest ROM from JYA's site and emu48 from Cristopher Giesselink's site, a real xmas gift if you never had a 49 in your hands :) Marco ==== looking for ascii as in <64><65><66> using a loop etc.. Icurrently progam in vb qb vbs and plc ladder logic but not too familiar with rpl as of yet. and ==== ok found a way to convert string to ascii in a list but how to put a string into a list ie abc={a b c}?? thanks in advance equivallent equivalent? ==== If you say how you do the former the answer may be as simple as put CHR here, otherwise you can have a second loop with CHR. Please, though, tell us 1) what you *really* want, unless your sole end is to get a program to do what you describe, and 2) why you want it, which ought to be a redundant request if you answer #1 correctly. Somebody may have a better answer to your deeper goal. ==== Huh? You lost me on that one. Given a string on level one, try the following program: << {} OVER SIZE 1 SWAP FOR s OVER s DUP SUB + NEXT SWAP DROP >> Note that that isn't optimised, and works on a 48 (S or G) series or 49G, and for that matter even on a 28S and I suppose a 28C. Which model of calculator are you using? If you're using a 48G series or 49G, then maybe it coud be improved with the HEAD and TAIL commands. If you're using a 49G, then SWAP DROP could be replaced by NIP. , James I was wondering where the WSLOG information is kept, because when doing ON+A+F all the data in the calc is erased, but not the WSLOG, and it just adds to this file the warmstart caused by the reset. I figure that it works the same way in the 48 & 49, but I'm not sure. I wish someone could explain it to me how all this works, where is this information stored, ... Gregorio CdU-ETSIG I would like to change a variable's vaule which is in another dir. For example I'm in HOME/WORK/, and I want to change HOME/foo's value. ,---- | 3: blah blah new value | 2: 'foo' | 1: STO `---- makes a 'foo' named variable in the current dir, so what should I put on stack level 2 to make 'foo' in my HOME dir regardless where currently I am? thanks in advance -- cyberbob ==== Level 2: blah blah new value Level 1: << PATH SWAP UPDIR 'FOO' STO EVAL >> EVAL will save the new value at parent directory and return to current directory. You may use any directory you like instead of UPDIR. Hope this may help. K.K. ==== On Wed, 26 Dec 2001 15:21:09, CyberBob If you are in home/work, then: UPDIR new value 'foo' STO will create an instance of foo in HOME with the value ''new value. But this is NOT necessary if 'foo' already exists in HOME. The HP48 searches UP the directory path (shown in the display) looking for instances of a named variable. I presume that you are trying to write a *new* variable in Home. If 'foo' exists it will be visible from home/work and should be updated without problem. If 'foo' does not exist on the path, then a new instance is written 'here' (wherever you are). Note that a variable 'foo' in home/bar/ will not be visible from home/work. In order to avoid these sort of problems it is easiest to create your working directory for a program and put all of the sub-programs *under* it, with all the variables in the working level except for the main program which launches everything. That way all of the sub-programs look 'up' for their variables without problem, and all of those variables are invisible from 'above' the working directory. AS in: Home/bar/vars and Home/bar/vars/progs (except for <> which is in vars, and whose first command is 'PROGS' to change to the progs directory. See Chapter 5 of the HP48G Series User's Guide (available as a zip/pdf from hpcalc.org thanks to Rechlin.) Geoff ==== CyberBob schreef: You can't you must change the directory to home first try this: << PATH HOME blabla 'foo' STO EVAL >> -- This message was written with 100% recycled electrons Pivo ==== (1) The command PATH will give you the current dir, (2)then go to the dir you want and (3) for going back, EVAL the result of step 1. Hope this helps. ==== >> I would like to change a variable's vaule which is in another dir. For >> example I'm in HOME/WORK/, and I want to change HOME/foo's value. > Level 2: blah blah new value > Level 1: << PATH SWAP UPDIR 'FOO' STO EVAL >> > EVAL > will save the new value at parent directory and return to current directory. > You may use any directory you like instead of UPDIR. Geoff, thanks for your explanation about 48's directory system! -- cyberbob ==== I would like to change a variable's vaule which is in another dir. For example I'm in HOME/WORK/, and I want to change HOME/foo's value. ,---- | 3: blah blah new value | 2: 'foo' | 1: STO `---- makes a 'foo' named variable in the current dir, so what should I put on stack level 2 to make 'foo' in my HOME dir regardless where currently I am? thanks in advance -- cyberbob ==== Hrmmm, since they are written in SysRPL you could disassemble em on the host calc and then recompile em on the 49 (just in case the entry points are different). What do you people think about it? ============================== I don't think you could automaticly convert an ApLet to run on the '49. You'll always end up w/ rewriting it to become a library or directory and you'll have a hard time to work around the 38/39/40 TOL I/F and task switching. Bye, Detlef -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel ==== host complex I tell you. /me calls out Detleffffff! Hmmmm I suspect JY is out riding his brand new bike ==== Oh, and anyway, there are so many cute aplets on your page I'd like to try them out on my 49 (hence my question). I have emu48 with 39/40 rom but I'm not very at ease with their user interface What about a compatibility layer written as a library to be installed on the 49 ? ==== Well, thank you kind sir. :-) If you don't mind trying them out on the emulator then download the package on my Utilities page (if you haven't got it from there). It includes ROM images containing most of the aplets. Unfortunately they are for the 38G rather than the 39/40G but the behaviour is identical in most respects. You'll need to download the aplets themselves to get the documentation on how to use them. Let me know if there's something I can help with. You can download an excerpt http://members.iinet.net.au/~ccroft/zipfiles/excerpt.zip . It's only small but it might help give you a start. -- ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== May (!) work but a) it's hell of a library to write (sorry, don't have the time for that) and b) you'll still have to recompile an ApLet for running on top of that layer. Bye, Detlef -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel ==== Don't forget that most (95%) of the aplets written for the 38/39/40G are written as uRPL (text) programs attached to copies of the standard aplets built into the BIOS. If I understand the situation (not being a 48/49 user), as such they would not have code to recompile. Can I suggest that you just buy a 39/40G? It would certainly cost you massively more in time costs to write the library even assuming that it was possible. -- ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== Well, that's not user-RPL. In fact, their user programming language does not have a name at all (AFAIK). Or buy an HP38G: The last pre-ACO machine! :-) That library (if limited to user aplets) can be written; though, as Detlef noted, not without great effort. Anyway, the text is already compiled in the 38/39/40 before running, so porting that part of system code could be crucial. If anyone is interested in porting a particular aplet can contact me. I've realized many migrations of software among HP programmable calculators, such as the HP-41C, HP-42S, HP38, HP48 and HP49G. , Bye. HPCC member #1046 ==== from a mother whose son is going into high school. She wanted to know if an HP41C was the same as the 39G that was on his booklist. Although I told her it wasn't the same, I wasn't actually sure what it was - I just knew it wasn't a graphical calc so I guessed a programmable scientific. Correct? -- ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== A bug in FPTR2 ^RSUBext seem to have surfaced: {'A' 'X'} {4 3} FPTR2 ^RSUBext -> {'A-4' 'X-3'} {'A'} {4} FPTR2 ^RSUBext -> {'A-4'} {'A*X'} {3} FPTR2 ^RSUBext -> {'A*X-3'} {'A/2' 'X/3'} {2 5} FPTR2 ^RSUBext -> {'A/2-2' 'X/3-5'} These are all ok, but see this: {'A/2'} {2} FPTR2 ^RSUBext -> {'{A-4}/2'} That is not a valid user object, and can be dangerous to further calculations. It seem to happen only when two lists with one element in each are subtracted, *and* one or both elements contain a division. It seems like a bug to me, unless there is something I have overlooked? ==== About a year ago, I said this: I've just spent a couple of hours being driven mad by the same problem again (on the latest beta rom), and have finally figured out what sets it off: having the res value in PPAR set to a non-default value. Reproduce by: 1) Plot some y(x) function, with 'Indep step' value set to something other than 'Default'. 2) In the 'Plot Setup' form, select 'Wireframe' or 'Fast3d' mode. Then go to the window editor form, and try to set 'Step Depnd' to something other than 2. Close the form, and inspect VPAR: the 'Step Depnd' value (last member of the list) will be set to 2, regardless of what you tried to set it to. Workaround: manually or programmatically set the res element of PPAR to 0. If there are any further ROM releases, I for one would be very pleased to see this fixed. James ==== i think about an error. XT? TK-? CP-500 ? MSX? Eniac :D ? IBM PC? ==== Beto escreveu na mensagem <9l4ru8$7ihb8$1@ID-95421.news.dfncis.de>... Good choice... mupad is very good, but is sometimes slow(on Win95+P233+64MB RAM+2MB Video+2 GB HD). ==== Roger Metcalf escreveu na mensagem ... it's very very good. ==== Does HP have a copy right on the saturn architecture or is it legal to design chips with the same architecture without copy right infringement? ==== what the programs? i know alg48+erable+cashp(i dont remember the correct name) any more? ==== I have updated the documentation of ClckAdjst. Automatic clock adjustement including change of time zone and daylight saving time. It is available under http://users.belgacom.net/EAA/Heiko/ClckAdjst.zip and soon on 's site. There is still potential to improve the documentation and the user dialog, I am working on it and would appreaciate your comments. I wish all of you, a happy new year ..Heiko ==== As it happens, I don't like MS software, because it is buggy and expensive and insecure. I don't like MS the corporation, because I dislike their ethics, as amply shown by the evidence brought forward before Judge Jackson, and even more amply brought forward in deposition evidence which was not explicitly put before the Judge, but which is public. Legally there was no reason, *aside from anti-trust law* why MS should have had to make the same deal with IBM as it made with Dell. But the fact that the 'deal' was so very bad for IBM, and so costly, is evidence *only* of MS' intent to leverage its market power to hurt competing products in their market, in any way possible. And if that was the *end* of the bargaining, I wonder what the original position was? You should know something about what you are writing about, before putting pen to paper, or electrons to phosphor. OS/2 runs perfectly well on 'DOS based' FAT. (FAT is not 'DOS based'.. It is a standard, DOS happens to use it. OS/2 can also use HPFS. HPFS is an extra which is not required in order to run OS/2. HPFS are invisible to Windows. (As it happens, NTFS partitions are invisible to OS/2..) There is no technical limitation why OS/2 and Windows cannot co-exist on the same hard drive (except for a size limitation: you need more space to have both, and also have some working space). It did not have to create anything 'special'. Which is exactly what I did: I chose OS/2 *before* Win3.1 came out, and I have never owned a copy of an MS OS since DOS 6.22. But you miss the point: MS used its leverage in the desktop market to force its most dangerous competitor to limit the choice of its customers to a one-timce choice, when there was not valid technical reason why IBM customers could not have had that choice every day and whenever they wanted. And technically oriented customers could achieve that choice: a Boot Manager has been available from IBM for years. But 'retail' customers were denied that choice. AND MOST IMPORTANTLY, Microsoft received payment for *every* shipped copy of Win95, *whether or not* the customer chose to install it. I will absolutely agree that IBM turned out to be the 98 pound weakling in this mess: but they were faced the other possibility, that they would end up selling a total of, say 50,000 computers instead of the 1,000,000 they thought they could sell *if* they could offer Win95. The economic argument in favour of succumbing to the extortion was quite strong. But that is what everyone who pays 'protection money' believes. Which does not excuse the extortioner. The difference between criminal law and anti-trust law is that the former requires threats or menaces of physical harm or damage. Absolutely! You have the timeline backwards. MS developed Windows 3 secretly, while it was a partner with IBM, and announced it, at a Comdex, the day before there was to be an announcement about OS/2. MS may have been 'small' but it was powerful because it owned DOS and everything based on it. OS/2 was a joint endeavour and MS could ALWAYS have some control over OS/2. And it still does as it jointly 'owns' some of the code base. This is wrong: IBM does not fully 'own' the OS/2 code. IBM still has intellectual property rights restrictions on OS/2 which preclude it from doing certain things with parts of the code (mainly in Presentation Manager code, some of the TCPIP layer code, not so much in the LAN code). Among other things IBM probably cannot release the IBM has lots to be sorry about, but I do not think that IBM ever 'attempted to destroy them' (MS) in the manner that MS acted towards its partner (IBM). The saying is, that it's not the size of the dog in the fight, its the size of the fight in the dog. MS is ONLY a software company. IBM is a hardware, software and consulting company. Take out the big hardware, take out the big OS software (OS/390's, RISC based, or AIX specialities, etc) and take out the consulting (mainly aimed a big ticket networking and big boxes) and you are left with a 'desktop software' arena in which IBM is the mouse in the arena with the elephant....even if you include Lotus Notes/Domino. Geoff ==== You won't convince him. Jacek is the very worst kind of apologist for Microsoft. He endlessly quibbles about this tree or that tree and carefully never, ever, looks at the forest. With such a mindset no honest discussion is possible, he is right, and everyone else is totally wrong and just picking on Microsoft. As if MS were the injured party here. With the metaphorical battlefield strewn with bleeding com- panies even now being savaged by the rabid jackel that is Microsoft you would expect some concept of common sense or fair play, but you would be wrong. Jacek, it would appear, believes a company should be allowed to do whatever it wants to, legal or not, and anyone who com- plains about the results of such a thing is obviously just picking on the successful company. Deny, deny, deny, obfuscate, deny and lie. It worked for Clinton, it's working now for Microsoft, and as far as Jacek cares, it's all okay with him. -- .-. .-. .---. .---. .-..-. | Wild Open Source Inc. | |__ / | | |-< | |-< > / | Making the bazaar just a `----'`-^-'`-'`-'`-'`-' `-' | little more commonplace. home: www.smith-house.org | work: www.wildopensource.com ==== You mean is not for free ? Mercedes Benz is also expensive. Do you hate them too ? Judge ? Judge is a noble profession that requires fairness. Judge Jackson is a disgrace for the profession and should be removed from the job. The so called judge denied every Microsoft document that was proving that the Justice Department accusations are false, and allowed himself to show in comments what kind of fair judgment Microsoft can expect from him. He-he-he. Since when public rumors and street myths are evidence ? You mean we don't need no stinking trial anymore to hang somebody ? Lynch was popular between uneducated masses in the past but rarely produced fair outcome. And you are a lawyer ? Legally IBM missed the date of the deal and was obligated to pay whatever the new deal required. There is no law that requires same price for the particular product all the time. Check your local gas station. Nonsense. It was IBM leverage over Microsoft to kill Windows by putting OS/2 together with Windows on the same machine in order to switch the users from Microsoft to IBM OS. In fact I wouldn't be surprised if Microsoft in fact demanded that the buyer has only choice of either Windows or OS/2 but not two of them together if they sell Windows under special price to hardware manufacturer. Otherwise IBM should pay street price for the Windows. Special deals are for those who do not openly plan to ruin Microsoft. There is no technical reason for Microsoft to provide dual boot especially for the product that suppose to put Microsoft out of business and is recommended to use different disk formatting. FAT maybe was allowed for OS/2 but it was an inferior disk format and was not recommended. Also why you demand dual boot from Windows and not from OS/2 ? Let IBM to spend some money and develop boot to OS/2 that gives user a choice of Windows if he wishes so. IBM is at the same fault here as Microsoft or even grater, because it was IBM who was trying to put new product on the market. Dual boot is an special option that Microsoft sometimes put into it's OS to provide user with possibility to run different versions of Windows. It's a feature, that Microsoft is not obligated by any law to provide to anybody else operating system. In fact I think, if Microsoft is selling it's product in a special deal below street price, it is allowed to put a restrictions of the usage of the product, that the buyer can accept or reject. In fact IBM by being manufacturer of OS/2 and the hardware had all kinds of restrictions against Microsoft. Initially they did not even bother to put Windows as a choice at all :-) So why are you complaining ? If you've chosen bad product and you still refuse to see it, it is mainly your problem :-) It doesn't matter that BetaMax was technically a better system than VHS. Those who bought BetaMax can blame mainly Sony for policy and themselves for the mistake. No, you missed the point. It was much more powerful company that was trying to destroy Microsoft competition by putting their own product into their own hardware and trying force out Microsoft from the market by using their power in hardware and software manufacturing. It was pure luck, that Microsoft survived that onslaught mainly because IBM never really learned how to sell competitive computers on the mass market. You had a choice when you bought computer. You've made OS/2 choice, rest of the world mainly Windows. In fact even if you didn't like the original choice it was just a matter of walking to the software store to change to choice. It is a nonsense to demand from competing companies to develop the feature that simplifies customer to switch to competition. There is no law that requires it and it is a nonsense from the business standpoint. and I did not heard that Microsoft ever did it again. The problem with IBM was, that they were arrogant company that could not compete successfully with new bread of high tech companies. When they started OS/2 with Microsoft they quickly leaked out the plan to eliminate Microsoft, then after divorce with Microsoft in their stupidity they did not realized, that they need Windows together with OS/2 from the day one to convince customers and they never managed to advertise and make deals with their competition to make OS/2 main OS in every PC that was coming to the market. At the time only and idiot was buying overpriced IBM's. Microsoft survived, come out later on the market with much better OS than OS/2 and destroyed IBM plan to rule the world. I can only imagine what IBM would do to the PC hardware architecture if they also owned OS. You would not buy PC for $2000-$3000 except Mac. Every company makes new products secretly. It is called competition. That is why magazines pay so much money for spy photos of the new incoming products. If you consider competition as back stabbing, then I'm sure you must agree that Lincoln Navigator was a Ford back stabbing of BMW :-) IBM by the fact of owning significantly more resources than Microsoft would slowly overtake OS/2 ownership with every upgrade of the OS. Products belongs to those who invest more in to it. I'm sure Microsoft would be left with bits and pieces here and there that have no significance whatsoever and later would be just replaced with add ins to the OS from different vendors. I've heard different opinion, but it was not a reliable source, so you might be right. IBM should replace those parts of the code with their own programming. They had enough resources to do so. on that subject about a plan of loosing Microsoft slowly later from OS/2. I/m sure, it was not a plan on paper, but rather a general business direction that IBM workers felt or knew about. IBM did not need Microsoft at all in the long term. They needed them only to smooth transition from Microsoft OS to IBM owned software in the future. That smooth transition was denied by Microsoft whose main business and future is to make software. So clearly it has much more resources that can use to extort anybody else. You cannot take big hardware out from the picture. It gives IBM money, programing and manufacturing resources, market influence to dictate standards, and of course experience in different parts of the business. It is not a coincidence, that today PC was basically defined by IBM with the few mistakes. But thanks to Microsoft :-) Otherwise you would only buy IBM with OS/2 today for let's say business type price of $10,000 ;-) as IBM is known to price it's products, that they do not have competition. Jack ==== I'm sure, you would be happy to hang Clinton or burn him on the stake because he was so IMMORAL. It's so typical for the crowd. Jack ==== I did not say 'free'. If it were a lot cheaper, it might be a better deal, but as it is, it does not work as I expect a OS to work, and therefore it is too expensive. And I do not hate Mercedes or Porsche because their cars are expensive.. But if you want to buy a Yugo $70,000 US, I *will* sell you one! What public rumours and street myths? The Washington Post, among other sites, obtained and posted the transcripts of the depositions of all of the people who were examined. Only a limited number of witnesses were called for public viva-voce evidence. Certain of the depositions were referred to in court but not all of them nor all of any one of them. Every person who was deposed, was sworn, upon penalty of perjury to tell the truth, and was examined by lawyers for both sides...And you think that that is 'rumours'? Ah so, now MS was lynched! Had the right to counsel, to examine witnesses, to name its own witnesses and to appear and argue the case, to argue as to the penalty (which opportunity it ignored) and to appeal (twice).....Some lynching. Yes I am a lawyer. There is a law which may require the same price for a particular product. That law is the anti-trust law.. which as you should be aware, does not operate prospectively but retrospectively. It is not only what someone does which gets them into trouble, but what they intended by what they did... and that is the great difficulty with anti-trust law. Ha Ha Ha... If IBM had any leverage, it would have ended up *either* not putting Windows on any of those machines (which would really have screwed MS if IBM in fact had had that leverage) *or* IBM would have ended up paying LESS than Dell for Win95 (because MS was sooooo anxious to get Win95 onto those machines). Get a real life, or better yet, get a grip on reality. Boy you really do have cause and effect backwards... But one further point: although MS may have ceased requiring OEM's to pay equivalently for every computer shipped (a fact I doubt), it has merely changed its tactics: it now signs agreements which REQUIRE the OEM to ship Wincrap code on EVERY machine. The OEM cannot remove it, will not remove it, and the customer ends up paying the 'MS TAX'. Don't believe me: try to buy a bare Toshiba notebook. It cannot be done (and the OEM agreement is so secret, that MS was not required to publicly show a copy of ANY OEM agreement at the DOJ trial...even with portions redacted... Another poster was correct: you are an apologist for MS and are highly unlikely to believe or accept anything which denigrates MS in any way. Your reaction about Judge Jackson is telling but you still cannot monopolist STILL STANDS. I would rather talk to a brick. It does not spout jibberish back at me and when I want the conversation to end, I can throw the brick away. This thread is dead. Geoff ==== (just like everybody who works for M$). Everything he uses is the absolute best and he won't hear anything bad said about them. I think he fits some of the conditions below: 1. works for M$ 2. owns stock in M$ 3. works for a company whose life depends on M$ 4. owns stock in a company whose life depends on M$ John Edry ==== How typical. I thought you are better than this. The fact, that you had never seen professional CAD system doesn't surprise me. I assure you, it is not made by Microsoft. Wrong Wrong Completely wrong Nonsense. Does you knowledge of Microsoft wrong doing has the same foundation in facts ? Jack ==== That is exactly what you don't understand. If you want cheap Yugo You mean, those who judge Jackson allowed ? I'm not surprised. I know that it was limited. All Microsoft defense materials were rejected by the judge. Deposition is still an accusation yet to be at least explained or proven. Witness might lie or might not know all the facts. Microsoft was planning to provide documents and witnesses to prove it's innocence. It's a shame neither judge Jackson bother to allow them to trial nor Washington Post to publish them. On the other hand, with your assumption of Microsoft guilt you probably wouldn't believe it anyway. You must missed something in the discussion. The word lynch is pointed toward people like you, toward Microsoft accusers that had never even got into detailed documents and have either minimal or none knowledge regarding business practices and law. Court system that found Microsoft guilty and in the legal process found appropriate punishment for the company is no lynch by any means. It is people like you that demand higher punishment based their opinion, than the court actually had given to the accused or demand of punishment without the trial at all. Something like I see here in your postings. I suggest you start suing oil companies. They not only vary the price of gasoline on the time basis but even at the same day the price is different for different customers depending where they live within the same area of distribution. And tell me when Microsoft had different street price for selected people depending where they live and when they buy the product ? But if you are a lawyer so sure about Microsoft guilt, why don't you just sue them on the civil trial ? Really ? IBM did not put Windows on their machines initially. I wonder, when people will start demanding to remove name David from the Bible because he was dirty immoral Jew (sarcastic) that killed poor Goliath using nasty tricks. I suggest make a Microsoft free zone around yourself. Believe or not it is still possible despite of falsely accused monopoly of Microsoft. You might find problems with the others who refuse to create software or hardware for people like you, but here you have to believe me, that there is nothing personal in it. If that disappoints you sue them. Microsoft has no influence on those others who refuse to support anything else beside Windows. Strange how people as so blinded by hatred that they do not remember even the simplest facts. Boy, what a nonsense. Toshiba sells Windows on all machines because they do not rely their business on people like you. They simply understand the business of successfully selling computers and are not interesting in misniscule market of misguided people unhappy with software package based on Windows operating system. If you want no Windows buy computer despite your accusation of Microsoft policy. I find it primitive to use name calling instead of facts and arguments but on the other hand it is so typical for so many Microsoft accusers. About the punishment - IT WAS GIVEN. Microsoft paid the fine. I don't understand, how lawyer like you cannot see the clear facts. Are all lawyers like that ? Go ahead, make my day. I'm sure the brick always agrees with you so you can be happy of being always right. Jack ==== I'd have been quite satisfied with kicking him out of office and throwing him in Federal prison for perjury. -- Wayne Brown | 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 ==== I did not liked Clinton either, and I wish him jail too, but I prefer to keep feelings separated from the judgment. Otherwise there is not much there that makes us different from Holly Inquisition. One should withhold the judgment without full knowledge of the facts. There was so much talk about prosecuting Clinton after he leaves office. Where are the prosecutors now ? Jack ==== I have used CV Cadds 4x (back around 1984-85) and Intergraph I/Ems. Are those professional enough? Pro-E has come a long way in the last couple of years. If you ask some Catia or Unigraphics salemen, you'd find out that they consider Pro-E a good competitor, and on level with them now. John Edry BTW The US Army is going Pro-E, along with several of the it's armored vehicle manufacturers. I strongly doubt they would do that if Pro-E was a toy . ==== Come on, that would really be a big joke in Europe. Clinton was human and very good for Ama. Your current president, Jack Spade (or is it Bush?), is a disaster. What if Clinton lied? Big deal, it wasn't important, and it would never have been brought up, if it wasn't for vengeful do-gooders standing in the wing waitng for their share of the carcass. If you look back on Aman history, the Clinton ordeal is a ultra-dim blip on the radar. I don't wan't to discuss this again though, since it's not that many months since it was last debated here. ==== I don't know. I'm not familiar with those. I did not said that Pro-E is a poor CAD. I said, that it is not that strong as CATIA or Unigraphics. CATIA and UG are much better CAD systems than Pro-E, but they are much more expensive. That is why little companies that do not have resources or need for stronger, expensive CAD use Pro-E as a cheap ersatz of industry strength CAD. Pro-E is a better choice than Autocad, but it is far cry from CATIA or UG. There is no coincidence, that companies like: Aerospatiale, Dassault, Boeing, Lockheed Martin, Sikorsky, Rayethon, British Aerospace, Fiat, Ferrari, BMW, DaimlerChrysler, Porsche, Renault, Volkswagen, Cosworth, Scania, most of the Formula 1 teams, Black & Decker, Frigidare, Samsonite, IBM, Sony, Samsung, use CATIA CAD system and companies like: 3M, Apple Computer, Textron, McDonnel-Douglas, Rockwell International, Osram Sylvania, Gilette, General Motors, General Electric, Subaru, son, Aiwa, Fujitsu, Minolta, Seiko, Robert Bosch GMBH, Siemens, Philips use Unigraphics for their product design. Of course, there is always possibility, that departaments that choose CAD system for those companies employ idiots and you are right :-) Could you please, tell exactly who from the vehicle manufacturers is going Pro-E ? Last time I've checked, Aman Motors with it's HUMVEE and General Motors with it's armored LAV are using full time Unigraphics. Also manufacturer of Abrams tank - Genral Dynamics is using either CATIA or Unigraphics depending on the departament. None of those companies is even considering Pro-E. Please, tell me, what exactly is desinged and manufactured by US Army ? Jack BTW Have you ever seen CATIA or Unigraphics ? x-mac-creator=4D4F5353 ==== Virgil - My Mac, like yours, has serial ports for printer and modem, and I almost never use the modem, since I've got cable, so that leaves a free serial port. These serial ports take nearly round connectors about a third of an inch in diameter, with 8 pins in them. But the serial port cable that came with the calculator is different--looks more like a small monitor cable, 9-pin, a row of 5 and a row of 4, probably about 1.6 inches by 0.5 inches (just guessing--not looking at the thing right now), which apparently fits into the serial port of a PC computer. So what is it that you plug into your computer? Do you have a PC-serial to Mac-serial converter cable thing? And do you need any software to make it work? And what software (if any) do you use to get stuff from the website onto your calculator? I found something called Link 48 installer, but have to confess that I'm not sure what it is. As you might guess the calculator is new, so I'm trying to figure all this out. -Doug Magnoli ==== The cable is one that I got from an HP outlet along with the Link 48 program, which also works with the 49. The cable connects to the HP48, and comes with an adapter which allows it to connect with an HP49 also. It seems to have een made in China. Down at the serial port end of the cable is a number, 8120-6735, and on the HP49 adaptor is a number, 1252-6635, over the word CHINA. It all came together as a HP48/49 to Mac adaptor kit. I cannot remember from which HP outlet I ordered it. but you might try http://www.wholesaleproducts.com/hpcalcs.html Or the HP corporation itself. Note that this connector and software do NOT allow you to upgrade your flash ROM. For that, find a nearby owner with the upgraded stuff on his HP49, and get it there. ==== Using Eqlib 1.00 on HP49 1.19.6 i get an error message STO error: undefined XLIB Name. Whats wrong? I guess i made a fault installing BZ. But what is the fault? ==== There are a few lessons here: 1. Not every expression has a symbolic solution; 2. Many expressions have num solutions, but the num solutions are often, at best, approximate; 3. Some solutions are transcendental; not just irrational. This is one (two) of them. A transcendental number is one that cannot be expressed as the root of a rational expression (one polynomial divided by another). Pi and e are transcandental numbers. Dan ==== There's one more: I don't know how to use MuPAD :) The symbolic solutions actually exist and have been found by a bug-free release of the 49 CAS (see other posts in this same thread) ==== Marco, what are you saying? I don't see any symbolic solutions in this thread. If there are any, I'd love to see them. I contend that the solution(s) cannot be expressed analytically: as a rational expression or even as the solution of a rational expression. What I do see are two *solutions* 1.07257978557 and -1.17749189498. These are approximate solutions: accurate to 11 decimal places. Dan ==== I've a made up a little confusion about it... there are actually two different problems, but I didn't notice, hence the mess :) ==== Yes, Dan, those are the only solutions we have found... ==== I am using the 3.2 version of erable for hp48 and I have noticed that the factors command gives some strange chars in screen instead of correct answer.Is that a bug or is a flag issue? If someone can help me to correct this it would be nice ==== since the 5^x equation didn't worked, i tried 5^A just for the heck of it. to my surprise, i got an easter egg with a twist :0 ==== gone :) test! I don't get a TTRM, but an error instead: SOLVE Error: Not reducible to a rational expression ==== You have to enter the expression exactly as given, including the typo (A instead of x): '5^A+7^x^2=15'. Using SOLVEVX (with VX='X') on '5^A+7^X^2=15' nukes the calc as well (tested with emu48, I don't want to try this on my real calc). Looks like this could be a real bug in 1.19-6 there. Maybe Bernard Parisse could take a look at this? -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag ==== I just tried SOLVEVX(5^A+7^x^2=15) in the CAS on my HP40G. Crash! Reboot and memory clear. -- ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== I do, and this is on the real calculator. I have entered it with the EQW, and the equation looks exactly like this: '5^A+7^x^2=15'. No additional parantheses or anything. I get the same result as with 'X' SOLVE - no TTRM, but merely an error. Maybe it's a bug in emu48? My flags are: {#2484038C83250FF0h #0h #1010400A200008h #0h} ==== I tested it with your flag settings on a real 49 w/ROM #1.17-2, and it gave back an empty list in RPN mode, and a more complicated expression in ALG mode. , Raymond ==== I'll do it at the beginning of next year, I don't have a debugging PC here... ==== $Zm5.522471@news000.worldonline.dk... Don't you think that it would be a bug anyway? 1.18 gives symbolic solutions, after all. I don't think so, I tried your flag setup in emu48 and got no TTRM: it puts the calc in complex mode. Switched it back to real mode and -bang!- I got the TTRM again ==== Nope, I tried with my real calc running 1.19-6, and it does the same (100% reproducible). With your flags I also get the error message, so it's obviously an issue with a certain flag combination. Just for the record, here are my flag settings under which I get the TTRM: %%HP: T(0)A(R)F(.); { # 2080038A04010FF0h # 0h # 801000000A000000h # 0h } I'm too lazy right now to figure out which specific flag settings are causing this, but I noticed that if I turn Complex mode on (like you have in your flag settings), then I get two complex solutions instead of the TTRM or the no rational error message. -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag ==== in the HP40G too? that is really strange. lets try 'the cobra equation' on a TI-89 :) it may kill it ==== A hole row of buttons on my HP 40G don't respond anymore (the one with the buttons alpha, 4, 5, 6, -).... Has anyone had the same problem? Because HP ceased his calculator section, does this mean I lost my guarantee on my calc? I already switched the batteries, but the problem persisted.... Has anyone encountered the same problem? Is there an easy way to fix this? thanks for your help, David ==== This is fairly rare but is a problem which HP is aware of. If you take it back to the place where you bought it then you should find that they will replace it under warantee. -- ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== Just to clarify: I can't speak for HP - I just know that we've had a few calcs (out of hundreds) replaced in our school under warantee for the same problem. -- ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== I just wanted to confirm with the knowledgeable people here if the cable for the 49G/48G (Part Number: F1897A#ABA) can be used with the HP39G. HP's site does not mention the HP39 being compatible and I have contacted HP on the phone and they are clueless on this. The HP39G manual refers to the cable and software as the connectivity kit but gives no part number. I am sure it works but I just want to make sure and not damage my calculator by using the wrong hardware. Also I found that Samson (www.samsoncables.com) sells just the cable (Part Number SCHP49) for $15.00. Samson states that the cable can be used with the HP39. Has anyone used this particular cable successfully on an HP39? I would prefer buying the cable only since the software can be downloaded free of charge. FlyerX ==== The number embossed on the serial plug of the cable that I use with my 38G, 39G & 40G is F1897-66000. It does suggest that they're the same cable but it doesn't seem to match exactly what you mention. Can someone more knowledgable comment? Personally I'd experiment. I've accidentally used the wrong cable before on 38Gs with no permanent damage. The screen blanks out but when you unplug and press ON the calculator resumes with no trouble. Perhaps I've been lucky but this has happened a number of times. Mind you, I feel safe in saying this since I'm in Australia and you're not likely to look me up and punch me in the nose if I'm wrong! :-) You can also get the software, along with aplets, from my site. -- ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== The cable from Samson Cables will work with your 39G. The 49G and 39G use the same cable. A cable for the 48G series only, should never be used for the 38,39,40 or 49 seris as it was originally designed to work with the Hp100 and Hp200 lx palmtops as well as the 48. This allows 12 volts through its extra pins into the the 38,39,49 calcs and CAN DAMAGE THEM. I too experimented like Colin and had the very same effect with no apparent damage. But others have not been so lucky. I bought the right cable from Samson for $25 and have been happy with it (after I experimented, like Colin). , Ron Ross ==== You actually can use the 100/200LX cable w/ adapters 5181-6643 (LX --> 48) and 1252-6635 (48 --> 38/39/40). Quite useful if you've a 200LX and a couple of different calcs... Bye, Detlef -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel ==== This is the exact cable setup I use... (db9--10pin)(10pin--4pin)(4pin--10pin). Works great, very flexible. Matt ps. [OT] Detlef, are you still in Corvallis (Firmware Systems)? Just curious. ==== I use my 48GX for navigational purposes, among other things. While the clock is not essential, it's handy. So I have begun to plot its accuracy...using www.atomtime.com .I wonder what accuracy I should expect, and the nature of the error...be it systematic or random? The error in my PC's clock (and most PCs) can be quite significant (a minute or more in a week), or it may keep near perfect time for a couple of weeks running. Lee Martin ==== Glendon schrieb clock of most ... The last experience I have made with a HP49G. I would say the nature of error is systematic, depending of the ambient temperature of the calc. or its clock. The clock slows down with decreasing temperature. How much? I hope, while continuing your plot, you will find out (tell me/us). I estimate it is in the order of 0.015 seconds per day and Kelvin s/(d K). I have never tried to figure out the difference between: keeping the calc. in hand and let the calc on the desk. Or running large programs, to heat up the interior :-) And it could happen, that the calc hangs up and the clock is sticking until you reset it. There are several programs for automatique adjustments available on http://www.hpcalc.org/ ..Heiko ==== be No. If no MK, install ERABLE1.LIB in port 1. works fine on my GX+2 MB Ram. ==== I want to buy HP48GX. What Manufacturer P.N.: 48GX#ABA means? Are those the latest models? Janek ==== AFAIK,...ABA just means something like 'Aman version'. Happy New Year, Raymond Jan Piaszczynski schrieb im Newsbeitrag HP48 PROGRAMS, CIVIL ENGINEERING, ANALYSIS, ON 1 CD. MAIL ME! WWW.GRADJEVINA.CJB.NET ==== After successfully downloading 2 programs from my PC to the 49g, I now seem to have a problem. The PC engages the 49g because I briefly see some sort of a packet being exchanged on the 49g screen, then the PC displays the can't make a connection with the Calculator, press right shift and right arrow to set the calculator into Server mode. The Calculator however, is already in Server mode. The physical connection seems fine and I've used the same PC for the previous successful downloads. Any ideas? Martin. ==== it is probably in your configuration. HP49G: check in apps, i/o, transfer settings-> kermit,ASCII, 9600, etc COMPUTER: check the settings on your com port: hardware devices, com-port and in the connectivity program also, try sending data (XMIT) to the hyper terminal after the configuration. if all fails, do a hard reset in the memory of the HP49G. good luck! Blennerhassett says... ==== Ok, it's late, but I found a Jornada X25 photo.. See it at: http://www.smh.com.au/icon/0111/28/news2.html Nice machine... J.Manrique CdU de la ETSIG ==== zz921174@opalo.etsiig.uniovi.es (Lopez de la Fuente, Jose Manri) Speaking of STRONGARM processors, I wonder how much STRONGARMING went on by MS to make HP drop a LINUX PDA. ==== Why, nothing -- HP can do than all by itself. Probably the Singapore PDA division acted sneaky in all this... Bye, Detlef -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel ==== Only a question: Is absolutly impossible a new Erable for the 48? ==== I'm glad to inform you that I've just finished v1.7 of IFB. Many have complained that it had a bug when resizing objects, and that was quickly confirmed by me. Fixing it, on the other hand, has been very difficult - one of the obstacles have been the fact that IFB is contained in one single source file, which is around 50kb in size (I won't do *that* again). I believe I have spent maybe 40 hours on fixing this bug, but I felt that I owed it to the many people using IFB. So, it's done. As always, if you want it sooner than gets his site The bug turned out to be a missing DUP BTW :-/ ==== Great!!!!!!!!!!!! Arnold ==== Using your library Polinomial v2.1 I found a possible bug. Just try: [ 1 'A/5' 1 ] [ 1 -2 ] PDIV What I get is: [ 1 '{ 'A--10' }/5' ] [ '1--2*( { 'A--10' }/5] [1 -2] not a valid polynomial. Have you seen this? Saludos Jorge M. Valenzani Oh my God, I can't belive what I read in PocketPCThoughts about HP cancel of Jornada X25, just some: The sub-$200 market is a volume market but is too crowded and profitability is negative. Going forward I believe users want more, not less. I understand the decision to stop, or at least hold, the project based on these factors. In the light of more overcrowded soon too, thus making competition even more fierce. Exciting! competition in the under $200 market would probably see it do poorly, which is probably why HP killed it. Besides, to be it isn't Microsoft, not because it's some revolutionary OS. I guess thumbing your nose at the leader is some sort of rebel is too unsecure to ever be a major player. It may be fine for a PDA like the X25, but not for the desktop. What do you think? You coudl see more at: http://www.pocketpcthoughts.com/archives/00003457.php#comments J.manrique CdU de la ETSIG ==== I don't know.This could have been a very interesting tool with the approrpiate softwares. However i don't like keyboard less tools. ==== it is hard to speculate on this with the ipag floating around. if i were to buy one rigth now, i would go for the ipag. i think you can run java (with all the free CAS sofware made out for it), linux (?), msdos and pocketPC. but i am not buying anything :) so it doesn'y matter. =?ISO-8859-1?Q?Timit=E9_Hassan?= says... ==== in more secure than Windows (in any flavor). Charles Perry P.E. ==== year old rewrite of unix which itself is over thirty years old now. But that indeed is its strength. In the security stakes any unix version will be more secure than any MS OS. No MS OS is designed with security in mind. The primary motivation for any MS OS is proffit. For MS ecurity is only an issue in so far as it helps sales. To MS the appearance of security is far more important that the reality. Open source software must be more secure (and consequently is) because the open source community take it as a challenge and take a certain pride in their abilities (which are considerable). If you want security then use Open BSD, probably the most secure OS on the planet and this is open source. A security consultant told us that there has never been a major security update for Open BSD. Just as an illustration - who has written the best SW for the HP calcs over the years and why? Apart from security there are many other advantages to using a unix OS which I do not wish to go into here. However the arguments for unix are probably not so compelling for the PDA, except for stability. Stability is probably the biggest issue for MS OSs on the small to medium devices (embeded, PDAs and desktops). because it has captured peoples imagination. There are better varieties of unix but any unix is better than any MS OS and people are beginning to realise this. This is getting RIGHT OT so I had better stop before I get roasted. ==== I don't think that he means unsecure in the sense of security but in the sense of market viability. Sure MK was a freeware. I don't think that there are many other advantages to using unix OS besides security or stability.Even stability is not really an advantage as Win2000 and WinXP seems quite stable to me. Server and other stuff like that.I strongly believe that MS OS are much better for the majority of users and for commercial softwares.Freewares are great for the users but not really for the computer business. ==== It is this way of thinking which will guarantee that we will still be using MS OSs ten years from now and loving every minute of ride. You would have thought that by Widows XP they would have made the OS secure. Instead it is an open invitation to hackers. Furthermore, in the follow-on to XP, you won't be able to buy the OS. You will be required to make yearly payments to MS or your OS will time out on you. That is the way MS defines customer loyalty. happens. ==== Hmmm. My first thought went to JAZZ, as it's open source, Just to correct this: MK wasn't freeware in the first years. You had to pay for the module, and the 'Demo' version was crippled. It became s.t. like freeware some months ago. And of course it's still not open source... Yes, WinCE 2.x can really be annoying sometimes;-) As always, it's good to have a backup. I used NT4 for years, and from my point of view, it's stable. I have an HP-UX workstation running UX 10.20, and from my point of view, it's stable, too. but that doesn't really count for private use;-) For most things, the NT machine is easier to use than the others, and that's what really counts for most people (IMHO) I know many people who aren't computer specialists, but don't have problems setting up a running Win** installation. use... , Raymond ==== There was some discussion a few months back regarding the HP49G keyboard. Was there any consensus that the new calculators have a softer keyboard ? My serial number is ID94702241. Although I only bought it 5 months ago, apparently this was manufactured in week 47 of '99. I find the keyboard heavy. Unfortunately nobody around me has a similar calculator for a comparison. As mine is still under warranty I would consider sending it back to HP for a more recently manufactured one if there is really a difference with the new keyboards. Aubrey ==== Is HP honoring warranty on let's say calculators bought new but manufactured over a year ago? If so I am sending mine in for a new Chinese version! Santos Lucero ÀProblemas con las news? Webnews, el acceso m207s r207pido y fiable http://webnews.aforo.com User IP: 207.212.230.83 With the downfall of HP I was wondering if Jean will come out with MK3.0? ==== I guess you're talking about the version for the HP48. The answer is : Surely not. Shortly after version 2.30 we ran out of space. So with the huge technical problems, my lack of motivation to just even think about it.... You get the picture. ==== Ricardo Blasco Serrano escreveu na mensagem <3B71ACCD.6357B7E1@teleline.es>... I agree too :) ==== My program is: Ç 1 1000 FOR X '6*X+1' Eval IF PRIM? 0= THEN HALT END È 1: Isprime is from Alg48; 2: on HALT, it goes to the HP default stack. What to do? ==== ValuComp It may crash and burn, but at least it is cheap. Charles Perry P.E. ==== BellyUp Inc.? ==== Mark Ringrose escreveu na mensagem <1009143386.2427.0.nnrp-10.c1ed3e79@news.demon.co.uk>... HPaq? HPCO? HewCom? ComPackard? ==== You will need to get your cable working to be able to upload the new ROM to your HP49G. The instructions for how to load a rom can be found in www.hpcalc.org, the full URL is: http://www.hpcalc.org/install.php I used these instructions, all very clear and consise but you must follow them line by line ! I'm somewhat surprised that you have this low ROM version.. what is the serial number of your calculator ? , Aubrey. ==== back in Oct01 I picked up a 49G from an Office depot. It also had the 10c rom. s/n ID95200668 ==== I'm a Clinical Epidemiology grad. student and, as one would imagine (Epidemiology), I do a tremendous amount of statistical analysis (linear and multi-variate regression analysis, correlation (uni- and multi-), ANOVA, confidence intervals, hypothesis testing, probability, etc.). I am also required to do some small degree of financial analysis (but a minor concern). I often use computer programs such as SPSS and Excel (for spreadsheets). I would really like a graphic calc. to help me and have been looking at the TI-89 and HP 49G. I would **really** appreciate your insight and help as this is an area in which I have virtually no experience. 1. Which of these calculators would be more helpful in calculating the type of statistics I mentioned above? 2. To what degree can you create simple spreadsheets? Is it possible to download Excel worksheets to the HP-49G or TI-89? I have heard that the TI-89 has a better spreadsheet function, is this true? 3. Does one have better statistical programs? 4. Which has better support? (What are people's thoughts re: the HP49's future what with the R&D wing of HP's calculator division closing?) 5. Which one would you suggest and why? The TI-89's larger & higher definition screen,and 12MHz processor are nice. The HP-49G on the other hand does have more memory and it sounds like (from some of the posts I have read here) certain other advantages. When it comes to graphing calculators I have virtually no experience. Really, any and all help in making my decision would be *greatly appreciated*! Lisa ==== Here you'll get THE BEST for Statistics: http://www.mtsac.edu/~sguth/ Download Stat48Pro or Stat49Pro and read the documentation. You can read these ones on line too: http://www.mtsac.edu/~sguth/stat48pro/Stat48Pro.htm http://www.mtsac.edu/~sguth/stat49pro/Stat49Pro.htm Then you will decide by yourself. If this helps, I'm happy. ==== Lisa-- If you go with the 49, i would reccomend getting Gilberto Urroz's book on science and engineering with the hp49g. Im not sure if its volume 1 or 2 (i think 2). But 1 of them has a section on statistics. I had a ti89 before i got the hp49g about 2 months ago. The ti89 is easier to use at 1st. The hp's RPN format takes a little getting used to. I haven't used mine for statistics. But i would highly recommend getting at least the one volume if not both. The manual that comes with the 49 is a piece of *shit*. You can purchase the book at greatunpublished.com. Then do a search by the author. I have both volumes and they are a reliable company to buy from. The ti89 manual is a pretty good one(that comes with the calc). Good Luck Brandt ==== My vote goes to HP49. The 49 supports spreadsheets both as is and with 3rd party applications. I have no knowledge of the TI89 but I would be surprised if things were different there. The existence of this newsgroup played a major role in my decision to get a 48GX. I have no reason to believe that HP wont honor its guarantee in case you need it. HP49 (because I am biased :-) This should not bother you. The TI has a faster processor but the HP is more highly-tuned, meaning that its software makes better use of its resources. The result is about equal. Neither did I before I got my 48GX. One thing you should know is that the learning curve for HPs is steeper at first, until you get used to RPN. Afterwards however, you will be more productive. Xronia Polla. X-Inktomi-Trace: arninkcluster 1009357671 21435 62.163.50.11 (26 Dec 2001 ==== What is the normal usage of batteries for my HP48G. If I use Duracell it is about 4-6 months, cheaper last shorter. This is the ultimate battery test;-) TIA Klaas k.s.leguit@hccnet.nl ==== is test;-) Both my HP48GX and my HP49G uses a set of Duracell's in around a month, and I've had cheap batteries run out in 5 days on the '49 :-/ ==== in my HP49G i measured the current a while ago. it was about 6 mA at iddle. when you 'press' any key, the current goes up to 18 mA (i used a cheap meter). that means that the more key preses you do, the shorter the life of the bateries. duracell website has a graph (hours- current from AAA) that indicates how long batteries last. says... ==== I suggest you consider this library: http://www.hpcalc.org/hp48/math/symbolic/alg48v42.zip and use the (unlimited length) integers in this library. PLEASE let us know if your algorithms are more efficient ;) Best Gjermund Skailand John Steele skrev i meldingsnyheter:QjkM7.9104$pP5.1188818@news1.rdc1.mb.home.com... effects for -- -- ==== As I see you are using erable 3.2 Are you experiencing problems with the use of Factors command? See you ==== Yes, as you say, FACTORS only returns rare things... but I don«t use it. Since I install this version (near two years ago) I only use FACTOR (Erable) and FCTR (Alg48). If you need more information, I will ==== it warmstarts infinite, so i do ON+A+F to reset it. ==== Raymond Hellstern escreveu na mensagem ... Ok. And to RPL on PC, not direct on 48 :o)? And docs about sysrpl/asm, where?? ==== for SysRPL programming on the PC, I recommend the HP Tools, which consist of SASM.exe, MAKEROM.exe, RPLCOMP.exe, SLOAD.exe, along with their sources, some header files, an entry table file, some examples, and a bunch of documentation. There exist at least two versions of the HP Tools, one 16 distribution, which you can use from DOS, and a 32 bit distribution for use with Win32. and I know at least one person with an OS/2 32 bit version. You can get the Tools at www.hpcalc.org There are many supporting documents around the Tools and SysRPL/ASM on hpcalc, too. For editing your files I'd use UltraEdit, or TSE/32. Both are very powerful but easy to use text editors. Notepad works, too, but then, I'd set the font of Notepad to monospaced (non-proportional). Last, but not least, I recommend to use DDE48 for easy transfer of your developed binaries to Emu48. ==== where this information is kept. The thing that really keeps hammering in my head is why when completely reseting the calculator is the WSLOG not erased. Is it saved in some special memory or ...? Gregorio <3C299E96.B0D429E0@public.mobidig.net> ==== According to the 49's RAM map, it's saved at the memory locations FAILSTK1 through FAILSTK4, starting at address #80010h and taking 18 nibbles per entry. I guess it's about the same on the 48. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ==== *This* goes directly to my greek kernel! No winner => Everyone is winner. Like the old olympic spirit, isn't it George? Participate and you already won! (But they had judges, anyway ;-) ) Spread the light to the world, ==== well, um... i think i was a little bit offensive earlier, i want to apologize for some words i said. Shouldn't do it. Won't happen again. In any case it seems people aren't very interested in no-goal no-prize 'contests'. I would like to just see what people can come up with, without thinking about winning something. I am dissapointed by it. greets, gT ==== Actually, I knew that the HP used the cordic algorithm to compute sines and cosines and other trig functions, plus log and ln, but I didnt know it could be used for square root too. Apparently, the HP Journal talks about this, but I've never read it. ==== Well yes, the HP8 doesn't try to evaluate *symbolic* arguments like A, B and the like when the flag -3 is clear. But it will *always* evaluate numbers, in algebraic expressions like SQRT(2) for example. There is no (easy) way to get,say 2*SQRT(2)/3 instead of 0.942809.... as a result of some calculation. Or am I mistaken here? ==== Actually, they aren't, however there's a bug in the software which won't start the recharge if the batteries are flat-out. So you have to charge them for a minute or two before you can use them in Xpander. They are standard AA batteries, just the industrial type, so you can replace them very easily if you want to. The reason the battery cover has been glued, it's because of another bug, the automatic detection between rechargeable batteries and alkaline batteries didn't work properly and the unit could have tried to recharge alkaline battery which is a bad thing. ==== schuchert schrieb im Newsbeitrag thanks a lot ! just moving from a hp48 to hp49 and have big problems to find all symbols ! *g* Christoph ==== Can anyone tell me if this company is still open for bussiness, their web site is www.dvtg.com the numbers on their site are not correct. I am trying to purchase the Electrical Eng Pro & Electrical Eng Reference combo card (13300-1A F) for my new 48gx Are they the manufacturers of this card or are they a reseller, if the later, can anyone suggest where else I can get this card from? Lloyd lecutler@bigpond.com ==== Go to Calcpro website www.calcpro.com. Can anyone tell me if this company is still open for bussiness, their web site is www.dvtg.com the numbers on their site are not correct. I am trying to purchase the Electrical Eng Pro & Electrical Eng Reference combo card (13300-1A F) for my new 48gx Are they the manufacturers of this card or are they a reseller, if the later, can anyone suggest where else I can get this card from? Lloyd lecutler@bigpond.com ==== If you are in MK's EQW with an equation, just press EQUATION (LS ENTER). . VFM escribi227 en el mensaje ==== Do you want to simply connect the points or get a smooth interpolating curve? ==== Try running over it several time with your car. This has been most effective in the past. :-) But seriously, try http://www.geocities.com/hp_upgrade/ Hmm. Why don't you make things really interesting and let people try to write algorithms to play the game. One approach could be that TETRI would take as an argument a program, which, on the other hand, would take as arguments information on the next two pieces and an array, for example: 0000000000 0000000000 0000000000 0000000000 0000000000 0xx0x00000 0x00x00x0x xx0xxxx0xx ==== We need to construct this list manually? That means pausing every time a new piece arrives on the screen :-) ==== TETRI itself does not record the game. That would make things too easy to win a HP49 :-) Clearly, one must be very clever to get a record without interrupting the game. One possibility is to create a global, REC say, isolating the key handler of the program and just adding to REC what is needed. But even for a smart programmer like you this will be not so easy because the main program is BZ-compressed inside a asm-code shell (Jurjen's trick). To isolate it one needs some extra tool or knowledge of asm. This difficulty gives the average 49-user a nearly equal chance to win the award. Wolfgang ==== Hehe - I *knew* it wouldn't be that easy ;-) ==== Probably not :-) A hint for those who not yet played TETRI. Try first successfully finishing level 8, 9 and 10, to get a feeling for the right choice. Pivo know this. He anyway has no time to read the NG now because I'm fairly sure he is playing non-stop to repeat his survival of level 11. At the time when he was lucky, TETRI didn't yet have the option Same to repeat a game, so he's a hard time to recover his game's seed ... mwahahaha ==== I have been lurking here quite a bit lately out of general interest in the world of calculators, and because I have become impressed with what seem to be very superior numal and symbolic computational abilities of HP calculators compared to the TIs with which I am familiar (although I am unsure about how TI92+ compares to HP in the numal department). If it weren't for the fact that 1. HP lags behind TI so much in screen resolution and CPU power, as well as the fact that Z80 and 68k are common architectures and I have interest in embedded development as well, and 2. HP is very unfortunately scrapping their calculator development, I'd probably buy an HP. Recently while thinking about this whole subject one day it occurred to me the idea that if HP doesn't plan any more calculator development there may be some chance to get HP to open the HP calculator architecture and ROM source, so that those who really love the HP calculator can take off with it, porting the code to other platforms, and perhaps even implementing the Saturn CPU on a FPGA or something. Forgive me if this has already been done and I just don't know it, but in any case I'd like to hear more about the current state of openness regarding the HP. Is there any chance that something like this could occur? -- _____________________ Chris Carlen crobc@earthlink.net ==== Urania asks QVSOP for help on any planet. QVSOP may refuse. See the QVSOPsrc.s, first line in QV_D.Y.n>LBR Ah, yes, and please don't change the ROMPTR order! However, there is a tradeoff between speed and memory consumption. The inner used in Urania, and would require lots of data in QVSOP's approach. Clear Skies! -- Georg Zotti gzotti(at)cg.tuwien.ac.at http://www.cg.tuwien.ac.at/~gzotti/hp/ http://www.ashproject.org ==== I've borrowed a TI83+ to evaluate for our school and I'm told that it is possible to download a flash app that will allow it to have similar CAS capabilities to the HP40G or HP49G. Does anyone here know if this is correct and, if so, where to find the app and more information on it? PS Quite a nice machine in many ways. Certainly infinitely better than the Casio I was looking at a few weeks ago. -- Colin Croft Old mathematicians never die; they just lose some of their functions. ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== Symbolic. Look for it in http://www.detacheds.com/symbolic/ Not exactly what you could call a CAS, but nice anyway. Colin Croft schrieb in im Newsbeitrag: 3C5BD028.E16F641F@iinet.net.au... ==== finished pointing that one out to me. I took a look at the list of functions that it adds to the calculator's capacity and, while useful, they don't even remotely compare with those of the HP40G/49G. Nor does it have the type of interface which they offer. Is there nothing else? I'd been given the impression that it was possible to download a really powerful CAS. -- Colin Croft Old mathematicians never die; they just lose some of their functions. ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== There's a Flash app called Symbolic: http://www.detacheds.com/symbolic/ Look under External Software for a pretty-print engine that Symbolic interfaces with. ==== There's a Pretty-Print application that interfaces with Symbolic: http://www.softheiss.de/projekte.html Powerful? No, fortunately for TI-89/92+/V200 users ;-) ==== Eerrrh, another one. But this is easy to fix. The code for back substitution and solution for the original variable in the cases where HALFTAN is used, will fail if the preceding SOLVE didn't find a solution and so returned an empty list. Now this is a big question for me. When HALFTAN is used, then the resulting thing is a polynomial (or ratio of polynomials) in TAN(arg/2). With complex and num factorization mode bpth enabled, such a polynomial should always be possible to SOLVE if it doesn't contain any other symbolic parameters except the variable to solve for. That was what I believed, but it is not true. In the next part I'll give an example for this. Anyway, the code must be changed to: IF subs NOVAL @If subs is not NOVAL =/ THEN @then IF DUP { } =/ @If solutions list not empty THEN @then Back substitution, and solution @TempSolVar=ArgumentOfTrigFun 1 DISP subs SUBST @substitute back the original subvar SOLVE @variable and solve. END subvar 'var' STO NOVAL 'subs' STO @restore subs, subvar to NOVAL NOVAL 'subvar' STO END flags STOF @Restore flags which could have >> @been changed through switch @to approximate mode Nick ==== :-) I was wondering whether THEM was a new symbol you had created ;-) ==== Yet another smaller bug! If the arguments of trigonometric functions are not all the same then the list of arguments os trigonometric functions must be droped, because otherwise it would appear in the list of sulutions. The code Must be changed to IF { 1 } @functions are + PILIST @the same THEN Type f(sinx,cosx,tanx) 1 DISP eqfact HALFTAN @apply HALFTAN EXPAND OVER HEAD 2 / @to the current EXPAND 'TempSolVar' 2 @equation factor=0 ->LIST |^MATCH DROP @replace each 'eqfact' STO @occurence 'TempSolVar' SWAP @of the argument HEAD 2 / EXPAND = @of trigonometric 'subs' STO var @functions with 'subvar' STO @TempSolVar, 'TempSolVar' 'var' @Store STO @TempSolVar=trigarg @for later substitution @Store variable to @solve for in subvar @Store TempSolVar @in var ELSE DROP @else drop list of arguments END ==== How can we thank u for the fantastic work that u do on the group! Congratualions and never endings thanks Vincent ==== Yes, like There is THEM (TI users) uns US (HP users) ? ;-) P.S. Only joking, don't take me serious. :-) ==== Dear colleagues & fans of the HP49/48 (the mother of all calculators), I have not been successful to make XCELL20.HP run in my HP49G... (This program code is available at www.HPcalc.org under XCELL48 for the HP49.) I have uploaded FNT1 (universal font) & placed it on port 1 of my HP49 memory doing a STO 0 operation... (per program instructions...) I have uploaded XCELL20.HP on the same identical way as above, and the calculator accepts & recognizes the code as a legit Library file... ...but the instructions read that after turning the HP49 off/on a new variable named XCELL should pop up in the LIB menu... (This XCELL variable should make the program run...) ...well in my HP49 this did not happen... ...instead I got a library file named 1220 which only places the XCELL20.HP code back to the calculator's stack... I must have overlooked something or missed a step... I need some guidance... TS There s a doc about what you want on www.HP-Network.com , 6c242c52.0201271911.4ecacdf5@posting.google.com... ==== RATS!!!270)(/Û#)(/)(/&%Û# It's ALL your fault, Wolfgang!!! I'm not even close to 10,000 points AND my 49G keyboard is now so soft that I get key bounce no matter what the value is that I feed to ->KEYTIME ==== can be found at http://education.ti.com/product/tech/30xa/faqs/faq83086.html. The description likely applies to HP products as well. Best , Tom Ferrio and could this, journal algorithm, :( I would show the time in ASM on the HP49G, HOW ? HELP , ==== I am making my way through Gilbert Urroz's 2 volume masterpiece on the 49G. It's very good, although very long. One thing that is a bit frustrating is that when I compare results on my calc with the ones listed in his book, the result IS the same, but displayed in a different way. I know the 49G allows many formats for display, which is great, but if you want to give someone an example result, it would be convenient to say make sure your mode value is xxxx, where xxxx, say, is something like a bitmask of ALL modes: FLAGS, CAS and DISPlay. This would allow completely reproduceable results, provided you have the same ROM, which *can* be checked using VER. Anybody have an answer? Mark. ==== Use CASCFG first If of no help then you have to downgrade to ROM 1.16 which was in the marked at the time Urroz 49G. displayed anyway to do it? ==== I was experimenting in Debug2 something about the new 49 CHOOSE engine. As reported in Carsten Dominik's tutorial (Using HP49 CHOOSE for more than just selection), this is done by the unsupported entry something like FPTR 2 72 but Debug2 compiler gave me an error in that line. I tried to change FPTR in FLASHPTR, but this won't solve the problem. So I will be grateful to the one that explains me about rompointers abd flashpointers calls in debug2 and MASD sintax (I guess they are different). /-----------------------/ Marco Tinarelli - Bologna - Italy tittiXXYYZZ_tt@tin.it (togliere XXYYZZ dall'indirizzo per rispondere) ==== I tried and it works: :: a b c 3 tyt 0 ' DROPFALSE FLASHPTR 002 072 ; Piotr Kowalewski. ==== Well, FLASHPTR 2 72 workes fine for me using debug2. Greetings AM ==== I want a smoth interpolating curv of values in the X-axes relatet to values in the Y-axes. The nearest funktion I found is the scatter funktion. But I want a curv instead of dots. Kent Nick Karagiaouroglou skrev: ==== After SCATRPLOT press |STATL| which draws the least square model that you have chosen Note: The Plot changes to Function now! I hope this is what you wanted... values in the Y-axes. instead of dots. curve? ==== How can you solve an Initial Value Problem eg.: dy/dx=exp(-3*x) where y(0)=5/3. I'm in RPN but can't find the syntax. Also, how about a symbolic solution of an equation with a domain restriction, eg. Integral(1/(1-t)) ; (t>1). David z. ==== A) You have to give them in a matrix [ d1Y(X)=EXP(.3*X) Y(0)=5/3] Y(X) DESOLVE B) 't' STOVX 't>1' ASSUME but I don't know if REALASSUME list applies to Integrals ==== Hey. I've been using these two libs for a long time, tried the 3.7b4 version of java too. Great programs. My question is ... does it appear to you that java is significantly slower at displaying algebraics (in pretty print, of course) than EQstk is? something is screwey here... 'cause EQstk's AGROB needs .226s to process 'Integrate(a,b,Sqrt(c*x^n-1/d),x)' (small font, just a random expression) while java's AGROB command (which is the same, right?!) takes .43s, Twice as EQstk! And java 36a takes .45s :( with large font it is more or less the same situation. Accidentaly i found out that java's AGROB on a *grob* takes .29s! longer than EQSTK processes the *algebraic*!! why does this happen? ...if only java was as fast... now i have to switch back and forth... :(( ==== moin, does someone know, where to buy a leather pouch/case for a HP 49G Calculator? I'm thinking of a pouch like the original accessoire for the HP 48... Anyone know where to buy it in Europe (or better Germany)?? Or does anyone sell his own? I tried at ebay and many shops, but didn't found anything... Matthias ==== Don't use it! It's based on an earlier 1.19-6 and only the CAS is better but you'll miss for example pretty-print e^x JYA & BP have promised to make a united 1.19-7 later on - just don't hold your breath (I guess a good release date would be 1.4.2002 :) ROM?? X X Every calculator in the history will someday vanish the production stopped and the official support will disappear, so that cannot be a reason to not to buy a calc. AND there is no mention officially from the HP that the HP 49G (nor the 49GX ) is not produced anymore. ACO is dissolved, yes. But still JYA & BP (at least) from the original developers of the v'ger team continue to support the 49G. And what about the active user group?! his goodies, while totally pissed off by ACO's death. (not unlike myself and many others) The TI 92 has new members in the family the latest which of has a whopping 3MB of Flash. Still I would not hesitate to buy any of the TI models if they would fit the bill. Or yet better: buy a TI 89 AND a HP 49G and get the benefits of the both worlds!!! It is not dead until the Hex Lady Swings (or something...;) ==== PS: I'm IN for open source or even restricted open source X snipped a long ranting about 49G is not dead (yet) ==== I just checked the source, and there might be some conflicts, depending on how people are using the library. QVSOP (the command) currently outputs only planets 5-8, all at once. If a library is using *this* command rather than the planet-specific version, it will break, unless other changes are made. Certainly, the preferred command is QV_D.Y.n>LBR. A new command should be added to indicate which planets are supported by a given QVSOP library, in the order that they appear in the output of the QVSOP command.. Or, another option: set up a QVSOP library with a different romid, dedicated to the inner planets. This allows for some flexibility in memory usage, at the expense of using up another romid. So, Georg and Dennis -- what're your thoughts since you two know your own libraries? ---------- Keith J. Farmer kfarmer@thuban.org http://www.thuban.org ==== any way to solve quad, equacao do 2¡ grau(it has that name in brazil)? solver does invalid data. ==== I'm thoroughly enjoying the Stat49Pro program. I'm having a little trouble with the one-way ANOVA, though. I tried using the data that came with the program and my own data, but the HP-49 keeps giving me error messages like undefined result or invalid dimension. I suppose I'm entering the data into the calculator incorrectly. Any advice/suggestions/tips? ==== The Euro symbol is used by MASD to enter special characters.. I guess the decompiler doesn't recognize it ==== The minus sign in the argument of the KEYEVAL ensures that the original key functionality is assigned. Using a positive number (no minus sign) indicates that you'd prefer to use a possibly *assigned* key (in User Mode _only_ - naturally) to be assigned to the designated key location. ==== it to my 49G. Could anybody give me some pointers?? Cam ==== I hope you do not mean my STARTEQW programs?! The documentation to that is of current HP standard that is: try it out! The programs are supposed to reside in the Home level and in CASPAR plus my PARDIR AND especially in the STARTEQW which is run every time you press [LS] [MODE] e.g.. [CUSTOM] while in EQW then you can select from a CHOOSE list a sublist (in my STARTEQW case) and then the final selection of an operation that you would like to perform on the selected subexpression. Veli-Pekka Nousiainen install Obfuscate: to deliberately make something confusing and difficult to understand something tells me this thread is already dead, tho ==== George Tsiros suggested: That sounds great! Like C! Here are some thoughts: * 48 or 49? Or why not 38? * Supported entries only? (Supported by whom? :-) * How will the code be executed? I mean, under what circumstances? For example: :: NDROPFALSE ExitAtLOOP ; When that program is the config code of a library, it will work as a config aborter on any 48. Weird enough? * Why not a prize? C'mon, everybody has something at home that might be a nice prize. Things such as HP brochures in local languages, a spare issue of the HP Digest, or PPC/HPCC/HP Journal, some 67/41 cards, a poster, an original box, a cable... Any prize makes a challenge appealing! , Bye. HPCC member #1046 PS: > The *program* gets the credit. Not the author. We need a Who's Who Among HP Users! ==== yep. i gave you the rules. ditto. cute one the only rule is the file to compile . If that is not clear enough, it roughly means: anything goes. i do not want to make the contest appealing by introducing a prize. that, for me, is lame. you should do things not because you expect something in return. what are you? a whore? do you also expect to get paid when a friend comes to your house to stay for the night? you don't seem to understand the philosophy here. it doesn't matter at don't make me put a second rule which says only anonymous posters allowed. Oh wait, i got a better idea. from now on, i am not the judge anymore and there will be no winner. ==== Neat! BTW, Aaron, my HP48 took more than 6 iterations to get SQRT(9) to 12 decimal places using the half algorithm. I don't have a clue how the hardware does it. Dan ==== What did you use as your initial guess? I used 1. Maybe also we had different modes set and mine rounded to 3 sooner or something. Also, there is a distinct possibility that I simply can not count anymore ;-) -- Aaron ==== I have read somewhere that HP employs the CORDIC algorithm to calculate rational functions, sin, cos, tan, exp, ln, sqrt. It's close relative to algorithm whitch is used in floating point units of x86 processors. One old Dr. Dobbs journal explains the algorithm in details. Sad thing is that to use the algorithm, one has to have tables of arctan(2^-n), and one other function, at hand :( Best wishes, Robert Tiismus ==== I don't know anything about HP49 but on the HP48, flag 3 is the key. If that flag is OFF, the calculator doesn't evaluate functions until you ask it to. That way you can leave the square root sign and mess around with the expression to your heart's content, switching A*B to B*A and A*(B*C) to (A*B)*C, etc. Dan ==== Could someone tell me why the following code that converts a real to string crashes my calc BAD? I believe that the calc cannot find itself after returning to normal OS mode at the =GETPTRLOOP function. ASSEMBLE NIBASC HPHP49-C RPL :: % 90.3050 ASSEMBLE CON(5) =TURNMENUOFF CON(5) =DOCODE REL(5) >end GOSBVL =SAVPTR * Save RPL pointers GOSBVL =DisableIntr A=DAT1 A ?A#0 A * address equal to 0 - no ob on stack? GOYES + LC(5) #201 GOVLNG =GPErrjmpC * Exit & Error - Too Few Arguments + D1=A * Point to address of string A=DAT1 A * Read first 5 nibbles LC(5) =DOREAL * Load C with string prolog ?A=C A * Do we have a string? GOYES + LC(5) #202 GOVLNG =GPErrjmpC * Exit & Error - Bad Argument type + GOSUBL GetptrEvalC RPL :: ( Convert real to string ) DO>STR COLA ASSEMBLE CON(5) =DOCODE REL(5) >end GOSBVL =AllowIntr * Enable interrupts GOVLNG =GETPTRLOOP * Exit Program GetptrEvalC C=RSTK * Get address of current command before A=C A * dropping to RPL GOSBVL =GETPTR * Get RPL Pointers PC=(A) * Start RPL sub program RPL GARBAGE ; ==== In such a case, you'd better come to your exams with a slide rule! and the prof should have a supply of them for those who don't have a simple calculator, but just a cell phone, on loan during the exam :-) ________________________________________________________ Prof. Karl Kleine http://www.fh-jena.de/~kleine Fachhochschule Jena kleine@fh-jena.de Carl-Zeiss-Promenade 2 +49-3641-205-502 [fax -503] D-07745 Jena, Germany ==== I think that these restrictions will limit me pretty much to the 10Bii only (what, an ALGEBRAIC calc!!???!?!?!?) or that 6s thing. Of course, there are several TI models that one could use, but then they're TI's. Roger Arlington, Texas standard disclaimer: The University of Texas does not necessarily represent my views about anything. ==== I have a question to calculating complex numbers on the HP49G. How can I Enter a complex number as |Z| and phi ? I can only enter them with realpart and imaginarypart. Please help me ! thanks Christoph ==== You can set the mode of your 48 from rectangular to polar in the mode screen (or just type -16 SF) When entering the two numbers between the brackets put an angle sign (CHR 128) inbetween. This will work in either rectangular or polar mode -- This message was written with 100% recycled electrons Pivo ==== just type (Z [ALPHA] [rshift] [6] phi) ... ready Christoph Krall schrieb im Newsbeitrag realpart ==== Is comp.sources.hp48 dead ? ==== I think that you cannot access the original equation writer within the metakernel. Maybe with syseval or libeval. ==== You could type HALT to suspend the MK environment. Not very elegant, but the only way here... Raymond Jos216 Alberto Machado schrieb im ==== Just do #6B48Eh SYSEVAL BE CAREFULL TO PRESS THE CORRECT HEXADECIMAL VALUE TO AVOID TTRM's Paulo ==== I wonder how to get a line that combine the doots in a scatter-graph? If this ain«t possible, do you have any Idea of a good solutions. Kent ==== I have a HP 48 GX... Kent Kent Jonsson skrev: ==== Can it help you if you define a function by ''intervalos''? other solution could be create a programm Kent Jonsson escribi227 en el mensaje ==== I am looking for the hard plastic case that was made by Astech Industries and sold by EduCALC. I bought one years ago from EduCALC for my 42S and I am looking for another. The case fit all HP calculators that used the same body design as the 20S, 32S/SII, 42S, 17BII, & etc. Are the cases still made, if not, is there another manufacturer that has something similar? Any help would be appreciated. -- Richard Garner rgarner@vidnet.net ==== Check the various cases offered by Calcpro website www.calcpro.com. I have bought from them several times and have had good dealing with its manager, Paul Nelson. Frank Travis. ==== John E (www.hp41.org) http://library.hp41.org/HP41/LibView.cfm?Command=List&Class=HP41&CategoryID= 12 ==== I'm glad somebody is looking at the future for calculators now that HP has dropped the ball. I can understand the disillusionment but I'd be sorry to see you desert the HP48/49 development. I think you would be short changing yourself if you are interested in developing calculators further. I think it is this community who will be the ones looking at what you might develop. Your right that the HP49 could have been a stepping stone for HP but now it can be a stepping stone for you and others. I am really hoping that someone out there takes up the challenge to provide the next level of advanced calculator beyond the HP49. I'm being selfish here because if someone doesn't I won't have a replacement for my HP49. The HP49 is an amazing tool but even I can see room for improvement. the 71 out '49 team. take, dB, take look the ==== to I'll still frequent this group, and I'll maybe once in a while release some programs, if it's any fun. You won't get rid of me that easy :-) ==== TETRI is a tetris game for mathematically minded people. In contrast to Win-like tetris versions, not growing speed but growing complexity makes the game fascinating. The first person winning 10,000 points gets as award either 100 euro or a HP49 ! Each TETRI game can be repeated and hence completely reproduced. The 10,000 points winner must prove his success by delivering a record of the game (like a chess game record). This is a list as follows: { #germ #c1 #r1 #c2 #r2 ... } where #germ is a bint, representing the germ of the game, more precisely, of the random sequence used to make a random distribution of dices in the first n lines in level n and the random choice of a dropping figure. #ci is the column the current figure is put (i= 1,...10), #ri is the rotation index of the figure (i= 0,..,3). The figure itself need not be encoded since it is determined by the germ and the first 2i+1 list members, assuming the game starts in level 1. In the directory version of TETRI, #germ is stored in the hidden file 'G'; in the library version it is stored Precise score rules of TETIR are in TETRI.txt We made a statistical analysis of the game at the Math Dpmt which showed that 10,000 points can be won with a not too small probability, assuming that no obvious mistake is made in placing a figure :-) Have fun - Wolfgang PS. As to my knowlegde, nobody (including myself) ever got 10,000 points ... ==== 11.467 and I successfully completed level 11 twice man that level 12 is a *BITCH*!!! -- This message was written with 100% recycled electrons Pivo ==== I do not believe this until I see the game record. As you're an experienced programmer, it should not be too difficult to record automatically while repeating your winning strategy :-) - Wolfgang ==== Juan Jos216 escribi227: Well, only 1.19-5. Hpcalc does have 1.19-6: http://www.hpcalc.org/hp49/pc/rom/beta1196.zip You can find it Avenard's own site: http://www.epita.fr:8000/~avenar j/hp/49.html http://www.epita.fr:8000/~avenar j/hp/product/hp49119-6.zip together with accurate instructions and a what's new section. ==== I«m looking for a file called vger.flash. It should be a new 1.19-6 ROM?? Juan Jos216 escribi227: Well, only 1.19-5. Hpcalc does have 1.19-6: http://www.hpcalc.org/hp49/pc/rom/beta1196.zip You can find it Avenard's own site: http://www.epita.fr:8000/~avenar_j/hp/49.html http://www.epita.fr:8000/~avenar_j/hp/product/hp49119-6.zip together with accurate instructions and a what's new section. ==== take a look: http://groups.google.com/groups?hl=pl&selm=146e5c77.0110241526.38dc8882%40po sting.google.com Piotr Kowalewski ==== Nobuddy can help me ? 3c533b1a$0$16596$626a54ce@news.free.fr... ==== You bet he is! How are you at MC programming? Georg was soliciting help for updating portions of map2 from HP48 to HP49. Any interested in participating contact me please! Dennis ==== I'll see what I can do.. I'll have to locate the source for QVSOP and set up the work queue for running all this. (what? you think I'm going to do this without a database? foolish mortal!) portions machine code's not my game yet ---------- Keith J. Farmer kfarmer@thuban.org http://www.thuban.org ==== I am basically looking for a descriptive stats package that handles more than one variable at a time and that also handles weighted values for said variables. Like when given a list of classes (x) and their weighted (y) values I should be able to enter this into a matrix and then just get the values of Sigma (x) and Sigma (y) and get the weighted mean, median and std. dev. values. ect... I have found that the HP49g does have some of these functions built in under the stats menus, but a lot of them are missing. Maybe I just haven't figured how to get them. I downloaded a few programs from hpcalc.org, but most of them are for inferential stats. While playing with the TI89 and specifically the Flash App for stats all of the functions that I was looking for were there and also I found the interface for entering data much nicer than the 49. the can an to ==== Let eur denote the euro symbol = character 160 on the 49. The name 'eur' decompiles and recompiles correctly with the UsrRPL-compiler on ENTER, but not with MASD. It decompiles as it should to ID eur @ But this recompiles with the MASD compiler V5.20 in ROM 19-6 to the empty name (NIBBHEX 84E2000). Character 160 was an empty character on the 48, but if the 49-designers agreed to put eur at its place, it should perhaps rigorously be used in this setting in the whole OS. - Wolfgang ==== Bug 1: Because restoring 'subs' and 'subvar' should be done only if a substitution and solution of the substituted equation has taken place, the above part should be corrected to: THEN Back substitution, @TempSolVar=ArgumentOfTrigFun and solution 1 DISP subs SUBST subvar SOLVE @return the original var subvar 'var' STO @and solve. 0 'subs' STO 0 @restore var and 'subvar' STO @restore subs, subvar to 0 END flags STOF @Restore flags which could @be changed through switch @to num mode >> After correction 1 go for bug 2: A 0 was chosen as a value for 'subs' indicating that no substitution has to be done. It is better to take NOVAL instead of 0, because otherwise the test subs 0 SAME NOT will not work if the calc has switched to approximate mode. Corrections: Instead of make it: NOVAL NOVAL RCLF -> eq var feq subs subvar flags and instead of make it: IF subs NOVAL @If subs is not NOVAL SAME NOT THEN Back substitution, @TempSolVar=ArgumentOfTrigFun and solution 1 DISP subs SUBST subvar SOLVE @return the original var subvar 'var' STO @and solve. NOVAL 'subs' STO NOVAL @restore var and 'subvar' STO @restore subs, subvar to 0 END flags STOF @Restore flags which could @be changed through switch @to num mode >> That's all I found until now. If I find more, I'll post it immediately. ==== Another bug! The DOSUBS-Procedure for filtering out factors that don't depend on the variable to solve for, will crash if it encounters an algebraic that doesn't contain any names, like for example 1+SQRT(3) . The following part Should be changed to: Filtering factors 1 DISP feq 1 @Put feq and a 1 on the stack @DOSUBS procedure starts here << NSUB R->I 2 DISP @display current @factor count IF DUP TYPE 9 =/ @If current factor OVER TYPE 6 =/ AND @is not algebraic @and not name THEN DROP @then drop it ELSE @else IF LNAME DUP { } SAME @if it doesn't contain THEN @any name DROP2 @then drop it ELSE @else IF AXL var @if it doesn't contain POS NOT @the variable to solve for THEN @then drop it DROP @else ELSE 0 = @build equation factor=0 END END END >> @DOSUBS procedure ends here DOSUBS 'feq' @Store the resulting STO @list of factors in @local variable feq. Mirror, mirror! Who's the maker of the buggiest code around? ;-) ==== Another bug! (Let's hope the last one.) Since for equations of the type a*sin(x)^2+b*sin(x)*cos(x)+c*cos(x)^2+d=0 there is another factor filtering procedure in the program, it should also be corrected, in order to filter out factors that don't contain any name. An additional correction in this part of the program is not to convert the resulting equations from the form A*sin(x)+B*cos(X) to the form A*tan(x)+B=0, for two reasons: When the original equation is transformed to the form (a+d)*SIN(X)^2+(b+d)*COS(X)^2+c*SIN(X)*COS(X)=0 it is possible that the factorization of this form returns also factors like COS(X), which then build equations like COS(X)=0. When dividing through COS(X) in order to bu222ld an equation that only contains TAN(X), we then get 1=0, which can't be solved and we lose solutions. The factorization of the above returns equations of the form A*SIN(X)+B*COS(X)=0, which can already be solved from the CAS. So because of the danger to lose solutions and because the CAS can solve the resulting equations, we only need to pass these equations to the built-in SOLVE without any further convertions. The code: should thus be changed to: @Another DOSUBS starts here @to filter out factors that @do not contain the variable @to solve for. The procedure @also builds equations @of the form factor=0 @which can be solved by the CAS << NSUB R->I 2 DISP IF DUP TYPE 9 =/ @If current factor OVER TYPE 6 =/ AND @is not algebraic @and not name THEN DROP @then drop it ELSE @else IF LNAME DUP { } SAME @if it doesn't contain THEN @any name DROP2 @then drop it ELSE @else IF AXL var @if it doesn't contain POS NOT @the variable to solve for THEN @then drop it DROP @else ELSE 0 = @build equation factor=0 END END END >> @DOSUBS procedure ends here DOSUBS 'eqfact' STO Let's hope, correcting will end soon! ==== Poor boy! You don't know what you missed! So many beautiful colors on the display. Almost 256-colors mode, though for some reson they don't seem to move with the contents of the screen. ;-) You can see the rainbow effect in the HP-49G microcomputer by Joseph Horn in : http://www.holyjoe.net/rainbow.htm Note: I have a one Hp-49G serial # : ID9... with this effect, but this is not problem. If you see all my programs in the Rechlin web page www.hpcalc.org), type ... caporalini in the search windows. This programs perform calculus for Engineering. Best .[:Â) Miguel Angel CAPORALINI HERK - Neuqu216n - ARGENTINA **************************************************************************** *** ==== Looking to buy a low cost HP48sx. Must be in good working condition -Bob rxg31@yahoo.com As it turned out there was a problem with the 40G CAS documents available from HPs site in that it used an HP keyboard font that had not been saved with the document. This meant that many of the explanations did not read properly. I've now obtained the font and re-saved with it embedded. If anyone needs it, the revised documentation can be found on my site. -- Colin Croft Old mathematicians never die; they just lose some of their functions. ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== Chinese-manufactured version. Like others mentioned you can specify the Chinese version. It cost me $155 ($163 with shipping) and it comes with the pocket guide, user's guide, and advanced user's guide. Also included is the calc-to-calc link cable, with an adapter for 49-to-48 linking, and the all-important serial (DB-9) adapter for linking with a PC. The last thing included in the package is 3 AAA batteries to use in the calc (I think they're Energizers). I've never used a 49 before so I don't know how bad the original keyboards and displays were--but I have owned a 48S and 48G in the past so I can compare with them. The screen is very nice (I like the black instead of blue) but I think I remember the 48 having a more anti-glare screen. This one is covered by glossy plastic so you get lots of glare and reflections when viewing on an angle. It doesn't ruin the experience for me, I just look at it dead-on and pay no attention to the problem. There's no trace of any rainbow effect as reported in the c.s.hp48 faq. The keyboard is fine. The keys are softer than the 48's and the click is a little cushy, but it isn't bad at all. It's just different. I know when I've pressed a button because the click is there, and I haven't had any problems with unintended double-presses (bouncing). Nor have I noticed any abnormal delay between the time I press a key and the time the machine responds to the press. The only problem I have with the unit at this point is the cover locks in a little too tight. It actually caused me to send my brand new $155 49G on it's first trip flying across the room. My hands were pretty sweaty and when trying to open it, the calc suddenly broke free of the cover and got away from me and landed on the floor (about 2.5 ft. drop because I was sitting). I'm glad to say the thing still works fine and there's no damage or scratches to the case. It did land on carpet after-all, but it's thin burbur with no padding so it did take a decent hit. I'm hoping after I've broken it in the problem goes away. Overall I'd say this is a good calculator to buy. It comes with the PC-link cable (which used to run you an extra $30 I think), the new screen cover, the new keyboard, and a hard copy of the advanced user's guide (which used to be sold separately, but is always downloadable in pdf form from HP's website). I still wish I had picked one up last fall when they were $75 after rebate, but then I wouldn't have gotten the Chinese-manufactured model with all the improvements. One last thing I thought I'd mention, is I HIGHLY recommend ordering from Samson Cables. They're the only place I've found that will guarantee you get the Chinese model, if that's the one you specify. Also the price is good, and if you do a search for 49G on ebay you can see they're always running auctions for the calculators, so you may be able to snag one for cheaper. If there's anything else you'd like to know, please ask. Mike I bought mine threw Samson cables. they had some from china but ran ought. know when one came in about three weeks went buy and I got my new china you. please reply soon. so.... I did ==== depends what screen. If we want back the 49 default screen as in case of a STARTOFF program then SysDisplay redraws it perfectly as was pointed out to me by JHM (thanks John). Let me return to the quite different problem in AskQuestion. Look at the program :: RECLAMINDISP TURNMENUOFF Turn menu off? AskQuestion NOT?SEMI TURNMENUOFF %2 dowait ; It first clears the entire screen. Answering YES we wait 2 sec to see whether the menu is really cleared. It seemingly isn't ! In reality, it is. IMHO, AskQuestion updates ABUFF in a wrong way. It seemingly updates only the upper part of ABUFF, not the lower part where in the 49 usually a copy of HARDBUFF2 is hosted. Also the complete clearing of ABUFF with RECLAIMDISP didn't help. The 48 would not behave this way because its ABUFF size is too small :-) ==== SysDisplay will redisplay the screen correctly only if you're in the HOME screen. If STARTOFF is called while you're in the Filer or in the EQW, then that's no good, you will have to call the display handler of these ParOuterLoops Are you discussing about STARTOFF or AskQuestion ? AskQuestion calls ChooseSimple, the screens saved are HARDBUFF and HARDBUFF2 and it will be restored to HARDBUFF and HARDBUFF2 . TURNMENUOFF will not clear the menu, it will just change the menu height. So if you want the menu to appear cleared you will have to clear it manually. What you see after AskQuestion is how the screen had been saved originally. It appears correct. It looks like a menu but it's really part of ABUFF ==== Since in my STARTOFF-program only the Header is concerned, it runs now perfect with ?DispStatus instead of SysDisplay, also in the EQW and in the Filer. It doesn't matter loosing the XXXX FILE MANAGER XXXX line but getting the first line of the standard Header instead after the count-down starts. It's even more useful :-) With my STARTOFF-program, I solved all garbage collection problems (except in the Mineshunt game; but Raymond is on the way of porting his ambitious ML-written 48-Minesgame to the 49, with no garbage collection anymore :-) Here my slightly revised STARTOFF count-down program (with a 4 nibble HXS # FFFF in TOFF): :: xMEM EditLExists? caseDROP 10 ZERO_DO DROP 98 1 9 INDEX@#- #>$ $>grob XYGROBDISP %1 dowait GETTOUCH ITE ZEROISTOSTO TRUE LOOP IT TurnOff ?DisStatus ; -------------------------------------------------------- I know that. Just for this reason I made a RECLAIMDISP in the program under question :: RECLAIMDISP TURNMENUOFF Turn menu off? AskQuestion NOT?SEMI TURNMENUOFF %2 dowait ; Also a CLEARVDISP instead of RECLAIMDISP would do the job. RECLAIMDISP completely clears the ABUFF (including its lower part with the virtuel menu) *before* AskQuestion comes into action. IMHO, even without the second TURNMENUOFF, there should be no menu displayed in the program's wait-phase, virtuel or not. Wolfgang ==== Change your program to Bye, Detlef -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel ==== Hallo Detlef, kommst Du mit dem Leser der NG nicht mehr hinterher ? :) ==== No, my news server is lagging behind :-) -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel ==== What I would do is: SetDA1Bad SetDA2aBad SetDA2bBad AppMode? NOTcase SysDisplay AppDisplay@ EVAL ; That way it will work in every condition. On the HP49, the Header may not exist (if the user did 0 ->HEADER for example) or like in the Filer or EQW. so you have to mark for redraw the Header and the stack. This program will also check if you're inside a ParOuterLoop and will then call the specific display program ==== QUOTE doesn't seem to work nested sums like GS(n=0,2,GS(m=0,n,a(n+1)*b(n-m+1))) . Even using QUOTE in such cases doesn't prevent the problems to appear, when the symbolic expressions which use the summation index of the outer sum like b(n-m+1), are list or array elements of the variable b, referrenced by index. On the contrary b(n-m+1) is evaluated correctly even in such nested sums, if b is a user defined function that takes one argument and returns one number. That means, it is not only the fact of using expressions like b(n-m+1) that use the index of the outer list. It is also the fact that b(n-m+1) is meant as a list/array element and not as user defined function. Anyway, good to know that summations are also possible through programming, although much less comfortable. Sum(n=0,2,Greetings(n)) ==== The 49G is just full of surprises! I wonder why it would convert the ZINTs to reals. Oh well, we can always use FOR ... NEXT loops for such cases. But after thinking about it more, that doesn't explain (to me) why it failed in the manner that it did. If it returned 'B^(A(n))+B^(A(n))+B^(A(n))+B^(A(n))+B^(A(n))' or '5*B^(A(n))', then I could understand. But erroring out with Non algebraic in expression? Could it be that 'A(n)' is being treated as the UDF (User Defined Function) 'A' along with its argument 'n' instead of as an indexed variable? I decided to try the RPL form: << 1 5 'B^A(n)' GS >>. I wasn't surprised that I got the same error message, but I was surprised at what was on the stack: level 6: 1 level 5: 164 1h level 4: 5 level 3: 164 1h level 2: 'B^A(n)' level 1: 'n' where 164 represents the character used to indicate a BINT. I had been thinking that perhaps this whole matter of strange results when using indexed variables in summations might be just an unfortunate limitation, but when a UserRPL program (even a buggy one) leaves BINTs on the stack, I consider it a bug in the system. , James PS: The below is just filler material. because it had more included text than new text. Neither would get the impression that http://www.squeakycleannews.com/cgi-bin/ mangles any thing posted from there. Hmm....I wonder whether they count by lines or by character - maybe if I set my right margin lower and rewrap my text... So what the heck, I've added some more new text; maybe it'll go through this time. from beautiful snow-covered China Township, Michigan, U.S.A. James ==== Have you tried doing : GS(n=0,2,QUOTE(GS(m=0,n,QUOTE(a(n+1)*b(n-m+1))))) It should do just like what you want. As Ivan explained, by default the HP49G tries to evaluate the summation symbolically which can't be done when working with lists like this. If you use QUOTE, then it will be done numally. But it should be slower than on the HP48S/G ==== Oops! What I really tried was << n 1 5 'B^A(n)' GS >>. This is with the original 1.19-6 ROM. , James ==== What about if when using reals as boundaries for the summation, it would call the old HP48 engine? That way you could do more things .. (but it will be much slower than on the HP48) ==== Yes; you can use the capital greek Sigma for this. ==== New to the group and wanted to start with the FAQ. Will someone point me in the right direction? I'm not sure where to look or how often it is re-posted here. Gil ==== www.hpcalc.org It' _the_ direction Ricardo -- http://iespgirona.xtec.es/~rblasco/ Un222os, hermanos linuxeros 640 Kb de memoria son m207s que suficientes (IBM AT Designers - 1982) Bueno, espero que esto no aparezca en la versi227n final (Bill Gates en la presentaci227n de Windows Crash Debug '98) ==== I have my anunciators (the shift indicators) turned ON almost all of the time when the HP48 is on. Sometimes one goes off, but never in the life of this HP48 did both of them function properly... Other than that, the unit is functional. Any ideas will be appreciated. Anyone with the same problem? mu. ==== I was trying to reset the memory of my HP 49, so I pressed ON + F1 + F6, then released F1 and F6 and then released ON (just like the user guide says). I don't know what I have done wrong, but the calc started to make a lot of test. So, I'm stucked in the KBD2 test, because I don't know the right order of the keys. I've started with F1 to F6, APPS to TOOL, VAR to NXT. My problem is that I don't know what keys are the following, I've tried with the arrows keys, but the test says failed. I also tried with the HIST key, but again the test fail. Please, help me with this, what is the rigth order? Camilo ==== To clear the memory, you have to release F6 before you release F1. Hold down ON, press and release F3, and release ON to stop the test with a warmstart. I don't know the order for Kbd2 either, but as long as your keys are working, don't lose any sleep over it. , James ==== I tried your differential equation but the solution is: -(LN(|Y(X)+2|)-LN(|Y(X)-2|))/4=(1/3)*X^3+cC0 (Absolute values of the arguments of the logarithms.) When I use LNCOLLECT and then try to solve for Y(X), I get the error: Not reducible to a rational expression If I replace Y(X) by Y and try to solve for Y, I get the same error. This is because of the absolute values in the expression. Could it be that you accidentally stripped away the absolute values? If the absolute values weren't present then you would only need to use LNCOLLECT. Then you could solve for Y(X). ==== Jim, I think you're forgetting the printed magazine of HPCC, Datafile (www.hpcc.org/datafile.html), which is sent to members worldwide every other month. Current issues of Datafile have nothing to envy from old and dead club journals. Datafile has information often not found elsewhere. Yes, there is a whole world outside comp.sys.hp48, with people that many of you have never heard of, and who are very keen on RPL programming. Try to get a sample issue - you won't be dissapointed (dispite my contributions!). Refer to Special Offer at www.hpcc.org/membership.html , HPCC member #1046 - ==== guess, after about a decade of intense HP handheld activity, I've been somewhat out of the loop since the late '80s. So it's wonderful to see any and all good activity out there. Jim Horn ==== As Steve suggested this will. However if you have limited space try a single OR gate in a SOT23-5 package such as fairchild NC7S32 Tiny Logic. This also helps with any timing problems due to addtional tp in the circuit. Colin. ==== Colin: The main reason for me to sugest the HC00 was that he can have the option of inverting CE or not, because 256 usually have an active low CE,and also, it's easier to get. :-) Since the first Ram's used in the 48 series had delays up to 150 ns, the propagation delays on the arrangement i made are negligible. If you like tiny logic, i sugest you check out the new NC7SP58 series... ;-) Steve Sousa ==== When they actually have both varieties in stock that is. I wonder whether the dealers get a choice. Also note the premium price: US$139.95 for the Indonesian model, US$164.95 for the Chinese model. Also note that Chris Edmunds, the founder of Samson Cables, also sells on eBay as cedmunds, and if he's selling the Chinese model, he makes a point of saying so in the description. Sometimes he gets less than the Samson Cables price, sometimes he gets more. , James -- ==== the inventory, I have two more college bound children in the next few years, so I have kept my eye on Samson. I have noticed the Chinese model does go in and out of stock. Given what HP is up too, it will be interesting to see if the Chinese models remain available, or if HP is purging their stock of 49s. ==== Well Nick, if you're trying to confuse me, you're doing a pretty good job. Being rather simple-minded, I tend to stick with RPL programs and avoid algebraics, but the GS (uppercase sigma) function demands an algebraic. Plus, when I look at your example, I wonder what n in the algebraic represents: the local name 'n' from UDF (User Defined Function) 'a', the other local name 'n' from UDF 'b', or the index n of the inner summation. To help keep things straight in my head, I changed UDF 'a' to << -> c << c 2 ^ >> >> and UDF 'b' to << -> d << d 2 * >> >>. http://groups.google.com/groups?hl=en&lr=lang_en&th=ee35cba5abfe0cee&rnum=1 that the 49G by default attempts a symbolic solution. In the original problem, 'GS(m=0,2,GS(n=0,m,a(n+1)*b(m-n+1)))', it looks to me as if it's a matter of not having 'm' defined with a num value and evaluating the indexed variable at the same time. But maybe I'll add more to your Another new (old) bug? thread. , James ==== Can you visit www.hpcalc.org? I can't since yesterday... ==== It works fine for me, using either http://www.hpcalc.org/ or http://ca-on.hpcalc.org/. , James ==== James M. Prange escribi227 en el mensaje ==== invalid syntax. For the arrow, we are talking about Right-Shift 0 right? ==== You are not typing this , are you?? In RPN, just type 200. ->KEYTIME And yes, the arrow is RS 0. Important: NO SPACES BETWEEN THE ARROW AND KEYTIME! If you're still having problems just get the command from the catalog (CAT), it's there. Enter the CATalog, hit RS 0 and look for ->KEYTIME just a few rows down. Type with care and take care! Diego. ==== Can someone explain how to use the message array of the libs to make ==== I am taking an Engineering Electromagnetics class the spring at Point Park College. I was looking for some programs on the subject including Vector Algebra or Vector Calculus. Any information that you can give me 22:45:02 -0600) ==== Probably because of the special included headers: X-this-is-too-long-winded-to-read: X-caution-may-include-Other-Topics: [ON][SPC] [ 48 only :( ] ==== 8.14 HMS-> SIN -Al ==== Working in a geometrical problem, I have come across the following expression: SIN(X) - X * COS(X) = PI/2 I have solved it numally with the 49 (X = 1.90569572931), but I have not been able to solve it symbolically; does someone know if it is at all possible? Rafael Mill207n. ==== Equations containing both trigonometric and non-trigonometric variable quantities do not in general have exact solutions. Yours is one of those that do not have exact solutions. ==== you may wish to have a look at www.ebay.de where HP49G are offered at 99Û. http://listings.ebay.de/aw/plistings/list/category9972/index.html G237nter ==== I am wondering why i 22 + 'i' STO can't be replaced by 22 'i' STO+ in local variable working? Best , Tal ==== STO+ does work on local variables, e.g. the program that follows returns 29, as expected. What exactly do you think doesn't work? << 7 -> i << 22 'i' STO+ i >> >> -- David Haguenauer ==== Could you inform me how convert LIST to 1 column matrix or SDAT)? jan ==== In a 49, the command AXL will switch between list and vector format, or between list-of-lists and matrix format. To switch a row matrix ( 1 row by n columns) to a column matrix (n rows by 1 column), use TRN on the matrix, which transposes real matrices. To switch a vector to a column matrix use 1 COL-> on it, where -> is the right-arrow character (keyed in by alpha, right-shift, zero). ==== Ok, 1 COL-> solved my problem, thanks I do not read Pocked Guide very carefully. jan ==== 1 COL-> is solved my problem. I have read Pocked Guide but .... jan ==== I was asked whether updating Unitman affects an already existing UNITS customization. It doesn't ! The only reason for updating was that we Europeans irreversibly stepped into the euro age. Hence, the description on using the 49 euro symbol in the new units category MONEY should be as unambigious as possible. It should be said that probably only from ROM 19-6 on the euro symbol in treated by MASD correctly, like the $-symbol. All former bugs are fixed. The fact that the 49 has the euro symbol but the 48 hasn't is one but not the only reason for me that I lately concentrated on the 49. The possibilities the 49 offers with respect to individual customization, a 48-owner can only dream of. The only problem is that ACO didn't document definitly and in detail the superiority of the 49 when there was still time to do so, right after ROM 19-6 appeared :-) - Wolfgang ==== This project is a very good idea! I think you should continue it - more examples with a very good comments ;) Piotr ==== Below the modified turn-off count-down in STARTOFF if you do not want any disturbing turn-off not only in Edit mode but in none of the concentration demanding applications like EQW, MTRW, the FINANCE dialog box etc: :: xMEM AppMode? caseDROP 10 ZERO_DO DROP 98 1 9 INDEX@#- #>$ $>grob XYGROBDISP %1 dowait GETTOUCH ITE ZEROISTOSTO TRUE LOOP IT TurnOff ?DisStatus ; This also helps you keeping the memory clean in long sessions which unnoticably generate a lot of garbage. - Wolfgang ==== I played around with this problem and found some things that I think should be posted here. As said before the sum GS(n=0,2,GS(m=0,n,a(n+1)*b(n-m+1))) with {1 4 0} stored in a and {2 5 0} stored in b, returns '2.*b(m+1.)+(4.*b(m-1.+1.)+2.*b(m+1.))+(4.*b(m-1.+1.)+2.*b(m+1.))' when evaluated. But if a and b are defined as UDFs, for example: << -> n '-(7/2*n^2-27/2*n+9)' >> for a << -> n '-(4*n^2-15*n+9)' >> for b then the sum can be evaluated to a number. Now the (even more) strange things: 1) If you define a and b as above in exact mode and the evaluate the sum GS(n=0,2,GS(m=0,n,a(n+1)*b(n-m+1))) then you get (after a long time) the correct result. 2) If you define a and b as above and evaluate the sum GS(n=0,2,GS(m=0,n,a(n+1)*b(n-m+1))) in approx. mode, then you get: '14.*n^4.+-50.5*n^3.+34.*n^2.+20.*n+....' This is already strange enough, but it goes even stranger. 3)If you define a and b in approx mode, << -> n '-(3.5*n^2.-13.5*n+9.)' >> for a << -> n '-(4.*n^2.-15.*n+9.)' >> for b and evaluate the sum GS(n=0,2,GS(m=0,n,a(n+1)*b(n-m+1))) in exact mode, you get an error: >> Error: Num input, where >> is the program end delimiter. Now, I have seen DUP Error, + Error and the like, but what is a >> Error? 4)If you define a and b in approx mode, << -> n '-(3.5*n^2.-13.5*n+9.)' >> for a << -> n '-(4.*n^2.-15.*n+9.)' >> for b and evaluate the sum GS(n=0,2,GS(m=0,n,a(n+1)*b(n-m+1))) in approx. mode, you get: '14.*n^4.+-50.5*n^3.+34.*n^2.+20.*n+....' Apart from the fact that this behavior is a bit weird, we have at least one possible way to evaluate such sums. We need functions that take an index n as argument and return the same value as the nth. element of a list. For example let's take a look at the list a which was {1 4 0}. If we have a function Fa(n) such that Fa(1)=1 , Fa(2)=4 and Fa(3)=0, then we could replace a(n+1) through Fa(n+1) in the sum, and evaluate the sum in exact mode. We can always create such a function Fa(n) because the HP49G has the command LAGRANGE. The list a is {1 4 0} so we need the Lagrange interpolating polynom for the matrix: [[1 2 3] [1 4 0]] The following program uses this idea: << RCLF -> sum ell flag << 1 ell SIZE FOR I ell I GET DUP RCL -> el lst << 'n' DUP 1 lst SIZE 1 SEQ lst 2 ->LIST AXL LAGRANGE XQ RCLVX 'n' = SUBST 'F el S~N + DUP (n)' + OBJ-> ROT DEFINE sum ' el S~N + (&n)' + OBJ-> ROT (&n)' + OBJ-> 2 -LIST ^MATCH DROP 'sum' STO >> NEXT sum flags STOF >> It takes such a sum like: GS(n=0,2,GS(m=0,n,a(n+1)*b(n-m+1))) and a list of all functions that actually are meant as list elements (in this example {a b} ). It then finds LAGRANGE interpolating functions, that when feeded with n, they return the same values as the nth list element of the corresponding list. It DEFINEs these functions and then it replaces for example a(n+1) through Fa(n+1) in the sum, using ^MATCH (upwards MATCH). (Can I program anything without MATCH?) At the end it returns modified the sum. you can evaluate this sum, but it takes a looooong time. Of course it is not the fastest/smallest possible, but it shows that weird ideas can also give the results you need. ;-) Interpolated ==== Yes, exactly. And in addition, there seem to be real difficulties when packing do much functionality in one command. Perhaps a split would avoid such difficulties. You mean TRIGSIN, TRIGCOS? ==== Hmmmm...apparently the same guy that just sold the last Xpander on eBay has another..... http://cgi6.ebay.com/aw-cgi/eBayISAPI.dll?ViewBids&item=1325544799 Roger ==== Or it was another fake auction to drive prices higher;-) Raymond Roger Metcalf DDS schrieb im Newsbeitrag has ==== I bet it won't be the lasr we see there -- looks like HP doesn't care and there're actually quite a lot around. Info for buyers: the batteries of all Xpanders are dead (if you don't use it it'll deep-discharge them). Yes, you can replacce them (the battery cover is glued on w/ double side tape). And no, they're not standard AA rechargeables... Bye, Detlef -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel ==== Folks--- Is there any way that you know of to manipulate (creating and allocating is the easy part) data and data structures maintained within the body of a CODE object, from within the same object? E.G: Say, for instance, each time I EVAL a paticular CODE object, I want the CODE object to keep track of the number of times it has been EVAL'd, using a 5-nybble number, rolling over upon overflow. How would I do this? Henry C. Gernhardt, III Remove _NOSPAM to reply direct ==== Hmm. CODE ... Counter CON(5) #0 * this produces five zero nibs D1=C C=DAT1 A C=C+1 A * increment DAT1=C A * write back ... ENDCODE I'd rather use a variable, maybe with a time stamp, in the hidden directory. IMHO self-modifying code on the 48 series is a mess;-) I remember that very nice big 3D pacman game, which modified itself. Because it didn't take care of it's own checksum, you had to hard reset your calc after each game... Of course it couldn't run from ROM (covered or not), because of the self-modifying 'feature'. , Raymond Henry Gernhardt/Andrea Winship schrieb im ==== I am developing a program that using multiple menus however rather than make a 131*7 GROB for each menu to display, I would like to create a GROB that contains a series of 21*7 menu lables one after the other. The problem is that the menus are not 4 bit (1 nibble) aligned on the screen. Does anyone know a way to draw the menus on the screen at pixel alignment rather that nibble without lots of programming? ==== X Elementary school went buy with 8 - 3 ---- 5 Later on we used slide rules, and then I had my HP 21: [ 8 ] [ENTER] [ 3 ] [ - ] 5 And everything seemed so natural. PS: I must admit that I have used Sharp BASIC programmable while in university together with HP 41 the single line equation editing was great AND that associates with the HP-71B and Calc Mode can anyone replicate that behavior with 49G? 01:57:09 -0600) ==== Why, you could no more remove Algebraic mode from the 49G than you could remove Internet Explorer from Windows; hey, mind your own business and stop trying to prevent HP from innovating ;-) --------------------------------------------------- I have a dream (and it's not about calculators :) ==== I can't find my manual and I wante to do a full reset on my 49g. I don't want to lose the updates, I just want to clear out all the programs I put into the different memories. ==== I'm using an HP48GX. I have a matrix on lvl 1 of the stack. I enter Solve linear system. I press edit. I press cancel. The matrix appears as if I had entered it after edit. It also dissapears from the stack. It's kinda useful. miz boundary=------------AE74378BA7ADCD702B2BF9A4 ==== --------------------------------------------------------------------- x-mac-creator=4D4F5353 I just got my hp49g, so I don't know how to use it to do this yet. But come on, you can solve this one without your calculator. It's just messy. Before noticing that this could be factored, I solved it as a cubic, but it's being factorizable makes the whole thing a lot more tractable. We can factor your equation like this: az^3 - 3z^2 + 3 - a = (z-1)(a z^2 + (a-3) z + a-3)) = 0 So we know that the solutions are: z = 1 and whatever the solution is to (az^2 + (a-3)z + a-3) = 0,. which we can get from the quadratic equation: z = (1 / (2a)) (3-a +/- sqrt[(a-3)^2 - 4*a*(a-3)]) Simplifying: z = (1/ (2a)) ( 3-a +/- sqrt(-3a^2 + 6 a + 9)) So there ya go. Pick a value for a and stick in it there. Let's say a=1: z = (1/2) (2 +/- sqrt(-3+6+9)) = 2.73205, -0.73205 Let's check these results: 1*(2.73205)^3 - 3*(2.73205)^2 + 3 - 1 = 0. The other root works too. Let's try a=2: z = (1/4) (1 +/- sqrt(-12+12+9)) = (1/4) (1+/- 3) = 1, -1/2 2*1^3 - 3*1^2 + 3 - 2 = 2 - 3 + 3 - 2 = 0 2*(-1/8) - 3*(1/4) + 3 - 2 = -1/4 - 3/4 + 1 = 0 So there's your solutions. See? You don't need a calculator for everything. -Doug Magnoli ==== I assume you mean a*z^3-3*z^2+3-a=0 ? My HP49G can solve this for z, giving: { z=1 z=-(a-3+SQRT(3*(a^2-2*a-3)))/(2*a) z=-(a-3-SQRT(3)*SQRT(-(a^2-2*a-3)))/(2*a) } ==== When that message appears in a box all over the screen, i should insert new batteries into my hp49. But what if I don't have them and I still have to do some calculations? (If Im on exam or similar) Is it possible to get rid of that message textbox? K. ==== Of course! Press F6 (Soft key for [OK]). ==== Just bought a 49G from Cynox, which had the better display, ROM 1.18, and yes, was made in China Impecable service. I. ==== We are at the seventh part of our trigo marathon already, if I didnÎt make any mistakes with counting. In this seventh part weÎll take a look at the inverse functions of SIN, COS and TAN. WeÎll also take a look at TrabakoulasÎ time travel. So we start by trying to stand on our hands to understand this inverse part easily. ;-) You of course know that trigonometric functions are periodic. They behave like doing after a while what they already have done. For example the SIN function keeps repeating itself, as it oscillates between 1 and -1. | * * * * | * o * o |* * * * --------+-------------------+---------------- | | * * | * * | | * * | * * | | * * | * * | | +-------2*Pi--------+ The distance between two x-coordinates which SIN(X) sends to the same y-coordinate is the period of the function. (Actually this is not quite correct, but it suffices for now.) For SIN(X) the period is 2*Pi. If we express this algebraically, then we have: SIN(X+2*Pi)=sin(X) You can do this on your HP49G: Enter SIN(X+2*Pi) , then press TEXPAND and EXPAND. The result is SIN(X). Or enter SIN(X+4*Pi) and press TEXPAND and EXPAND. Same result! The period of COS(X) is also 2*Pi. That means, if you have some quantity and take its COS, then adding 2*Pi to this quantity and taking the COS, returns the same number. The period of TAN is Pi. Now, why is all this important? Imagine that you have for example SIN(X) and you want the quantity X. There are more than one X which have the known sine. In fact there is an infinite number of such X. This was what Trabakoulas found difficult to understand. The aliens told him all about periods and the like and he applied this to time coordinate. He found that if he starts at t0 assuming a SIN(w(t0+t)) function with frequency w for his picture of the world, then when t=2*n1*Pi/w has passed by, he actually has reached the point where he started, though he never went backwards in time. But this contradicts his observation of steadily getting older. ;-) The functions that return an angle when fed a value of a trigonometric function of that angle are the inverse trigonometric finctions. The inverse function to SIN(X) is on the HP49G ASIN(X). It finds an angle, or arc, which has a given sine. If you solve an equation like: SIN(X)=a for X, then the HP49G answers with: { X=-((2*n1-1)*Pi+ASIN(a) X=2*n1*Pi+ASIN(a) } Both solutions contain an arbitrary integer n1. The infinite number of values that this integer can have, show that there is also an infinite number of angles X, that have the given sine a. It would be nice if we could show that the found solutions really have the sine a. So press SIN while the solutions list is on stack level 1. Now you have: { SIN(X)=-SIN(((2*n1-1)*Pi+ASIN(a)) SIN(X)=SIN(2*n1*Pi+ASIN(a)) } Both right hand sides can be TEXPANDed, so press TEXPAND. Now press OBJ-> DROP explode the solutions list. On stack level 1 you have: SIN(X)=SQRT(-(a^2-1))SIN(2*n1*Pi)+a*COS(2*n1*Pi) HP49G *returns* solutions that contain some arbitrary integer, but afterwards it *doesnÎt know* what variables are assumed to be integers. If we had this, then the above formula would simplify to a because SIN(2*n1*Pi)=0 and COS(2*n1*Pi)=1 when n1 is integer. The same is true for the other solution. We can of course meke a program that replaces SIN(2*n1*Pi) with 0 and COS(2*n1*Pi) with 1, but having such a feature like INTEGERASSUME would be better. Anyway, the other inverse trigonometric functions are ACOS and ATAN. There are also commands that convert between them. These are: ACOS2S - convert ACOS(X) to Pi/2-ASIN(X) ASIN2C - convert ASIN(X) to Pi/2-ACOS(X) ASIN2T - convert ASIN(X) to ATAN(X/SQRT(1-X^2)) ATAN2S - convert ATAN(X) to ASIN(X/SQRT(1-X^2)) For example, if you have ATAN(X) and want to convert to ACOS, then you press ATAN2S and then ASIN2C. The result is: Pi/2-ACOS(X/SQRT(X^2+1)) The HP49G has some automatic simplifications when it deals with inverse trigonometric functions. For example, enter ASIN(X) and then press COS. The resulte is SQRT(-(X^2-1)). Try also other combinations. Enter an inverse trigonometric function (ASIN, ACOS ATAN) and then press the key of a trigonometric function (SIN COS TAN) in any combination you like. Look how the HP49G gets rid of the, returning expressions with no trigonometric functions at all. LetÎs do some examples: 1) Show that ASIN(X)+ACOS(X) is a constant. Press ACOS2S and then EXPAND, to get Pi/2. 2) Show that SIN(ACOS(X))*COS(ASIN(X))=1-X^2 Enter the left hand side and press TEXPAND, EXPAND. Or enter the left hand side and press EVAL. 3) Show that SIN(ATAN(X)+ATAN(1/X)) can be used as a kind of definition of the function SIGN. Enter the expression and press TEXPAND. Then press EXPAND. The result is |X|/X which can be thought as a definition for SIGN. (What happens at X=0?) 4) Given an angle a, find all angles that have a sine equal to -SIN(a). The equation SIN(X)=-SIN(a) must be solved for X. So enter the equation, enter X and then SOLVE Example (4) shows where we are going to go in the next part. Yes, you guessed right: Trigonometric equations and their solutions! So if you like, send me any trigonometric equation that you find hard/impossible to solve with the HP49G, and IÎll try to solve them and take them in the next part. P.S. You can re-inverse your position and stand on your feet now. ;-) ==== If I want to solve a linear system (AáX=B) where B is all zeros, how do I find the solution. The calculator is only giving me the trivial solution (x=0). miz ==== Mithrandir schrieb: As far as I can see there *is* only this solution (given that A is a nonzero constant and X is the variable). Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ==== Try to say HCompaqP twice! No need for hurry in that one. :>) K. ==== iIm wondering, if the hp49G can be home made overclocked, like the Ti's, as i found in web pages, im pretty hapy with my calc, but sometimes, specially in def & lim, it takes a lot of time to get an answer, i already contact cynox but they dont sell the double speed module for self installation,... . ==== The HP49's Saturn processor is already overclocked (2X I think) ==== I think that you can run it at 6MHz (wow my PC is a 1.4GHz) Fielding <@accesscable.net> a 216crit dans le message de 3c4dd448.16403145@10.0.0.1... ==== :>) Not ALL Europeans. (I live in Croatia, Europe) ==== Also Dolar is short for Aman Euro. :>) Greetings! K. European ==== Yes, you are all riqht. Sorry if I offended you. ==== no matter whether we are living in Europe or not - no matter whether we are living in a country that have the EUR or not...: We all are newbies with the EUR as we have only had the coins and bank notes for the last month...:-) , Martin ==== <@hp-network.com> schrieb im Newsbeitrag ==== i've installed that program, but i only can transfer files to hp. i used to do it with hptalx, which is better? but what i'm looking for is a program where i can edit my user-rpl programs i can view them in any editor but i can't use symbols :( , or i don't know how to use them! thanks for your help! vicent ==== Yes! An Algebraic Expression Evaluator in any CPU uses a kind of ALG->RPN (Reverse Polish Notation) converter. The operands and operators are stacked... In Computer Science you have to program such a beast! (at least here in Finland in 80's & 90's in some universities) ==== NK = Not Kompatible (K instead of C, another incompatibility ;-) ) ==== remains that algebraics (at least on my HP49G) seem to evaluate slower than programs that do the same. I'm not complaining here3, ==== first of all thanks for the detailed explanations! Yep, I should. ( But inner workings are not as interesting for me, as for example what I see when I try something the one or the other way. :-( ) The execution time difference can be also greater. You are right, in most of the cases I'll not get grayer than I am already, but there are also cases where you want the calc to be as speedy as possible. ==== Megaleeee! :-D Quite the same with what I observe. And John has given a very detailed explanation why it is that way. ==== Nick: Only if you're not German! :) Henry C. Gernhardt, III Remove _NOSPAM to reply ==== Dann aber richtig: NK = Nicht Kompatibel ;-) (Incompatible) ==== Im learning to program some minor things for a 39g mostly because there are relatively few games or programs compared to the 48 and 49g calcs. But is there any way to convert all or some of the other games to 39g format. I have seen a 38 to 39g converter on Colin or Jean's site I think(not sure), but never have seen anything about 48/49. If it isn't possible, what makes it not possible. Different languages? -Mike ==== Totally different OS and Languages except for the 40G & 49G CAS ==== Does anyone know where I can get a list of all the shortcut key strokes for the 49G. Such as: Press and hold right shift and then ENTER to switch between approx. ands exact modes. ... Ed ==== just have a look at: http://www.hpcalc.org/hp49/docs/faq/#ss4.14 :-) Ed Brawley schrieb im Newsbeitrag for ==== If anyone is interested I've created an extensive help file for HPGComm, the HP38/39/40G version of the communications software. It's a bit rough around the edges in places but it should certainly help anyone just starting off with the software. Unfortunately it's not linked into the program - you have to run it separately rather than via the Help menu. That may change. You can find it at http://members.iinet.net.au/~ccroft/zipfiles/HPGCmHlp.zip -- Colin Croft Old mathematicians never die; they just lose some of their functions. ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ================================== First of all, I have an HP-49G from China with ROM 1.9-6 (the new one released on Jan. 11th I think). I was wondering if anyone has a program that would let me display a grayscale splash every time I do a warmstart. I read the previous thread on this (about four days ago), and I tried the program that was suggested there, but it displays the grob 4 pixels to the right of the edge and it cuts off 4 pixels from the left. It also does not display the grob in grayscale, it just displays the top half for ~.5 seconds and then switches to the bottom half. This could be just a problem with me, but I checked the program three times to be sure I typed it in correctly, and I did. Dylan S. ==== My HP49G behaves the same way :-( boundary=----=_NextPart_000_0022_01C1A7F6.C888DA90 ==== --------------------------------------------------------------------- I use the UFL 1 minifont as I did with my HP 48. Well, how could you display a minifont map when wrinting anything in the 49 and 'echo' the characteres in the text you are writing? When you press CHARS in the 49, the system big font map appears, and of couse it is not the same chars correspondence. Is there any program to do this? Alberto de la Vega ==== Ever get the feeling that the physical universe is unfortunately much less interesting than we wish it could be? -- ____________________________________ Christopher R. Carlen Principal Laser/Optical Technologist Sandia National Laboratories CA USA crcarle@sandia.gov ==== being a student... Jim Horn (not overclocked - just half fast) ==== No! It was not a black hole. You are wrong. Remember the hum that you can hear when you listen very closely to your 48? These are bugs (nanotechnology) that do repairs and checking. Read another c.s.hp48 about this issue. Now, remember the operating temperature range of the 48. As you can see, liquid helium is of MUCH lower temperature than what is allowed. When the previously mentioned bugs take note of the extreme conditions, they immediately initiate SNRBLFD(*) to warm the calc up. Sadly, this was not thoroughly tested technology (beta) and is used only as a last resort measure. Side effects include a quantum leap thru time. Good job, you just gave Neanderthals a toy to play with. At less extreme temperatures, the bugs (that's why the calc can withstand relatively low temperatures.) use other, more conventional means of warming up, namely micro-fusion. I'd like to say more, but already i have said far more than what i am allowed to...Wait a little, the doorbell... Actually, it is far more interesting than we'd like. Ever heard of entanglement? *: I am not allowed to tell you what this acronym means ==== Christopher--- Universe''. Add that to my warped imagination, and all of a sudden the universe goes from being finite but without boundary and no holes to being finite, without boundary, and having one hole. Now try and figure THAT one out! Henry C. Gernhardt, III remove _NOSPAM to reply direct ==== Oh my! Please, not the BORG nanorobots! Not again! Uga uga! Can eat? No! Nick toy play! Press button! [Cracle!] Kaputt! Harharhar! Is that why sometimes the screen starts turning red? Uga uga! Tsiros kaputt! Harharharhar! It is so interesting that some parts of it are completely incompatible to every other part. ;-) Uga uga! Speak! Uga! ==== The as....e of the universe? ;-) ==== I called HP today to see if I could get an advanced users guide. I was told they quit printing them long ago and they're not available anymore. Is it true they come with the HP49Gs manufactured in China? I also thought I saw a post where someone said they got one by asking for the ROM upgrade kit. Anyone know anything about this? thanks ==== Maybe you already know this but you can download it in Adobe PDF format from the following site: http://www.hp.com/calculators/techsupport/graphing/49g_userguide/english_aug .html ==== I just received an HP49G it's the Chinese eddition, it has both manuals, the link cables and batteries included... ==== HP must have reacted quite quickly...Pity they don't react similarly to customers desires! Geoff ==== R. G. Newbury schrieb im Newsbeitrag http://www.hp.com/calculators/techsupport/graphing/49g_userguide/english_a ug Try to type ... /english_aug.html. It works here. I ordered the German manuals by HP service in Frankfurt. They'll send them for nothing to my adress. I hope I get them tomorrow. Micha. ==== advanced users guide. I thougt that i would never get it! But two weeks ago, i got the book (in german)! This don«t really helps you... Richard Brian Chamberlain schrieb im Newsbeitrag ==== I have a HP 49G Indonesian model with manual ver 1.0... 3 weeks ago, I had a phone call with HP Service center, and I asked if I can get an advanced users guide for my calc (I had the AUG from the Homepage, but wanted a printed one), and how much it would cost. They told me, i can get it for free if I give them my address... perhaps 4 or 5 days later, i got 2 books (AUG and a new manual ver.1.2)!! good service I thought... I hoped to get a PC link cable for free, too, like someone other posted in this NG. perhaps next time... I'm living in Germany, so i don't know if they do it the same way where you live. The books came from the Czech Republic, the new HP Headquarter??? so long... MH Brian Chamberlain schrieb im Newsbeitrag ==== Moin Matthias! Matthias Herntrich schrieb im Newsbeitrag Would you post the result in this NG? Micha. ==== For sale assorted HP programmables HP 28s ------ HP's first graphical calc, and also then the top-of-the-range programmable calculators. Complete with both manuals Used, but not abused? Best SENSIBLE offer (in pounds sterling) secures !!! Buyer collects from LONDON NW2 or ads postage/SAE Jiffy bag. HP 41CV & Stats R.O.M --------------------- In it's day, a top-of-the-range programmable calculators. Used, but not abused? Comes with non-HP case. and also:- Hewlett Packard HP-41 stats module, part number 000041-14001 Used, but not abused? Two HP 71b's, one with HP-IL...Open to offers on these HP 35 (a EARLY model), no battery or charger Best offer (in pounds sterling) secures !!! Buyer collects from LONDON NW2 or ads postage/SAE Jiffy bag. Stephen 07956-544202 ==== How about the old HP-71B with a Financial Pac and ofcourse some RAM expansion filled with your own BASIC programs to handle everything that you think that might be needed? calcs calcs we tall! wireless during large long will calculators ==== Hmmmm....I guess my school didn't consider that a very old HP model might be more better than a brand new Palm!!!! Roger ==== Sorry, but I just have to say that this discussion again shows the lack of understanding of tools in the real world. We use a very simple but effective rule here: You use your calculator (in 'single use mode', of course, just to answer the IR-question :-) but you can't have any form of handbook or instruction manual with it. Worksmen know their tools and master them, for others it's just a dump device for maybe simple additions and multiplications, but never the real thing. http://www.fh-jena.de/~kleine ==== How about using the Rest Pin and then erasing all of the user Flash as well as system and down-loading the new 1.19-6 system then instead of Q Q take the Batteries off and put them back on having only one battery (the middle one) the right way for one minute then put them back the right way and finally reset with [ON] [F3] IF this didn't make sense and/or JYA sayz that this will certainly harm your precious 49 I think you better wait until somebody else (who might think before answering) answers to your problem too! . ==== I have a questions in this type of form: CASE k 18 == i 90 < AND THEN 'i' 22 STO+ END k 11 == j 15 > AND THEN 'j' 15 STO- END k 16 == i 22 > j 32 < AND i 66 > j 46 == AND OR AND THEN 'i' 22 STO- END k 17 == j 17 < j 32 < i 44 > AND OR AND THEN 'j' 15 STO+ END END I tried to convert these questions so they will operate faster, in this way: { << i 90 < IF THEN 'i' 22 STO+ END >> << j 15 > IF THEN 'j' 15 STO- END >> << i 22 > j 32 < AND i 66 > j 46 == AND OR IF THEN 'i' 22 STO- END >> << j 17 < j 32 < i 44 > AND OR IF THEN 'j' 15 STO+ END >> } { 18 11 16 17 } k POS GET EVAL but it doesn't work. Why? Best , Tal ==== It works on my HP49G! Did you perhaps have stored the list with the small programs in a variable and use that variable instead of the list in your program? If you did so, then the local names i,j,k will not be evaluated to their values, when the command EVAL executes, because these variables were not part of the local variables structure of the program. Interesting way for branching though! :-) ==== It's time for sysrpl. If you want to see significant improvement in the execution time of your program, learn sysrpl. It will be much faster when you have used nulllams from i, j, and k. Check hpcalc.org for sysrpl documentation. -- Aaron ==== Exactly! And that's why I'm really happy with the HP49, having both possibilities. Algebraic objects *and* RPN. :-) (I'm not talking about ALG mode here, but about the algebraic objects themselves.) Yes, it takes a lot of key presses. Perhaps you should try cursor mode in the EQW? It can save a lot of clicking. ==== This flag was introduced by Parisse while I was doing the same kind of thing as you except that I had problems with Complex Numbers thus -120 SF really helped me. To my surprise it does not suppress your question and thus the problem remains in your case. I'm sure that Mr. Parisse can easily broader the functionality of this particular flag. ==== e: In my specific case, the problem has now been bypassed, for I finally realized that the best solution for the TRI1 triangle solver (which can't afford to let its necessary variables be deleted) is to hide the variables instead, via RENAME before and after, which ought in any case to be less total overhead than PURGE (before ISOL) followed by STO (to restore after ISOL). It was also necessary to rename VX as well, because in the case of commands like ISOL and SOLVE, the variable whose existence trips up the command is only that variable named on level-1, while the variable named in VX has absolutely nothing to do with the command at all, yet the CAS doggedly deals only with the name in VX, and never pays attention to the name on level-1, which in the case of these commands is exactly the opposite of what is really needed. I do not mean to be hard on the author(s) for how things have turned out, for I understand that they have been pushed to release this to the market in somewhat less time than ought to have been devoted to it, and there exists some minimum amount of time and development, in less than which no one could accomplish the very best possible outcome. Many of the engineers who currently use HP calculators and who frequent this newsgroup will have similar encounters in their ongoing careers with the conflict between opposing responsibilities, and I call attention to a previous post [link below], and further links therein to real-life major examples (starting where Citicorp Center is mentioned), as contributing to a useful perspective, and possibly offering something to remember, in case your projects, unlike mere calculators, whether you act in engineering or management capacities, are vital to the life and safety of many people, as is faced every day these days by our governments, whose decisions need to be based on the fullest level of universal human awareness, most especially on the depth of their own human consciousness. Citicorp, NASA, engineering, and life: http://groups.google.com/groups?selm=955lok%24bf7%241%40nnrp1.deja.com ----------------------------------------------------------- With best wishes from: John H Meyers http://www.mum.edu t.es... X ==== Could you please tell us what the error message was. And also if you have stored anything in variable x? When I do x^2 0 FOURIER in RPN mode, I get 8*Pi^3/3/2*Pi, which after EXPANDing goes to 4*Pi^2/3. ==== tes: Type 116 MENU Go to hpcalc.org and download eqlib, or similar in hp49 software section.. #1078 HPCC Member ==== I recently bought a new? ID935 HP49 and find that my favorite HP48 tools are not available. I used the left arrow, solve, root, solver function/menu all the time on my 48, it let me pull my values off the stack for my eq variables, also if I had more than 1 equation I could push the next equation button an not have to leave the problem I was working. Kind of a manual MES. Can I get to that kind of menu now? Speaking of the MES, I upgraded the rom to 1.18 which is supposed to have the MES in it but is there a menu that supports it? How do I get to it? And finally, although I am a long time Hp user, I started with a 28c, I am not much of a HP programer. How do I get the HP48 eqlib and its menu on my HP49? ==== . Welcome to the club andd don't worry, most tools of the HP48 are still available. :-) On my HP49G w222th ROM 1.19-6 you can get the equation solver menu with 30.01 MENU. Same functionality as the HP48. Perhaps you should assign << 30.01 MENU >> to some key, so that you can get there more comfortable. Again, on my HP49G w222th ROM 1.19-6, I use 116.01 MENU to go to the MES-Menu. Sorry, no eqlib in the HP49G. But 113.01 MENU takes you to the old built-in libraries menu of the HP48. There you get EQLIB (which doesn't work), COLIB, MES, and UTILS which work. Hope this was a little help for you, ==== Like on the 48, you get MES with 113 Menu in the menu |EQLIB| |COLIB| |MES| |UTILS| || || which we get with leftshift EQLIB on the 48. But |EQLIB| sets the empty menu since the proper equation lib is not integrated in the 49 ROM. I could write a tool EQLIB to appear in the APPS choose box on the 49, setting just the above menu but with FULL functionality also of the EQLIB option. I didn't do this hithertoo because I believed that reintroducing the fine Equation Lib in a new ROM version would be only a question of time. Adding it to the ROM (for RPN mode only) would be much easier and more efficient than any tool :-) Since the proper equation lib is already ported (though not very efficient as far as could see) I only need advice of people who compared the various portings and tell me which is the best and fastest one (not necessarily the smallest one; I'll reduce the library's size ... - Wolfgang ==== HPs. I guess it's because I didn't know about the tools available (or that should be available) when I first started. You say most tools, what else besides the EQLIB is gone? Since you are a 1.19 user I'll ask you, or anyone who has a thought. What's to be gained with 1.19 vs. 1.18? calculator for it's lack/change of fuctionality. Now maybe I can get some work done using only one calculator. Well, at least until I need my trusty EQLIB. ==== I'm using eql49 ver 2.5 which seems to have all the HP48 equation library implemented fairly well. I just find the creation of a user equation library involved. One feature I'd like added to it is the abilty to run an initialization routine for the equation. It also only responds to the F6 as Ok and ignores the Enter key which is a minor annoyance. I've d/l the souce code but it's a bit beyond me at the moment to modify or extend it. ==== . Hmmm, let me think! I can't remember anything else that was on the HP48GX and is no more on the HP49G. There are some changes, like that there is no more RULES menu in the EQW, but you have the whole CAS which is available there, so I think that this doesn't count. Some important differences are that the HP49G has a current variable which can interfere with the CAS, if you have anything stored in this variable. But if you avoid to store values in that variable while you use the CAS, then everything works well. There are many twice as many flags and a huge lot more of functions/commands. Also many of the not programmable operations on the HP48 are now programmable commands. I think there are substitutes for the EQLIB at www.hpcalc.com . But I can't tell you anything ybout them, because I never used one. It is admitedly a much more complex system, but once you get used to its idiosyncracy, you can do things that were impossible on the HP48. The question of course is, what you need more: To be able to do everything or to be able to do exactly those things that you need. Generally speaking, things that you could do with the HP48, you can also do with the 49. It's only that sometimes the calc doesn't behave the way the HP48 behaved. ==== my old HP 27S has died on last friday! I want to buy a new old one. ==== try www.hpmuseum.org section ' . ==== ==== Harold, on which CD is the manual? You can get one disk for $15,00. But, I'm afraid that the manual that John Ensworth is looking for, is not yet included in the MoHPC CD set. Fortunately, John can download it from Warren Furlow's website (www.hp41.org) ==== ...like HP, for example? On HP's site, you'll find that you can buy from them, and they also list retailers that carry their wares. You might try HP's online store. You just missed being able to get an HP 49G at a sale price from HP. ==== Sometimes I use HALT CLEAR ==== I've seen this in HP website today: developer's kit (SDK) and show a prototype smart handheld appliance using HP Chai-LX software. In addition, HP experts will consulting services. perhaps calculators.. See the note complete at: http://www.hp.com/hpinfo/newsroom/press/29jan02a.htm ==== I had never heard of this difference before: -55 CF 'XXXXXXXX' 123 IFT -55 CF 'XXXXXXXX+1' 123 IFT Here we are causing an Undefined Name error, and flag -55 will return to the stack the arguments causing the error. The 48G seems to consider the error to occur during the evaluation of the name or algebraic, as if SWAP ->NUM SWAP were first being performed by the user; this does not return the original arguments to IFT, but rather returns the specific detailed arguments that failed in the middle of ->NUM. The 49G, however, returns the original arguments from IFT (and similarly returns an original argument if ->NUM fails, rather than 'XXXXXXXX' and 1 where + fails, for example). This probably has little effect on program operation, since in most cases programs abort upon such errors, and generally don't try to continue using the returned arguments after an IFT error (but it could puzzle a user during debugging :) There are programs which do try to continue after certain errors, re-using the returned stack arguments; the APLY program from the HP48G TEACH command is one such example, where if a final ->ARRY (trapped by IFERR) fails (due to symbolics in the final matrix), the program then tries to make a list of lists instead. If you have flag -55 set before using APLY, then the arguments will not be returned, however, and APLY will then really fail; this is an example of the fact that you should always attend to flag -55 whenever you use IFERR, as well as the other fact that programmers who almost exclusively write SysRPL/ML, even if they were the original Corvallis 48S/G team, might not have been as careful at programming in UserRPL :) Say, given the numbering of 49G versions, I wonder what had originally been envisioned for version 2.0 ? -- http://www.mum.edu http://www.tm.org ==== Sorry for asking this stupid question, but I simply do not know how to store a lib in my hp 49g. I would like to install the Stat49 Pro libary. Could someone please give very detailed instructions about installing it. (I just do not know what to do after uploading the libary to the homedir. I opened the filemanager and selected the lib by pressing enter. then i pressed F4 (RCL). AND THEN????) ==== ... Exactly! I wonder if there is a chance for both sides to come closer to each other, so that perhaps some day a calc will face the earth, which has the advantages of the two. Is this too much of a dream? ==== As long as no calculator exists that has the advantages of both and disadvantages of none, it is the only total solution. But it would be a little difficult to carry one calculator on each hand. How are we going to press buttons then? ;-) ==== . OK! The PDF that contains all parts is numbered of course. EVAL and SIMPLIFY won't make 1-X^2 out of -(X^2-1). If you use DISTRIB you can convert -(X^2-1) to -X^2+1, but that's all. To display this expression as 1-X^2 you must use flag -114 (ascending or descending powers in polynomials). You can temporarily set flag -114, then EXPAND, then reset the status of flag -114 to its previous state. Nice idea! I' ll do that :-) Actually next part of the trigo marathon is about solving trigonometric equations and next next part is about trigonometric functions in conjuction with complex exponentials. So when the trigonometry marathon is over (2 or 3 parts more), it will be a good point to start a complex marathon. Yupiyeeeeh, the marathon continues through the ages!!! :-) ==== That is already in my plans for the next marathonial postings. Seems that I can read minds? :-) . ==== I own a Ti89, and the step by step and Equation mode was to tempting so I ordered my HP a few days ago it arrived today, I was very happy to see the Link cable & chinese writting on the batteries :) the unit is much nicer to look at, I have already machined out the screen and added a foam insert to seal the contour (looks factory) as for the buttons, I dont mind a bit although there is a difference in pace between my Ti and the HP. I love RPN mode :) dont ask me why I just do I like it so much that I downloaded an RPN addon for my Ti89 and found it quite addictive although the crashes were somewhat pathetic -hehe all in all this calculator came well packaged, 2 large manuals (user and advanced) pocket guide, batteries & USB link cable with adapters. the screen is smaller but its actually clearer with the modified screen and smaller fonts than my Ti89. Tomorrow I am taking it down to the shoemaker for some measurements he already agreed to make me a leather pouch with a belt clip and velcro (screen flip) for 60.00 (well worth it :) ) You should try RPN by Lars Frederiksen: http://www.perez-franco.com/symbulator/download/rpn.html It is not known to have any crashes. Using it with an assembly kernel/shell might not work, but most people now prefer nostub programs anyway. ==== Wish you all possible fun with your calculator(s)! solution, that is, having both flag ships. I also think that this is much better than the eternal convert them battle. ;-) Please stay tuned and report observed differences in features/capabilities/behavior. . ==== Hallo where do i get the new 1.19.6, and what is changed from the first 1.19.6?? regreeds Martin ==== There is place in Florida USA that will re- build them. You might be able to get parts too. I do not know for sure. 1-800-535-5692: (Toll Free USA and Canada, I think) Their web page is: http://www.internationalcalculator.com/index.html Internationaal Calculator & Computer 2916 Corrine Dr. Orlando, FL 32803 ==== looks like an empty solution list, but maybe you could tell us A) VERSION of your ROM (should be 1.19-6) B) RCLF your flags, maybe CASCFG can reset them to a suitable value? C) Your problem including the steps you made in order to solve it. . ==== : It is by definition impossible to algorithmically generate random numbers. If you'll settle for pseudo-random numbers, there are many ways to do it. See _The Art of Computer Programming, Volume 2: Seminumal Algorithms_ by Donald Knuth. ==== Hmm, but if I recall it from a variable, I know which variable it was, so why should I search? (Nick easy minded doen't seem to get it ;-) ) Well, got this, but the question remains, if I recall a variable, then I know which variable it was, so what should I search? ==== For those who are interested in such things... Using the methods that lead to the QVSOP data for Zotti's Urania library, I've managed to make a python extension which will allow expansion of those tables. Not just for the outer planets, but also for the inner ones, to within parts-per-million or better of the VSOP theory. While this is intended to be useful in making updated QVSOP libraries, I suppose it would have use elsewhere. http://www.thuban.org/projects -- see AstroMath http://www.thuban.org/astro -- web interface ---------- http://www.thuban.org ==== This is a remote control program. I believe that there is room for improvement. I would appreciate it if you could give me any hints to optimization regarding to increasing the rate. %%HP: T(3)A(D)F(.); << GROB 18 10 FFFF30FFFF30FFFF30FFFF30FFFF30FFFF30FFFF30FFFF30FFFF30FFFF30 -> black << 89 1 0 RCLF -> i j k flags << 2400 BAUD rc48.gro PICT STO { # 0d # 0d } PVIEW { # 89d # 1d } PICT OVER black GXOR WHILE 1 REPEAT #41F65h SYSEVAL @ WaitForKey (for a very low power consumption, accept [ON] key also). #1950Bh SYSEVAL @ UNCOERCE2 (convert to real number). DROP 'k' STO CASE k 45 == THEN flags STOF 9600 BAUD HOME 2 MENU CLOSEIO DROP KILL END @ [ON] key. k 29 == THEN OFF END @ [<-] key. k 18 == k 16 == OR k 11 == OR k 17 == OR THEN @ arrow keys. 1 SF @ x position: 1; 23; 45; 67; 89; 111. @ y position: 1; 16; 31; 46. CASE k 18 == i 90 < AND THEN 'i' 22 STO+ END k 11 == j 15 > AND THEN 'j' 15 STO- END k 16 == i 22 > j 32 < AND i 66 > j 46 == AND OR AND THEN 'i' 22 STO- END k 17 == j 17 < j 32 < i 44 > AND OR AND THEN 'j' 15 STO+ END END i R->B j R->B 2 ->LIST PICT OVER black GXOR PICT ROT black GXOR END k 25 == IF THEN @ [ENTER] key for confirmation. j 46 == i 111 == AND IF THEN CLOSEIO OFF 89 'i' STO 1 'j' STO END @ off. j i R->C { (1,1) (1,23) (1,45) (1,67) (1,89) (1,111) (16,1) (16,23) (16,45) (16,67) (16,89) (16,111) (31,1) (31,23) (31,45) (31,67) (31,89) (31,111) (1000,1000 ) (46,45) (46,67) (46,89)} SWAP POS 6 + 'k' STO END END 1 FS?C NOT IF THEN @ keyboard table: AAABBBCCCDDDEEEFFFGGGHHHIIIJJJKKKLLL NNNOOOPPPQQQRRRSSSTTTUUUVVVWWWXXX YYYZZZ 777888999/// 444555666*** 111222333--- 000... +++ k 3 * DUP 2 + SUB XMIT DROP END END >> >> >> Best , Tal ==== i really don't want to dissapoint you, but for decent speed in those things you have to use assembly, or at least sysRPL with parts in assembly. you can learn those to by reading tutorials on www.hpcalc.org and the sasm.doc and rplman.doc Don't worry about losing mem , you can always use the emulator to testrun your programs ==== i've been trying to run emu48 under linux and i was able to do it! but there is a problem... :( when i load 48 rom all is ok! but when i try to load 49's rom all the screen becomes black and i can't see anything! i tried with contrast contrlos (on +-) but they don't work! can anyone help me? ==== this program can make this into an 'inline machine code', and store it in the variable 'SEND.OBJ' +---For example-(HEXIFY and CHT were posted by Alonzo Gariepy)---+ | | | 8F180508F72F401BD0FFF30D816C45E ;do not include newlines | | 0712074206B007C107110A4E53E77008 | | D505B1300660030215403182A6E5CF01 | | | | 'SEND.STR' STO ;chk [A27F] | | SEND.STR HEXIFY #2C96h CHT ;make 'inline' | | 'SEND.OBJ' STO | | 'SEND.STR' PURGE | | | +--------------------------------------------------------------------------- -+ I find the HEXIFY program and now i search the CHT program Ambroise. .. ==== 1- yahoo.com 2- search for HP28S CHT 3- Click on *first* match ==== If you're interested in 28 ML Programming, you can have a look to the book 'Voyage au centre de la HP28' (in frecnh, freely avalable on: http://ng-csh48-20020128.courbis.com/online-28.html A PDF version is also avalable for downloading ==== I've been out of the loop for quite a while; I used a 48gx in high school, about 5 years ago. It served its purposes well, and now that I'm about to take another math course, I'm wondering if I should upgrade to the 49? I'm concerned about not being able to directly download my 48gx programs to it, but intrigued at the new capabilities. I think this would be a great thing to put in the FAQ too. What are your opinions, any sticklers who swear by the 48? or is it worth giving up my expansion cards and programs to convert? ==== *almost* a 49... You must decide ==== If you are comfortable with your 48 don't spend the money unless you just gotta haveit. The 49 is a fine calc but if you're used to the old HP feel it's not there. I've been playing with my 49 for over a you who answered my postings) I sent my first one back because of the lousy key feel, it was a early ID model, they replaced it, with an even earlier ID model, it was no better. I still hate the key feel. I guess it just wasn't the same as my old reliable. I bought my 49 preloaded with software but if I could do it over I would have just bought the card for my 48gx. ==== I know this is childish, and it's not a real need but I'm windering if the HP can do this as the Ti does... is there a way to jump back and forth between Dec <-> Fractions? its simple enough to convert them but I was wondering is all, I saw the complexe vs simple mode output and the manual bypass option -> |enter| but not the other way ... can it be done? your help is appreciated ==== . I assume that you're using the HP49G... To go the other way use the command XQ i.e. type XQ and then validate with ENTER :-) [ ->NUM takes back] ==== Well, you can toggle approx./exact mode for decimal numbers/fractions by pressing the red-shift key, holding that key down and pressing ENTER. ==== with So, Typing XQ .5 |enter| would give me 1/2 ? ==== If you are in algebraic mode: XQ(.5) |enter| gives 1/2 XQ(3.1415926535) |enter| gives pi In RPN mode: ==== I thought that a call to STARTERR was incorporated into the 49G (SysErrorTrap > ErrorHandled?) to attempt to offer a means to make the 49G act more like the 48, no? Now, what seems interesting to me is that the exceptions (where STARTERR isn't called on the 49G), which happen both when certain applications (e.g. Stack Browser) display their own messages, and certain error numbers (e.g. #0) bypass the generation of any message strings, are the very same exceptions that occur on the 48 (where ERRJMP is either never called or exits abnormally, without making strings); this seems to me to at least rationalize why the exact same conditions are exceptions on the 49G, even though it doesn't satisfy you, because you want all the internally trapped or independent exceptions to also be changed and to go through STARTERR anyway, which is a bit of an extra demand upon nice guys like JYA, who don't even work for HP any more :) Well, on my 49G emulator, pressing CANCEL (during a running program) and performing 0. DOERR in a program both produce the same special Interrupted message box (which doesn't need any OK, because it goes away on any next keystroke, just like any status area error message), so I would say that these are still equivalent to each other (on the same calculator); in fact, when I trap general program errors using IFERR, when I am finished dealing with my stack cleanup, I conclude with ERRN DOERR, and this never changes the behavior of the error, even when it is caused by CANCEL, and my error trapping program needed no change between my 48G and my 49G. The fact that CANCEL now leads to ERRN being #13Fh and ERRM being Interrupted rather than #0 and is the only massive change that I see (yet both CANCEL and 0. DOERR still give the same results on the same calc :) Well, your postings have supplied plenty of documentation; perhaps they should be stapled into every new user manual :) Yeah, maybe that's why the all the special-case errors (which are the same ones that were special cases on the 48) still don't call STARTERR ;) What about #DFF, #13F, #13E, #123 ? Wow, once (only) I got two External's from #123h DOERR on 49G (but other times I get the same as on HP48: nothing at all). Who knows, maybe we've got another random bug to track down ;-) For this, ERRN is always #70000h on both 48 and 49G; in fact, #70000h DOERR will repeat whatever was the previous text message (which is also not in the manual ;) Well, some threads, when you keep tugging at them, just seem to keep unraveling indefinitely :) [r->] [OFF] ==== John, I'm not demanding anything. We only can appeal to the former ACO to document how things are really running on the 49, the more they never documented at least the differences to the 48 :-) It depends whether you run a UsrRPL or a SysRPL program. Since the program delimiters of a UsrRPL program execute CKOATTNABORT (which aborts after some other manipulation) we'll get with high probability the Interrupted message and the terrifying error sound when pressing CANCEL. But nothing happens if aborting with CANCEL the SysRPL program :: BEGIN NOP ATTN? UNTIL; It depends as is indicated by the above program. But I agree insofar as a SysRPL programmer is responsible for error trapping or other escaping means by himself. I think that your postings supplied much more information and reason for stapling them in a new user manual :-) Sorry, I may have been too fast in my conclusions regarding the behaviour of Do#Exit :-) ====: UserRPL looping/branching/structure words also invoke CKOATTNABORT, since you may not run into another program delimiter until the very end :) You are mis-using the word abort, because you never did abort [ABORT,ERROROUT,ERRJMP etc.] in that program! Your program merely *tests* for the key (*traps* the key), and then exits *normally*, generating NO error, just like << DO UNTIL KEY END >> (and then pressing any *other* key, to satisfy the key test). Insert an ABORT at the end of your program, and then it will of course produce both an error and your dreaded Interrupted! -- exactly like << DO UNTIL KEY END 0. DOERR >> Once again, the 48 and 49 are identical in this respect, and CANCEL -- when it is recognized to produce an actual ABORT, without which there has been no interruption and no error -- leads to exactly the same place as 0. DOERR That HP manual you were quoting was a UserRPL manual, after all, which of course you can accuse of lying about everything, e.g. it is a lie that programs encountered in the runstream are put on the stack, which illusion is perpetrated only by the UserRPL compiler generating an invisible xSILENT' for you, while in SysRPL every program object that is not quoted will immediately execute. I think you're just trying to be a contrarian, W.R. ;-) If I'm elected President, I promise you that we're going to have a new Attorney General! [Richard M. Nixon, in a meaningless campaign statement] [r->] [OFF] ==== I'm just trying to get it straight in my mind (in UserRPL terms, that is, remember that I'm a bit simple-minded) just when STARTERR will actually be executed. If I use DROP on an empty stack to set up a known last error, and then run the program: << 0 IFERR DOERR THEN ERRM ERRN END >> I end up with and # 0h on the stack. Similarly, if I instead run the program: << 0 1000 IFERR FOR n n NEXT THEN ERRM ERRN END >> and press CANCEL before it finishes, then the last two objects on the stack are and # 0h. shows that in general, errors trapped by IFERR do not invoke STARTERR. #13F DOERR does invoke STARTERR for me, even though error #13F caused by 0 DOERR or CANCEL doesn't. The other three don't cause STARTERR to run and also have a null string for an error message. I think I'll just try to avoid using #123h DOERR in any programs. ;-) Would the following be correct? With the exception of error #13Fh caused by 0 DOERR or CANCEL, errors which set a new last error with a non-null error message string invoke STARTERR unless trapped by an IFERR structure. Any other 'errors' do not invoke STARTERR. It seems to me that ACO did a good job on this, except for the documentation. But as we learned when we were toddlers, the job isn't finished until the paperwork's done. That does seem to happen when loose ends are left untied. , James ==== IMHO, an empty dispute on words. I could have said also But nothing happens if interrupting with CANCEL ... Where it is stated that to abort a program is to mean to make use of the commands ABORT, ERROROUT, ERRJMP ? I use the BEGIN .., ATTN? UNTIL loop in many of my programs, e.g. NOON, and the corresponding text file says CANCEL aborts. No Aman user of NOON ever complained about this word choice (there are several such users :-) In German, abort (abbrechen) and interrupt (unterbrechen) have more or less the same meaning, in English not? Unfortunately, unterbrechen has also one of the different meanings of the English suspend (temporarily), hence, in the problem we are talking about, a German instinctively prefers the word abbrechen which he translates in 99% of the cases as abort. ==== Addendum: Since TRI1 performs TYPE to check for real arguments, the prior post will not recognize integers for the lengths of the sides of the triangle, unless the 49G is in Approximate mode; you can amend OVER TYPE { 0 13 } SWAP to { 0 13 28 } PICK3 TYPE in the source text, to improve this (sorry I missed that!) About the speed: Well, everything added to the 48G version to make the 49G version is within that box drawn around the original ISOL EQ-> commands, so let's see whether those add any significant execution time. No doubt Emu48's authentic calc speed is not exactly right, but that's what I'm using here to get some idea of the timings; you can post follow-ups to note whether real calcs are comparable. Let's start with RENAME; if the name lengths are different, then this needs to shift a large area of memory, but when each name is just one character, it takes only .02 seconds per RENAME, so performing RENAME six times per triangle solution costs us a total of only 0.12 seconds (you could hardly notice!) Doing RCLVX and then STOVX (of 1-character names), three times per triangle, adds another 0.42 seconds to the total (it might be less if 'VX' were in CASDIR, but mine's in HOME). How about -120 SF RCFL STOF ? Add another 0.3 seconds total. Well -- what's left? How about ISOL? To get rid of extraneous dialog boxes, please prepare the 49G by first doing -120 SF 'X' DUP STOVX PURGE (so now 'VX' contains 'X' and there is no variable named 'X', and any dialog boxes should be rendered silent) On the HP48GX, I'll run 'X=5' 'X' ::ISOL DTAG TEVAL (using my own TEVAL); this gives me 0.09 seconds per ISOL (a rather simple ISOL, but let's just go with it for now). On the 49G, however, in Approximate mode, it's 1.35 seconds! (exact mode is slightly longer); well, that's 15 times as long! So, three ISOLs will cost us an absolute minimum of 4 seconds total, which all by itself just about doubles the execution time, as compared with the 48GX! Now let's change that equation to 'SQ(c)=SQ(a)+SQ(b)-a*b*COS(C)*2.' 'C' ::ISOL DTAG TEVAL (this is one of the equations in the actual triangle solver) and now we're up to 7.9 seconds for one ISOL on the 49G! (compare vs. 0.52 on the 48GX -- the 49G is still 15x as long!) Since we have to use ISOL three times per triangle, I think we now have the answer to Who knows where the time goes? http://www.informatik.uni-hamburg.de/~zierke/sandy.denny/songs/whoknowswhere thetimegoes.html If I replace ISOL with SOLVE, it takes only slightly longer than ISOL to get the same result. Well, then, if SOLVE, which is a new 49G-only CAS command, includes as a subset of its operation what ISOL already does, why couldn't ISOL (a command already existing on the first HP48) be left alone, to act as it did on the 48? No doubt some of the original symbolic handling of the 48 was stripped out to make way for the new CAS, but note that even the HP18C/19B/17B perform the equivalent of ISOL every time their own num solver is invoked (and only if ISOL fails does it resort to making guesses, etc., which is just like what we are doing here), so how big was it? The only 49G general triangle solver I see today at hpcalc.org is a port of QTS, which is probably now faster, whereas TRI1 emphasizes the automatic, idiot-proof user interface instead (and also shows that a 48G[X] can run rings around a 49G :) I may send a self-extracting compressed version of TRI1 to www.hpcalc.org some day, and perhaps even replace the seven equations of TRI1 with up to 21 new equations, pre-solved for each variable, enabling the hobbled ISOL to be skipped entirely (or would someone else rather do this for me?) [r->] [OFF] ==== Addendum: Since TRI1 performs TYPE to check for real arguments, the prior post will not recognize integers for the lengths of the sides of the triangle, unless the 49G is in Approximate mode; you can amend OVER TYPE { 0 13 } SWAP to { 0 13 28 } PICK3 TYPE in the source text, to improve this (sorry I missed that!) About the speed: Well, everything added to the 48G version to make the 49G version is within that box drawn around the original ISOL EQ-> commands, so let's see whether those add any significant execution time. No doubt Emu48's authentic calc speed is not exactly right, but that's what I'm using here to get some idea of the timings; you can post follow-ups to note whether real calcs are comparable. Let's start with RENAME; if the name lengths are different, then this needs to shift a large area of memory, but when each name is just one character, it takes only .02 seconds per RENAME, so performing RENAME six times per triangle solution costs us a total of only 0.12 seconds (you could hardly notice!) Doing RCLVX and then STOVX (of 1-character names), three times per triangle, adds another 0.42 seconds to the total (it might be less if 'VX' were in CASDIR, but mine's in HOME). How about -120 SF RCFL STOF ? Add another 0.15 seconds total. Well -- what's left? How about ISOL itself? To get rid of extraneous dialog boxes, please prepare the 49G by first doing -120 SF 'X' DUP STOVX PURGE (so now 'VX' contains 'X' and there is no variable named 'X', and any dialog boxes should be rendered silent) On the HP48GX, I'll run 'X=5' 'X' ::ISOL DTAG TEVAL (using my own TEVAL); this gives me 0.09 seconds per ISOL (a rather simple ISOL, but let's just go with it for now). On the 49G, however, in Approximate mode, it's 1.35 seconds! (exact mode is slightly longer); well, that's 15 times as long! So, three ISOLs will cost us an absolute minimum of 4 seconds total, which all by itself just about doubles the execution time, as compared with the 48GX! Now let's change that equation to 'SQ(c)=SQ(a)+SQ(b)-a*b*COS(C)*2.' 'C' ::ISOL DTAG TEVAL (this is one of the equations in the actual triangle solver) and now we're up to 7.9 seconds for just *one* ISOL on the 49G! (compare vs. 0.52 on the 48GX -- the 49G still takes 15x as long!) Since we have to use ISOL three times per triangle, I think we now have the answer to Who knows where the time goes? http://www.informatik.uni-hamburg.de/~zierke/sandy.denny/songs/whoknowswhere thetimegoes.html If I replace ISOL with SOLVE, it takes only slightly longer than ISOL to get the same result. Well, then, if SOLVE, which is a new 49G-only CAS command, includes as a subset of its operation the same thing that ISOL does, why couldn't ISOL (a command already existing on the very first HP48) be left alone, to act as it did on the 48? No doubt some of the original symbolic handling of the 48 was stripped out to make way for the new CAS, but note that even the HP18C/19B/17B perform the equivalent of ISOL every time their own num solver is invoked (and only if ISOL fails does it resort to making guesses, etc., which is just like what we are doing here), so how big was it? The only 49G general triangle solver I see today at hpcalc.org is a port of QTS, which is probably now faster, whereas TRI1 emphasizes the automatic, idiot-proof user interface instead (and also shows that a 48G[X] can run rings around a 49G :) I may send a self-extracting compressed version of TRI1 to www.hpcalc.org some day, and perhaps even replace the seven equations of TRI1 with up to 21 new equations, pre-solved for each variable, enabling the hobbled ISOL to be skipped entirely (or would someone else rather do this for me?) [r->] [OFF]