A64 Also, on the 49G there's the graphical toolbox. Here's the post: http://groups.google.com/groups?selm=860b4s%24bac%241%40web1.cup.hp.com Look for Sys Rpl Gxor somewhere near the middle. The entry is called Gxor. == GraphWriter 2.1 is now available on http://www.musikwissenschaft.uni-mainz.de/~ag/hp49/ This time I only improved the online help function, and fixed a silly typo http://www.musikwissenschaft.uni-mainz.de/~ag ==== Just one question. Why is the function zoom limited to three levels??? In fact, apparently in all of my graph, i cannot zoom in more then three times... Yes, GraphWriter won't allow PICT to grow beyond 400x400 pixels, that's a limitation I put in to prevent the display from eating up too much memory. Note that, unlike the builtin graphing functions, GraphWriter still shows the entire display range after zooming, so you can still scroll through the entire graph... With the default settings (1.0 x 1.0 rectangle, resolution 0.01 x 0.01 units/pixel) the display needs 100x100 pixels, after zooming two times with a zoom factor of 2 the display size already becomes 400x400, so a third zoom won't be possible. Of course you can view a restricted area at higher zoom levels. E.g., setting the range to 0. - .5 in both dimensions in the SETUP dialog would give you another zoom level. It is also possible to change the zoom factor value to something less than 2, that would give you more zooms in smaller steps, but that's probably not what you want. Finally, you can also go into the sources, increase the 400 bound in main.s (it's a DEFINE near the beginning of the file) and recompile the library. But be warned that to give you the default 1.0 x 1.0 range at zoom level three would already need 800x800 pixels or some 80 KB of precious main memory. Hmm, maybe a kind of clipping zoom function would be needed to automatically reduce the display range when zooming in? Opinions wanted... http://www.musikwissenschaft.uni-mainz.de/~ag ==== Ok, I put this on my TODO list. Note that GraphWriter does not attempt to be a general drawing software, it's just a graph editor, designed to facilitate playing around with graph algorithms on the 49G. I'd use something like Corel Draw for the purpose you mentioned. ==== The calculator has a built-in Gauss-Jordan elimination algorithm. If you're looking to solve linear systems in that fashion, I suggest you look into the commands REF, RREF and rref to reduce matrices into their row-reduced echelon form. For example, if you want to solve the following linear system: 4 X1 - 1 X2 + 7 X3 = 1 5 X1 + 2 X2 + 8 X3 = 4 2 X1 - 3 X2 + 3 X3 = 6 you would enter the following augmented matrix: | 4 -1 7 1 | | 5 2 8 4 | | 2 -3 3 6 | RREF ..and you would get | 1 0 0 87/7 | | 0 1 0 -11/14 | | 0 0 1 -99/14 | Therefore, X1 = 87/7 X2 = -11/14 X3 = -99/14 The calculator can do base changes easily without any program (RS 3), however, this little program, BACO, from Wolfgang is convenient to display all (binary, octal, dec, hex) base changes on a single screen. http://www.hpcalc.org/hp49/math/misc/baco.zip ==== The Algorithm is described a THE main - algorithm in algebra, the one to solve any system of equations, that's why I'm confused, I can't find it anywhere. Short Description: It is the one where you look for vector-elements un-equal zero in the given set and make it the 'pivot'. After that, by simple multiplication and addition you get one row of the system of equations to zero. this do as long as a element un-equal zero is in the set. from the end-sheme you can read the attributes of the set, for instance: Is one or more of the given vectors erasable to construct a 'basis' ... (if the coloums has ended to zeros and can't provide a 'pivot') Algorithm is available in 4 Types: Type 1 only shows the input - vectors and just tells if a vector is eraseable and looks like this: I II III IV I x x x x I x x x x I x x x x I x x x x I --------------- where the roman digits are to identify the vectors, the x's just represent the vector-elements. Left and down is a seperation which will be important for the othertypes of the algorithm. Type 2 also provides information about the linear - combinations. If a coloumn contains only zeros in the end-scheme you can read, by which combination of the other vectors this (or these) one/s can be constructed. the firs scheme looks like this: I II III IV I x x x x I x x x x I x x x x I x x x x I --------------- 1 0 0 0 I 0 1 0 0 I 0 0 1 0 I 0 0 0 1 I Type 3 can combine the given vectors with other, to show if some of them are included in the given system or other things which I hope to see in the right moment in the test ;o) I II III IV I A B C D x x x x I y y y y x x x x I y y y y x x x x I y y y y x x x x I y y y y ----------------------------- Type 4 combines 2 and 3 and shows us everything we need to know about a given system of equations I II III IV I A B C D x x x x I y y y y x x x x I y y y y x x x x I y y y y x x x x I y y y y ----------------------------- 1 0 0 0 I 0 0 0 0 0 1 0 0 I 0 0 0 0 0 0 1 0 I 0 0 0 0 0 0 0 1 I 0 0 0 0 0 0 0 0 I 1 1 1 1 - jens ==== After some hours of searching, I guess, the correct name is 'Gauss Elimination' and the was to do it in coloums is an 'invention' of our professor and it is normally done in rows. ==== Also known as Gauss-Jordan (row) reduction. It comes in two forms, (simple) reduction to echelon form, which must be followed by back substitution -- this is the method preferred by computer programs bacause it is quicker -- and reduction to reduced echelon form -- from which the solution can be read directly. ==== As far as I know, the entry points are different on the hp48 and hp49g, so you can't transfer the programs directly. However, if you have the source code you should be able to recompile them for the 49g.  ==== . I looked in HPcalc and it seems that the program you are talking about is LINALG by Andreas Daumann. I checked the programs that implement the Gauss Spalten Algorithmus and they don't seem to include any SysRPL code in them. That means that you can transfer it to your 49 (in ASCII mode) without messing with entry points or recompilations. I don't have a 49, so I can't tell you if it works, but hopefully it does. So my recommendation is. - go to www.hpcalc.org and look for LINALG (use the search field). The zip file includes the documentation (auf Deutsch nat237rlich) and separate files for each group of functions. - The file for Gauss Spalten Algorithmus is GSASYM. Download it to your calc using Hpcomm or whatever, but make sure you transfer it in ASCII mode. ==== I put the math.lib - which is labeld as String in the File Manager - into the stack, put in a '0' and press STO>, but all I get is a 'STO error: Bad Argument Type' ... the .lib in the Stack shows up like some compiled code, as far as I see it (how can I scroll to the left?) exept the Comments like copyright an programm-name. ==== The math.lib is not needed in the 49!!. Now I see your problem. Like I told you in the previous post the only directory that you need is the GSASYM. It's the only thing you have to upload to your calc. If executing the program it gives an error, then check if there is a command in 'ticks' left in the stack. Look for that command in the math.doc document and see what it does. Then write a small program with the same name as on the math.lib using the HP49 internal commands. For example, << + >> MADD STO will replace MADD math.lib command. If you need more help doing the conversion just tell me. I'm really sorry I don't have a 49 to help you more than that. ==== Note: use the exact punctuation shown, including the trailing semicolon This depends on where the program came from; for user programs created on or for the 49G, it is best to set the 49G to Exact mode, to compile the correct numeric object types (for those programs where it matters), since the presence or absence of a decimal point is significant to the 49G user program compiler, and the correct object type is significant to some 49G commands, particularly in the CAS. For user programs created on or for any HP48, it is best to set the 49G to Approximate mode, to compile all numbers as real type (for those programs where it matters), just as all numbers were on the HP48 series of calcs. In the case of the small program I posted in this thread, the results will be the same, regardless of which mode (Exact vs. Approximate) was set before downloading or typing, because none of the commands in this little program distinguishes between exact integers or real numbers. There's yet another advantage to posting user programs with all the exact ascii translation codes, which is that if you are using EMU48, then you can copy the program text and paste it right onto the emulator stack, without bothering to store it first into any file; similarly, you can copy text from the emulator stack directly back into any text file or post, with the aid of small ascii translator programs, which run in the emulated (or real) calc itself, as you can find within this post: http://groups.google.com/groups?selm=3AB195F4.6E36010D%40miu.edu ==== The calculator takes the name of the list from the header: x02 is the length of the name: 2 chars or bytes x4C is ASCII 'L' x31 is ASCII '1' So the name is always taken as 'L1'. ==== http://www.greatunpublished.com/bookstore/search.php3. I >have the both ==== ==== Try gnuplot. Octave (a free MATLAB clone) has an interface to it. For real data visualization with all bells and whistles, there's OpenDX from IBM. All these are available as freeware. ==== I just downloaded two sets of KML scripts from HPcalc.org. Cannot get any of them to run even though the ones I was using before work fine. I get the message below. I'm running Emu48 version 1.2, could that be the cause? Title: !best emuhp49 screen hp268x128pxl zoom 2 pc600*800pxls Author: Jaime Fernando Meza Meza Calculator Model : X Cannot open Rom hp49g.rom Patch hp49g.beep ignored. Please put the Rom command before any Patch. Cannot Load Bitmap EmuCalc HP49 [screen pc600x800pxls] zoom 2.bmp. Debug Off ==== Change hp49g.rom in the kml file by the name of the rom file (rom.e49). Title best emuhp49 screen hp131x64pxls pc600*800pxls Zoom 1 Author Jaime Fernando Meza Meza Model X Rom hp49g.rom <---here Patch hp49g.beep == That it does, even with flags -2 and -3 set (minor bug?); trying pi ->NUM won't help, either. You can, however, press NXT and then the CALC menu key, then enter pi on the stack; if flag -2 or -3 is set, then the numeric value of pi will already be shown; otherwise use ->NUM on the symbol. Finally, press OK. The 49G, thanks to its tendency to evaluate anything that you enter on a form, will perform automatic conversion unless the numeric flags have not been set, in which case you would still have to resort to the NXT CALC menu. In the 49G (1.19-6), going to NXT CALC for the Indep Low: Default setting puts a bint on the stack (minor bug?) [okay on 48GX!] Can you make a pi? No? Neither can I! http://people.deas.harvard.edu/users/students/Rebecca_Hwa/lyrics/berlin.html http://www.rnh.com/williamson/indexb.html ==== Any of these board games available for the 49? ==== Recently I gave my 32sii to my son then picked up a 48GX...... Assuming anything a 32 can do a 48 must do better! Right?....erk! I can't figure out how to enter Hex numbers directly at least hex numbers using A...F digits. I have gone thru the manual over and over ditto ditto ditto.. beat dead horse into bloody mush... :) I can put her in HEX mode but now I want to enter say F2A3h...... If I go the Blue? arrow # route I can enter numbers that are HEX but only hex numbers limited to numbers containing 0.....9. but if I try to use any of the A....F keys (the top row) the dang related function occurs (right now!) This doesn't seem very intuitive or logical to me....if you've declared # and are putting in a number why the heck aren't the functions suspended until you finish entering the number and hit ?!?!?!?! ==== Just press ALPHA before the hex digits A to F. I understand your frustration. I always keep the assignment that Simone Rapisarda suggested in his Smartkeys documentation (a program that automatically enters the alpha keyboard when entering a string: , and also when entering hex numbers: #) Refer to page 20 of SMTK.DOC on Goodies Disk #9 (in UTILS/SMTK.EXE): http://www.hpcalc.org/hp48/compilations/horn/horn9.zip Last Smartkeys version (but with abridged documentaion): http://www.hpcalc.org/hp48/utils/interface/smartkey.zip Hope it helps. ==== You have to put the calc in alpha mode to enter the letters A-F. Or do like me, have quick shortcuts to the letters: Keybindings: A 11.2 X 11.3 B 12.2 Y 12.3 C 13.2 Z 13.3 D 14.2 E 15.2 N 15.3 F 16.2 T 16.3 or something like that... :-) ==== ! Well, you must press the alpha key twice to lock alpha mode. (Or once if flag -60 is set.) I think it is that way because you can accumulate many additional things, like commands, functions and so on, on the command line, before pressing ENTER. The command line acts like a program waiting to be executed as soon as you submit it with ENTER. You might want to assign << {# ga } INPUT OBJ-> >> to the key # . (ga is the greek alpha } Then, when in user mode, pressing the key # will put a # on the command line and set alpha mode, so that you can directly input numbers and characteres. ==== Three ways to do it amazing! I like the automatic method. and George I did read the manual. If you go to the chapter 15 as the index directs you...you see even an example of how to enter a hex number! It's just wrong if you page 5 earlier I tried the alpha key to no avail before I discovered that the # was required. ====  Chapter 15 is binary arithmetic and number bases Chapters 1 and 2 are about entering characters and objects Don't get me wrong. I'm trying to direct you so that you don't get only this one question answered but many similar. Because all these automatic ways are just fine, but none of them will tell you why it is so. The easiest answer, for me, to give you is of course press the alpha key before pressing A,B,C,D,E,F). But i didn't want to give you that, or at least, not ONLY that. This attitude that i have has repeatedly made people to misunderstand me as a nasty, angry guy, while it is the exact complete opposite. I want to point people to where they will find the most info and what is most valuable to learn. As i said (and this is the whole point) while myself after i read the other guys' posts- it is very convenient) they won't teach you proper object editing and the alpha/shift keyboard. ==== How do I evaluate the following or even similar terms? Integral from minus indefinite to definite of e(-x^2/2) ? xy''+2y'+xy=0 searching for y ? ==== / Use the command | (red-shift, TAN) / The command needs four arguments: lower and upper intgration bounds, expression and integration variable. If you do this on the stack (RPN Mode), then: Press [blue-shift],[0] to put +infinity on the stack. Press [+/-] to chenge it to -infinity. Press [blue-shift],[0] to put +infinity on the stack. Enter 'EXP(-(X^2/2))' Enter 'X' Press [red-shift],[TAN] Alternatively you can enter the integral in the EQW. Press [EQW] to go to the EQW. Press [red-shift],[TAN] to enter the integral function and the placeholders for the limits of integration, the function and integration variable. The blinking arrow is now at the lower integration limit, so press [+/-] and then [blue-shift],[0]. Press [arrow-right] twice to go to the upper integration limit. Press [blue-shift],[0]. Press [arrow-right] once to go to the function that must be integrated. Enter the function. Then go to the integration variable and type X. Now you can press [ENTER] to put the integral in the stack. The HP49G doesn't automatically try to find the integral if you enter it from the EQW. You must press EXPAND to do that. Result: The HP49G can't find that. The command for solving differential equations is DESOLVE. It needs two arguments: The differential equation and the function to solve for. One of the notations for general derivatives on the HP49G is d1Y(X), which means the derivative of Y(X) for the first variable of Y(X) which is X. The second derivative is then d1d1Y(X). So you enter 'X*d1d1Y(X)+2*d1Y(X)+X*Y(X)=0' (the differential equation), then enter 'Y(X)' (the function) and then enter DESOLVE. Result: The HP49G can't solve this and you get an error: Unable to solve ODE Well, I wish I had better news but unfortunately the HP49G doesn't solve these problems right out of the box. I don't know if there is another way to do that, but perhaps some of the gurus out there can help?. ==== But sadly you are right. If I only put in this there is no solution. Maybe some one knows how to deal with them? Our Teacher said we have to use the square of the Integral if we do it by hand. I(e(-x^2/2)) -> (I(e(-x^2/2)))^2 = I(e(-x^2/2))dx * I(e(-x^2/2))dy For the calculator it is the same: no solution. And for the differential equation we have to use a Sigma-Series. A friend of mine has a TI89 which does these two equations right away. Maybe it is luck (I strongly believe that :) ) but sure there must be some way to solve the problem with the HP too?! ==== I can get my TI89 to solve the differential equation, but not the integral? The correct result should be 'SQRT(pi)*SQRT(2)' if I remember correctly... ==== I am trying to use the HPIDE for programming but I need HP-Tools to make it work. Where can I find HP-Tools? ==== HP-Tools to make it Simply where you got HP-IDE from: www.hpcalc.org Full set: http://www.hpcalc.org/hp48/pc/programming/sysrplpc.zip Nearly full set: http://www.hpcalc.org/hp48/pc/programming/tools.zip The V3.0.6 Win32 binaries: http://www.hpcalc.org/hp48/pc/programming/hptools-3.0.6-win3 2.zip I'd recommend the latter, 'cause they're the latest release ==== Schmidt's excellent InForm Builder creates Input Form for the old Input Form engine (DoInputForm) which can not handle the minifont. If you want to use the minifont you have to use the new engine of the 49G (IfMain), which can be coded directly on the 49G as a sys-rpl source or on the pc using the SDK. Search this newsgroup for IfMain to get details on how to do this and read the file informbox.h. The decomp object for minifont is BINT3. ==== SH> Has anyone (?) figured out how to implement minifont fields SH> within an Inform box? I know it can be done and, yes, I have seen SH> JYA's description of doing it using a PC-based program. SH> I develop entirely on the 49, using Schmidt's excellent InForm SH> Builder. I am, therefore looking for a way to do this within the following SH> workflow: SH> 1. Create Inform box using InForm Builder. SH> 2. Save as SysRPL code. SH> 3. Add/modify that code to specify minifont for the fields. SH> Any ideas, any takers? In the SysRPL code generated by IFB, just replace DoInputForm with FPTR2 ^IfMain. Then make sure that the DECOMPILE argument for the fields you want to have displayed in minifont is odd, i.e. if the current value is BINT4, change it to BINT5 (in other words, make sure that the least significant bit of the bint is set). Then run the code and look at it. It is possible that you need to change the position and size of the minifont fields by hand to make everything nice. Hope this helps. DoInputForm and IfMain are not fully compatible, but for me the simple procedure above has worked well. , I think it would be nice if IFB could be updated to use IfMain and to allow this minifont option. Clearly, it seems to be hard to support all the message handling stuff in IFB, but the minifont issue would be simple. ==== It would be nice if IFB would use the new inform box routines, but the problem is that the source code is one file @ ~50kB, and it takes ages only to find a simple bug. If anything major should be changed, I might as well write the program again from scratch. That ain't happening, I'm afraid. If anyone want to use the screen layout of IFB, or can use the source code for anything - please feel free to do so. ==== It is almost impossible to crash your 48 if you do not use software in srpl/asm. There are a couple commands that under certain circumstances can cause a memory clear. take a peek at this: http://www.hpcalc.org/hp48/docs/faq/48faq-3.html#ss3.5 ==== display. Maybe that was what did it, but I was disappointed... VW ==== Don't use the ticking clock. Experience says you're better of without it. ==== for your row vector [10 0], make it [[10 0]]. ==== I don't exactly understand the example. [10 0]*[5] [3] What is the [3] that stands alone? Should this be the result? ==== no the 5 and 3 are column vector. ==== Oh, I see! Then Aaron got it right, right away! You should enter [[10 0]] (row matrix) and [[5][3]] (column matrix) and then press *. ==== apparently one needs the extable library for that. ==== Neat function, I didn't know it existed. BTW 's program should read: << -> val << 1 0 1 val UTPN - >> >> ==== That was really my point, I wouldn't just tell you that I know two people :-) Yes, an acre is a chain by a furlong by definition I just use the football field unit when I need to measure by eye, soccer fields are 100 m long Yes, as the Germans would say: Bei uns ist alles grosser, besser und billiger. - Our stuff is bigger, better and cheaper. I picked it up in the Czech republic, it means beer. Wouldn't ,for the life of me, know why. :-) ==== Try a Google search on MIKTeX. ==== I got my ME package preloaded on a new hp49 from http://www.chotkeh.com. I don't however think it's worth the money. The steam tables are nice for avoiding interpolation but by the time you get the answer you probably could have done the interpolation anyway. Past that, most of the time the value you need is a straight lookup, which would take less time and be more accurate. sorry for not responding earlier, I was too busy. On the bottom of p. 23-15, it says Display the conic example plot... (x^2+y^3) mod 2 < 4 Anything mod 2 must be smaller than 4 (aside from the fact that the concept of a real number modulo something is a bit dubious), so that gives you a black screen (on the HP 48). Of course, it should be (x^2+y^3) mod 4 < 2, which produces the screenshot on the next page. I would be interested to know which formula the EXAMPLES directory on the 48 contains, just out of curiosity. The error has been copied to the HP49G User Guide. On the HP49G, following the advice from this thread, you need to type (x^2.+y^3.) mod 4. < 2. and set the step size to 1 pixel in the setup window to get the same thing. ==== Rom version R holds the record for most stable; after all, it hasn't changed in more than 8 years, and it's a sure bet that it never will Too bad that it won't run in an HP49! ==== Now, what do you want so many commands for? Just use EVAL ;) 3 5 << + >> EVAL You see, you can do it with just one command. Now, if War Starter counts << + >> as being another command, we should get another ROM with only the EVAL command, that should get a number from the stack which would represent what we now call commands. For example, if EVAL gets a 1, it should get two integers from the stack and add them, thus 3 5 1 EVAL would yield 8. We should really have that ROM available, just for pleasure. Hey, wait. We have SYSEVAL. [...] [About customability, specifically being able to change the keyboard layout] ==== It is,i agree but it is a third party software when such a function should be built-in in the TI89. But you certainly don't know that the main target of TI is high school. In fact the TI89 fails to find the antiderivatives of any functions of this kind: -(f'(x)+g'(x))*exp(f(x)+g(x)) -(f'(x)+g'(x))*cosh(f(x)+g(x)) -(f'(x)+g'(x))*sinh(f(x)+g(x)) -(f'(x)+g'(x))*tanh(f(x)+g(x)) And it also fails on antiderivatives of functions such as: -(f'(x)+g'(x))/sqrt(1+(f(x)+g(x))^2) -(f'(x)+g'(x))/sqrt(1-(f(x)+g(x))^2) And many others. Those antiderivatives that even a high school student should be able to find. Then why you come here with some selected few examples that the TI89 can do and where the HP49 fails ? You should know that both these tools are very limited in this domain in comparaison of software such as Mathematica,Maple,etc... The TI89 is even more limited because it doesn't have Risch type algorithm and has the 64 KB limitation despite 188 KB of user RAM. Sure the TI has an interactive Integration by parts function. Sure.Have you seen any real improvement of the ROM since the TI89 release,3 years ago ? New math commands or significant improvements of a built-in function,besides solve command ? And do you know why i give you this example ? I give it because the it show 2 recureent weakness of the TI89: -the use of inapropriate algorithms -The 64 KB limitation This problem could be solved with any of this weakness removed. I just wonder why TI hasn't removed at least the algoritm problem. Do you know that there was a program which enable to solve this problem on TI-Cas.org ? The author has even given for free the right algorithm to TI but they have refused to implement it. Just to show you how reluctent to improvements they are. Hey,calm down. Well despite repetitive command the HP49 has still much more function than the TI. The TU has less than 500 user availlable commands even with the possible combinaisons used. Which is far from the over 800 user availlable commands of the HP49. After 6 years of patience i have decided that TI doesn't really care about the category of users i belong to and that they kiss my a**. Sure,when the keyboard layout and the Text editor are as poorly designed as the TI89's one,it is. But when you have a great keyboard layout it is as easy to type numbers as letters and off course programs. It would be the case if the TI89 really takes advantages of its C.P.U.And if it was the case everything on the TI would be much faster especially numeric computations,TI-basic,symbolic computations,GUI,3D plotting,etc... Btw how much time needs your TI to solve a 25*25 numeric linear system ? This is the first question my math professor asked me when he saw for the first time my TI92. We have been both very dispaointed by the result. We are still disapointed by the result with the TI92+ HW2 and A.M.S 205. Sure if you like to use short awful name it is your problem. But actually,we are in the computer age and most pepople used to long names and could at least expect it on almost every computer related tool. And this included so called advanced calculator. I think that if we were able to rely on TI R&D team to improve the TI89 software,they would have already done major improvements on the software. And after a discussion with a TI engineer,i have come to the conclusion that they aren't interested in improving the built-in software at all. And why don't you try the example program i have given ? Are you affraid to be disappointed by the TI89 ? With newer ROM perhaps but it wasn't always the case. I do remember a time where ROM where very unstable and where you can't even rely on Flash ROM to recover your datas. Btw the HP49 is incredibly stable with the later O.S too. And you haven't answered to my question. This exactly what third party developpers are getting for all the work they have done for TI calculators. Then why do you raise the subject,then ? Generate revenue for TI in what way ? Expecting people to pay poorly designed Flash apps when they can use third party free programs which are much better than Flash apps despite all TI limitations ? Pisssing off developpers which have done incredible job to bring exceptionnal software to the TI calculator,thus allowing like you to have such attitude ? Then i expect that TI relies on other revenues sources. On a programmable calculator the alpha mode is very important. And it simply sucks on the TI89. Sure can your TI beat the HP49 for numeric computations,3D plotting/rotation and zoom,user language speed,taylor series,etc... ? Partial truth is not the truth. According to the vorlons(B5) the truth is a three edges swords. Your side,their side and the truth. What you have express their is only your side. Neither me nor anyone esle can claim to have the truth but what we can do is to be as objective as possible. And your post was extremely biased. My point is that both calculators have strengths and weaknesses. It just happens that software wise,the HP49 is a far superior calculator because of many factors. One is that it has been developped over a tool which was designed for professionnal users while the TI89 was designed with high school students in mind. Off course the third party programs significantly improve the TI89 but they are some points where they can't or where is no third party programs to od something. And this where you show your bias because when the TI89 can't compensate a HP49 feature you find that it is minor and not important. But as soon as a third party give it to the TI89,this features becomes important. This was the case for both EQW and RPN. I am sure that you were one of those who were minimizing the importance of HP49 EQW and who have completely changed their mind with the E.W EQW release. I am sure that if you knew that there was a third party program which enables the TI89 to have virtual tree directory structure you wouldn't have found this feature minor. Your name is an insult because your attitude is completely at the opposite of what it means. ==== Just for my interest: how long did it take? And did you ever check out, how long it takes on your HP49? ==== TI 89 With a random matrix created through randMat(25,25)->a and a vector with randMat(25,1)->b and solving with simult(a,b) in aproximate mode it takes 203.84sec!!! HP48 with 25 1 RANM and 25 25 RANM / takes 45.42sec!! ==== My 49g in Approx mode took 65.69s to divide the 2 matrices. My 89 in approx mode took 285.26s to do simult(a,b). Almost 5 minutes!! ==== I tried four different sets of random matrices. On the TI89 HW2, I get: simult(a,b): 211-213 seconds a^-1*b: 399-401 seconds On the HP48GX, I get: b/a: 19-19270 seconds a^-1*b: 53-53270 seconds On the HP49G, I get: b/a: 19-22 or 68-71 seconds a^-1*b: 55-57 seconds On the HP48GX, the timing differences are almost non-existent. On the TI89, the situation is the same - very stable execution times. The HP49G is a somewhat different story. The timing is pretty low and consistent, but in some cases, the '49 suddenly spends three times the normal execution time. The reason for this is simple: The HP49G has two types of matrices - symbolic (Type 29) or numeric (Type 3 or 4, depending on real or complex contents). The symbolic matrix type can hold symbolic as well as numeric elements, so it may not be possible to detect any visible differences between the two on the stack. RANM on the '49 creates Type 29 matrices in both exact and approx modes, so to make certain that a matrix containing strictly numeric values only, is Type 3 or 4, apply the ->NUM command on it. After this, the execution times are comparable to those on the HP48GX. The solve() commands on the HP49G and the TI89 can also solve these types of problems, but at these system dimensions they take ages to execute. ==== You must be right there. Maybe my test was slower because of free ram?. I only had 12kB available. BTW I measured the time with teval... could that interfere?. ==== As you can see in my previous post, I've got same time (20s) in my GX with 32K of free memory. I've measured with TEVAL too. (Erable 3.2) ==== TEVAL is pretty good - I don't know what you have done extra? Even if I add the generation of the matrices with RANM, I still only get 28 seconds. I have 77 KB RAM free in my HP48GX. ==== Ah, I don't have Erable. The TEVAL I use was included with some utilities I don't remember where. Now tell me that the speed improvement is because of Erable. Otherwise I've got the slowest 48 in this Newgroup >-(. ==== Erable doesn't improve the speed of the HP command << / >> that we are using for this test, but I'm not sure if your 12Kb of free memory are the guilty... Sorry if this doen't help What most CAS commands start off doing is to call ^CASFLAGEVAL, which in turn calls the mis-named ^SAVECASFLAGS, which conducts the famous friendly dialog asking you all those questions, such as Radian mode on? and Purge current variable? etc. And if you reply yes to that last question (which isn't optional :) then it goes ahead and does PURGE (a venerable, very clean command :) whose permanent, supported address (even in the HP48) is: 08C27 ! Therefore, you can't avoid doing PURGE, because it's built right into every CAS command anyway, so just grin and bear it :) Why do I invoke PURGE myself, instead of waiting for the CAS command to do it anyway? Firstly, because ^SAVECASFLAGS (and/or ^SAFEPURGE) currently refuses to be silent about that last question (and silence is golden, right?); secondly, so that I can test whether there exists another variable of the same name in any higher directory, which is subsequently just as troublesome as if you hadn't purged the original variable in the first place (this is another small logical flaw in the CAS and/or OS), and if so, I generate a name conflict error to explain what's wrong (as much as can be explained in 22 characters or less :) It's curious that employing SYSEVAL in a UserRPL program inspires so much fear, whereas no one pauses for a moment before using anybody's SysRPL program, which generally calls *only* internal ROM functions :) > Cases like putting DEG or RAD in the menu can be a little annoying. > But then one can mix it, like { DEG RAD RECT CYLIN } > { XQ EXPAND } XCASM + MENU, to have both mode-changing > and other commands in one menu. Most extraneous commands can be included harmlessly, but for commands meant to change flags (thwarted by our current diligence to preserve and restore all original flags), you can also do this, since lists within menu definitions are unaffected by XCASM: { { DEG DEG } { RAD RAD } ... EXPAND } XCASM [T]MENU Note also that XQ never goes through ^CASFLAGEVAL etc., but is actually *intended* to change to Exact mode (you can use XCASM to *prevent* that mode change, if you like, or else treat XQ just like DEG or RAD, etc.) You can also use separate RAD mode toggle and POLAR mode toggle functions, formerly printed on the keyboard of the HP48G[X]: First of all, is it a TDS Cogo card? If it is, you may not have installed the card properly. Try it again, with and or without the memory card. TDS Cogo cards are very stable, and have been for quite some time... If you have a memory card you may need new batteries for it. If the cogo card is not TDS, it may need batteries too. If you got it from a pawn shop do the three finger salute on it. (ON+A+F). If you have no need for traverse and closure programs, intersections etc. you may want to get rid of or shelve the cogo card. If you just want hms triangle programs please try Survey Triangles,(I think that is what I called it),which is, of course, available at HPCalc.org. If you do not have sufficient memory to run the programs on your card... you probably have too much barf or garbage stored in RAM or in your hidden directory. ON + A + F and start from scratch... > I picked up my 2nd HP48 at a pawn shop a couple weeks ago for $30. It looks to > be in good shape, functions OK and came with a Cogo surveyors ROM card. > > While I am not a surveyor, the Cogo application appears to have a set of > triangle solutions routines which I would like to get working. Only trouble > is, I don't the manual which came with the original card. Can anyone help me > get this running? > > The Cogo card is installed into slot-1 (the 32K-128K slot nearest the keypad). > I also have a 1MB memory RAM card installed in slot-2 (ports 2-9 avail). I > suspect I have to do something to the memory allocation to get the program > running, but none of the obvious ideas (to me at least) seemed to resolve the > problem. > I'm trying to build an external device that I can connect to my hp49, but it needs at least two inputs. Or in other words, the calculator would need at least two different pins that can transmit data to the device. My research turned up nothing but some documents that focused on creating some calculator cable (hp->pc, hp->hp, etc) and thus only included pins 6-9. Does anyone know what the other pins are and which, if any, can be used to transmit data? > I'm trying to build an external device that I can connect to my hp49, > but it needs at least two inputs. Or in other words, the calculator > would need at least two different pins that can transmit data to the > device. My research turned up nothing but some documents that focused > on creating some calculator cable (hp->pc, hp->hp, etc) and thus only > included pins 6-9. Does anyone know what the other pins are and > which, if any, can be used to transmit data? 1.11 What are the serial port pinouts? Top of Calculator 6 7 8 9 10 +-----------------------+ | o o o o o | o o o o o / --------------------- 1 2 3 4 5 Bottom of Calculator Pins 1 to 4 are used for RS232, and pins 5 to 10 are used for the overhead display connection. Pin Signal 1 Shield 2 TX 3 RX 4 SGND 5 VCO 6 XSCL 7 LP 8 LD0 9 LD1 10 GND > I'm trying to build an external device that I can connect to my hp49, > but it needs at least two inputs. Or in other words, the calculator > would need at least two different pins that can transmit data to the > device. My research turned up nothing but some documents that focused > on creating some calculator cable (hp->pc, hp->hp, etc) and thus only > included pins 6-9. Does anyone know what the other pins are and > which, if any, can be used to transmit data? You can too see the various pinouts configuration in: http://enrico.freeweb.supereva.it/hpcables.txt for 49? > implemented in the 49 > calculator to an excellent calculator. > However, I found a library which might interest some of you. For java there is a package called the Colt library and it's designed for scientific calculations. Because I heard that people were thinking of using (strong)ARM based RISC processors, this could be really interesting. As some of you might have noticed is Java on PDA's a lot quicker than on most PC's and seems to be a viable platform for developing calculators. A link to this project is: http://tilde-hoschek.home.cern.ch/~hoschek/colt/index.htm Can someone tell me if it's possible to compose a complex number from the stack using a modulus and argument (magnitude and phase)? I wouldn't know how to enter a phase on the stack. if I type 1 enter 45 and then real to complex command. I get (45,<88.7) instead of (1,<45) what I would like. > Can someone tell me if it's possible to compose a complex number from the > stack using a modulus and argument (magnitude and phase)? I wouldn't know > how to enter a phase on the stack. > if I type 1 enter 45 and then real to complex command. I get (45,<88.7) > instead of (1,<45) what I would like. , 1 45 << ->V2 RECT V-> R->C CYLIN >> EVAL http://www.hpcalc.org/search.php?query=Cannon http://www.geocities.com/SiliconValley/Heights/5888/ > > > I don't blindly trust a company. Two days after buying my > > > HP49, I installed 1.19.6. It was completely painless. > > > > Were you able to determine in two days, that ROM v1.1.9.6 was stable? > > There is simple not one single bug in the '49 software, when it comes to > > v1.1.9.6? I can see that you have looked thoroughly indeed. > > > > I switched from TI-89 to HP49, and found the former marginally faster > > > for some operations. Certainly not enough of an improvement to > > > influence which one I use. > > > > > An ARM-based calculator would be nice instead for on-board power, but > > > not if I have to recharge at the day's end. > > > > How about every other month? You're very skeptical about the ARM cores, if > > you think such a device necessarily needs to be recharged every day. > > > So, can you tell me the how SOLVE works, and what its limitations are? > > What's the maximum order you can develop a series to? Why? Why doesn't > > > Why are you throwing bug reports at me? If your point is that > occasional bugs make it hard for the new user to learn, then yes, this > is the case with all software. Fingers crossed, the source will at > some point be out for community fixing. > > > > 3. I downloaded various other docs from hpcalc.org, e.g. a UserRPL > > > tutorial, which I worked through during spare moments. I grabbed some > > > other docs that I will/have read when I'm ready, such as a SysRPL > > > manual. > > > > And all that in the couple of days it took you to decide to use ROM v1.19.6? > > Bravo! > > So I spent a weekend plus an evening learning how to use a calculator, > and getting more docs to learn more. I guess I must be sad. Oh well. > > > > sexy apps to install, such as Emacs, which made my life easier. > > > > Because you already code in SysRPL, or...? > > I'm talking about *everyone*, not just the official developers (and > JYA seemed to give a lot more insight while working than the TI > workers did). Random example... ticalc.org TIcalc has an IRC help > channel on EFnet, #ticalc (or whatever it's called, can't remember > now). It doesn't. You go in there, talk about TI, and are told YOU ARE > NOT ELITE and kicked. Grow up people! > > No, students who weren't willing to put in extra effort should have > realised that they shouldn't be buying a complex calculator *at all*. > Or if they do, then they should ensure they buy from a company with a > good returns policy. > > > The HP49G can show you the steps of all calculations made, but it requires > > a broad knowledge of calculus and French ways of teaching. You need to know > > for example the Risch algorithm, any applicable heuristics and all other > > algorithms normally used. > Ē { # 0d # 0d } PVIEW ERASE 1 'a' STO 1 'b' STO .5 'c' STO .5 'd' STO WHILE '1==1' REPEAT '(a,b)' ->NUM PIXON a c + 'a' STO b d + 'b' STO CASE a 10 >= a -10 <= OR THEN c -1 * 'c' STO END b 6 >= b -6 <= OR THEN d -1 * 'd' STO END END END Č Put both the object(s) and name(s) into a list: @ 'name' OR { names } ==> { obj1 'name1' obj2 'name2' ... } << { } + 1. << DUP RCL SWAP >> DOLIST >> @ Store objects from a list like the above: << LIST-> 2. / 1. SWAP START STO NEXT >> ==== Finally i have found a 48gx, but now i have a doubt: the 49g costs only 30 EUR more. The calculator must replace my 48sx (with 128kb HP original ram card and equation library card, which i can use on the 48gx), used only for work: numerical calculation, hydraulics, vessel, pump,heat exchanger calculations and so on. Lots of units conversions. No symbolic, no algebra, no games. I am a chemical engineer and work for an engineering and construction company, so i use my calculator very hard, at office and on field. My programs are written in user RPL and i use my 128kb card as a backup floppy (not merged memory) in which i keep programs and important data. I used it for about 60% of its capacity.. I use only RPN (i would even on the 49) and love the 48's keyboard. I need a good unit of measure management, at least as the 48sx's, absolutely nothing less. ==== No symbolics you said? And units conversions that *always* work as expected? And a keyboard with which you can do speedy input? Then buy the 48GX. == You can toss the eq. library card and download a free periodic table. The Periodic table and Tetris are the only two things that don't come built in with the 48GX eq. lib that Hp now includes in the ROM (that library is not available on the 49 except as a downloard). The 48Gx is everything the 48SX was and more. It has 128K and with your 128K RAM card it would have 256K and an extra card slot for extra RAM if you want or need it. You get the same keyboard and a calculator that is over 50% faster. Now, for the reasons to buy a 49G. It is probably tougher (All things considered, the screen cover protects the LCD, which the 48G series is exposed and that is the weakest part of any graphics). 1.5 Meg RAM and a faster calc when using EQ Writer. The keyboard alpha keys are better laid out. It does have lots of nice extras, but if you are happy with the 48? The drawbacks are the keyboard. Rubber keys do not last forever, like a plastic molded key will. The keyboard feel is also inferior. I have all of them and use a 48G on my desk. Its cheap, easy to use, keyboard layout with the enter key where ITS supposed to be, units conversions, and eq library that sometimes comes in useful. It has a clock with seconds tick vs the 49 doesn't out of the box (you have to download a clock to get seconds tick). The keyboard is great vs only good for the 49 (I won't knock the keyboard, it is okay, it just isn't as good as a 48, and I don't mind using it, I just like the keyboard of the 48's better). ==== i have heard something about two different screen types (the classic green/blue and a white/black one) and different keys (molded and painted) which are on GXs from different factories. As i can choose between many GXs in the shop :-PPPP , the questions are: 1) what is the difference (apart the color) between the two types of screen? which is preferrable considering readability, durability and scratch resistence? 2) how can i distinguish between the two type of keys? (of course, i would prefer molded ones) 3) it is possible to determine from the serial number the type of keys and screen fitted on the calculator (i know the meaning of the numbers and letters of the serial...) ==== I don't think you need a 49 or a gx for that. Have you looked at the lower level calcs? sure, a g+ (or gx, if you want to put that ramcard to use) would do the job for you, with EQstk in it (for pretty viewing of units), and you get to keep your progs, but if a lower calc suits you better? In any case, try what you want to do on a friends calc and see what suits you better. ==== The newer display bacause of it's higher contrast! If you hit it hard on the floor, it will break in most cases, of course I never saw an 48GX with scratched display. The display is slightly recessed, and it's covered by a thin 'anti-reflection' plastic sheet. unfortunately you have to decide to get the new display *or* molded keys. AFAIK both features in one machine is nonexistent. But I'd prefer the much better display of the newer chinese units. The key painting may be acceptible, you still have hard plastic keys with somewhat like the HP key feeling. Someone recently asked a similar question in the MoHPC forums. Here's the link : http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/forum.cgi?read=14858 Of course displaying *all* labels would make things extremely difficult. However it would be handy if we could display one or two letters next to the node (instead of the number), say, with a special label. That wouldn't need much more space than a single number, and could possibly be turned off with the same flag, but it would allow us to identify nodes more easily. ==== That must've been the abnormal winter olympics. Where were they held? I've heard of places where men are men and sheep are nervous, but I've always imagined that they were nervous for a different reason. == Well, having taken a second look at it, displaying the first node label shouldn't slow down the graph drawing too much. But then ppl will also want the edge labels which are a bit trickier, since there may be many edges between any source-target pair, in both directions... But I'll see what I can do about this. ==== this is the first time I post in this ng, since I have just bought an HP 49G calculator. I have a very stupid problem but I can't figure out the solution. I carefully read all the documentation I received with the calculator. Let me explain it to you. Suppose that I have to do the same calculation a lot of time. For example I have txarge number of couples of A and B numbers. I can use the way: Num Solve - 1. Solve Equation - Solve Equation I enter a dialog box and I put my equation in the first field 'EQ'. I put my A value (for example 2) in the second field (A:). I put my B value (for example 3) in the third field (B:). I move the cursor on the first field(EQ) again and i press EXPR=. I exit the dialog box and I found the result in the main screen. In the main screen I read: External External Expr: 11. If I have to do the same calculation (A+B^2) with different values of A and B I have to enter again the Solve Equation editor and exit it to find the solution. But if I have many values to enter this proceeding is very time consuming. I am sure there is a faster way....(I saw something in the HP48, but not in the HP49). I would like to enter my formula (perhaps in the EQ variable)and have the calculator ask me for the A and B values respectively, giving me the result. And then ask me the next A and the next B and so for. Is there any way to speed up this process? I thank you very very much for your kind support and I hope to hear from you very soon. first thing: If you see any 'External' entries on the stack (main screen) when using built-in stuff, you should upgrade your ROM. For your equation: You could write a short program that does just what you need, and store it in a variable. Enter << SQ + >> , then press the ENTER key. Type a name, like 'AandB2', then press STO. Now you should press the VAR key, and a menu key labelled 'AandB' should be there. Then enter your value for A, press ENTER, key in your value for B, press the menu key labelled 'AandB'. You should see the result of your formula on the stack in level 1. Every subsequent calculation of new values can be done that easy now ==== One small addition: My example assumes you are in RPN mode The fastest way to solve your data entry problem is with the use of lists. I'll assume you're in RPN mode, if not, I'll suggest you switch fow now, it will make things easier an faster. type this program and store it in a variable (ie. MYEQ) << -> A B << A B SQ ADD >> >> MYEQ STO where << >> = program delimiters (RightShift +) the forward arrow is RightShift 0 To use the program, simply put the data for A and B in the stack as lists and run the program. if A = { 1,2,3,...} and B= { 4,5,6,.. } you would enter on the stack: {1 2 3} {4 5 6} MYEQ the output will be in list form also, like {17 27 39} ==== First of all, thanks a lot for your suggestions. I will take care of upgrading the ROM in the next period. I was in algebric mode and I switched, according to your instructions, to RPN mode. I created a very simple program called L, based on the very versy simple formula L=A+B. I entered first A pressed enter (4). I entered B and pressed enter (5). I pressed the menu called 'L' and the calculator displayed the correct result (9). And I saw that this method is quite fast....thanks a lot. Now I have another problem. I have no problems with very simple programs, like L=A+B. But what if the formula is not a simple addiction but a more complex formula with a lot of parameters. Which is the ordern in which I am supposed to enter the parameters? And... Is there a way force the calculator, like I saw in some Sharp calculators, ask me the parameters, in order to be sure to enter all of them in the right order without forgetting a parameter? I saw that in the Sharp calculator you write the formula, you execute it and the calculators prompts for the parameters, for example visualizing in the top of the screen A? for the A parameter and B? for the B parameter. In this way you can follow the parameters entering procedure to be sure to enter all the parameters and the correct values for each of them. ==== I'm glad that you've seen the light. RPN is much easier to use, but if you've been using an algebraic calculator, it may take a little getting used to. Definitely. See the PROMPT, INPUT, and INFORM commands right offhand. ==== I think you need to have a look in the manual about how the numeric solver works. EXPR= dosn't solve the equation, what it does is evaluate the left and the right sides seperately so you can chck the calculation of the formulae you entered. In your case if you enter 'A+B^2' what you are really asking the HP49 to solve is 'A+B^2=0' it just happens that EXPR= will give you the answer your looking for as the evaluation of the left side. Enter the equation as 'Y=A+B^2' in the numeric solver, enter your values for A and B then put the curser in Y and press SOLVE. For quick and dirty entry I use 'X1=X2+X3^2' to save having to change to Alpha mode to enter equation. As with most things on the HP49 this is only one way, there are sure to be other ways that may suite how you work including on the rarely discussed dark side of the 49, Algebraic Mode. I don't think you should be getting external on the display during norma operations. As Raymond suggested check your ROM by entering VERSION. Latest ROM is 1.19-6. ... ==== me step-by-step what to do? Just suppose I have the simple formula A+B and I would like the calculator prompts me for inserting A, then B, and then it calculates A+B. (I just want the calculators explicitely asks for the parameters in order to avoid forgetting a parameter or put the value of a parameter in an other one). ==== In the numeric solve I entered: 'Y=A+B^2' After entering the values for A and B in the right fields, I move the cursor to the Y field and I choose Expr=, it calculates Y without leaving the dialog box, so this is faster then my old, WRONG, method (have to say that HP manuals in my language are very poor and they do not explain very well as I upgraded my ROM to the 1.18 version but I keep on getting External errors in the main window, but I read that 1.18 is the last official ROM released by HP (I downloaded it from their website and there was not the possibility to download the 1.19), I am not sure about putting on a beta version. Are these errors very bad? You shouldn't be getting those external errors with that ROM version. Nevertheless, you should definitely upgrade to version 1.19-6. You can get it here: http://www.hpcalc.org/hp49/pc/rom/beta1196.zip Even though it is beta, users tend to agree that is even more stable than 1.18. You get many more features, better stability and less delays due to improved garbage collection. You can keep asking in this ng but from what I know, the vast mayority of users use 1.19-6. In this ng there are many expert users (I exclude myself, I'm also rather new to this but I learned a great deal from this ng), all use 1.19-6. Follow their lead. Make a note of that site (http://www.hpcalc.org). You can get a LOT of great stuff for your calc there. As for your original problem, have you tried the simple program I suggested? I tried it and it accomplishes what you're asking quickly and efficiently. ==== While we're in the realm of complete fantasy, I may as well suggest that it would be much better to have them reestablish Corvallis Division. Corvallis Division was originally established in 1976 specifically for the calculator product line, but now it is responsible for ink jet cartridges. The transfer of calculator responsibility to Singapore was clearly the point of no return. ==== In fairness to the HP support, they seem to have done some basic check, as first they forwarded me to http://www.epita.fr/~avenar_j/hp/39.html made sure I've read http://www.hpcalc.org/goodbyeaco.php and took another day to check if there were any engineering specs at their disposal. Only after this they redirected me to the newsgroup. ==== Ok, so first they directed you to a web page of an ex-employee whom HP either indefinitely laid off or terminated. Very nice to know that they still have a high regard for him. So that you'd know that you really couldn't expect much support from HP. So they can try to support their product if they don't find first someone else to it for them. Oops, they should've thought of that before checking to see whether they had any helpful information available. Great support policy. I'll keep it in mind if I'm ever tempted to buy anything from HP in the future. ==== Ah yes, tried that one as well: zipped newsletters in PDF format with calculation examples and no design specs. In the meantime I figured out that investigating the saved lists with a 1-digit element is a waste of time, as they're stored in byte with offset 0xE starting from binary 1101101 and skipping every value ending with binary 00, while the byte with offset 0xD works kind of a tag with binary values xxab0000, where x-bits change in some undefined way, a-bit changes in cycle 1-0-0-1-1-0-0-.., b-bit changes in cycle 1-0-1-0-.. and only 0x2B3100 trailer may be a variation of trailer(s) used in multi-digit lists (0x2B3100, 0xB01203). Playing with x-bits can make the calculator reloasd when attempting to open the list. Multi-digit lists leave the 0x0D:0x0E bytes at 0x3093 and pack the numbers 2 digits per byte (1 decimal digit in 4 bits). If I recall correctly, it's similar to packed format in old IBM mainframes. The digits are recordered in little endian (least significant first) with sign byte jumping around the most significant digit depending on whether the entry number is odd or even and the delimiter changing between 0x309302 and 0x3329 in same manner. Mostly positive integers were used so far. Looks kind of .. well, I'd better keep comments about the format to ==== The following applies to the HP38G. It should also apply to the '40G mutatis mutandis. Lists (and arrays) must be in binary mode, with this header: HP38BinE*Ln... The E is for lists and arrays, the asterisk is the byte x02, which is the length of the name: Ln, n can be 0 thru 9. And then, what follows is a list of real numbers, with the very same format as the HP48, same prolog too. Object structures are described in many places. For example, refer to this classic doc: http://www.hpcalc.org/viewzip.php?id=233&file=HP48SX.DOC Note that numbers in the list can appear as five nibble quantities, which are pointers to reals in ROM (just frequently-used values: zero to nine on the HP38G, IIRC) I think that the quickest way to transfer a list of reals is create it on the HP48 or download it to the HP48 (or emulator) and add the proper header. The commands ->ASC and COD in the hack library will be useful. Just remember that no pointer to built-in objects should be in the list, their addresses were changed in the HP38G! And finally send it to the HP38G via JYA's S38: http://www.epita.fr/~avenar_j/hp/hp38/aplet/hp38.zip [For the '40G, send the resulting list to the PC in binary mode, and modify the header with a hex editor, my favourite is Mikael Klasson's HexIt: http://www.geocities.com/thestarman3/tool/dl/hexit157.zip ] ==== I was sitting in my DSP class tonight and learning about partial fraction expansion for the zillionth time and I was trying out (with my 48) one of the examples he was working by hand just to make myself grin, and it happened that he used some denominator that had complex roots, which alg48's PF doesn't do.. sorta. I factored it by hand and made the denominator the product: (x+alpha+beta*i)*(x+alpha-beta*i), and tried PF coefficients of the numerator and denominator in [ ], and it does partial fraction decomposition on the quotient of the two polynomials, and it works even if the denominator has imaginary roots. So.. I was just going to ask, would anyone find this useful? If so, I'll upload it to hpcalc.org. I didn't check to see if there is already something like this so I might have re-invented the wheel.. but it passed the time. =) It requires alg48 because I just use PF at the end. As it is now, I'm also using QPI to make things easier so you need that too (which all HP48 users should have in their arsenal anyway ). The program is really small.. 103.5 bytes. ==== Store the following code in RPLASM.A for example : STARTRPL MACRO GOSUBL startrpl CON(5) =DOCOL RPL ENDM ENDRPL MACRO COLA ASSEMBLE CON(5) =DOCODE REL(5) end ENDM GOTO + startrpl C=RSTK A=C A PC=(A) + Then, insert an INCLUDE RPLASM.A at the beginning of the first code object the uses the macros. This should work with Jazz and the HP-Tools. ==== S> !@#$%^& New I'd be embarassed if I'd missed a command. In my defence S> SCROLL's spelling is not even remotely related to VIEW and there are over S> 700 commands to look through. That's right. I remember having trouble to find SCROLL as well. S> The Command Listing by function I copied from this group also lists SCROLL S> as unknown in the miscelaneous section. A very usefull resource but I S> didn't save the authors name and he didn't sign it so I can't thank him. Of course you can. http://www.google.nl/advanced_group_search?hl=en and search for ==== I'm looking for a way to make a command like N(2) which gives (or is equal to) me the (normal) cumulative probability in the right hand tail. I.e. I enter this: N(-2.2) I would get 0.0139 Are there any software out there which adds this feature (or perhaps it's allready there) ==== how can i calculate few simple complex numbers calculations with my HP49g? how can i insert numbers in there. example.. 2*e^(pii / 4 j) + 6*e^(-pii / 4 j) answer is 8... how do i insert those?? answer needed very quick.please.. PS. ja jos suomeks viel212 sais niin kyll212 ois hyv212. mutta kyll212 englantikin k212y. ==== dam i was stupid =) .. yep now i got it but hp49 give answer like (8,0) but how can i change it like 8<0degree i mean in the polar coord <=> rectangular form?? i hope you get it. ==== but If you hit the MODE key, you see a choice of Coord System. Choose here which form you'd like to see complex numbers (and 2D/3D numeric vectors) displayed in. When you're highlighting that field, a shortcut to changing its value is the +/- key ( the W-alpha-key). ==== How about this? : << :&: 2 EVAL @ BINT 2 # AAB2h SYSEVAL @ PORTSTATUS DROP2 DROP2 # 5380Eh SYSEVAL @ COERCEFLAG .. ----- ==== On Tue, 12 Feb 2002 18:46:50 +0100, Albert Graef Here's a fix. Insert the following code immediately after the % get next object in the list comment : AD0EX C=A.A RSTK=C A=DAT0.A LC(5)PRLG A-C.A C=RSTK ?A#0.A { D0=C } == As far as I can remember, it is because ROM objects in a list are referenced by a 5-nibble pointer, not included. The ML code thus must solve the reference - but ForAllComposit does not do that and thus can only be used with embedded objects. The SysRPL 'R command will work with embedded and referenced objects, so there's no problem there. Werner ==== in the thread I cited, really saved me. I use different variations of it to do all the list processing in my GraphWriter program, so thanks again. :) 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 <3C6996D7.DF0AFE97@iname.com> <3C69DBFC.A226E899@miu.edu> ==== John H Meyers schrieb: So that's actually an entry point, not a memory location. I heard (hehe) the area is 25EC-B3C7 (and the other at 25565-40000). So, yes, it's in the stable area. ==== . Hmm, I don't exactly remember but I think you could use EQ-> - to bring both sides together and then COLCT ISOL ? (I don't remember if the command is speleld COLLECT or COLCT) This is again EQ-> - 0 SWAP = or EQ-> SWAP - 0 SWAP = if you like it more. It is implemented as the command EQ-> Can't COLCT (COLLECT) do that? There are ALG48 and ERABLE at www.hpcalc.org. Perhaps you should try them, they are very very good. ==== I guess this is really the command I was looking for, thanx. I found out that the proper isolation command is QUAD, as ISOL does not allow the variable more than one time. ==== The header's right, but try inserting C$ $ (without the quotation marks) right after it to tell the calculator that the rest of the file is a character string object. See the 48G Series User's Guide page 27-17 for details of character translations, but this applies to characters 10,13 pairs, , and characters 128 through 255. There actually is a supported entry point that does the work for xGXOR internally. Namely, this is GROB+#. It has the following stack diagram: flag grob_target grob_source #x #y -> If flag is TRUE then an OR is performed otherwise an XOR is performed. GROB+# does all the necessary range checking so it isn't mandatory that the grobs be of the same size. ==== around In UserRPL you could do something like the following: IFERR XLIB_Name THEN Error_message ELSE continue... END In SysRPL it's similar: ERRSET XLIB_Name ERRTRAP :: Error_message ; Note that ERRSET/ERRTRAP each expect a single object as 'parameter', hence the DOCOLs ('::') and SEMIs (';') Additionally, ERRSET/ERRTRAP should always be used as a pair on the same program level. More info on this topic is in RPLMAN.DOC, which is available on www.hpcalc.org Raymond ==== On Wed, 13 Feb 2002 01:42:04 -0600, Jonathan Busby There is a small bug in this code that causes it to crash if the bankswitching routine portion of the ACPTR is zero. (ie. it's just a normal pointer encapsulated in an ACPTR. ) While this really doesn't have any bearing on your question since you're interested in covered ports I would feel kind of funny if leaving loose ends. ;) CODE * Unsupported entries valid on the 48G/GX only. POPACPTR EQU #702A2 RSTK=CRTN EQU #71603 GOSBVL POPACPTR GOSBVL =SAVPTR C=R0 ?C=0 A GOYES + R0=A P=P-1 GOSBVL RSTK=CRTN GOVLNG =PUSH#ALOOP + GOSBVL =GPPushA GOVLNG (=OSIZE)+5 ENDCODE ---------------------------------------------------------------------------- --- Jonathan Busby - before replying. ==== I have been wondering where is the key combination for the Greek symbol phi? I have found the symbol for theta (alpha, right shift, F), but I can not find the symbol phi. I must be blind! Any information that you can give me will be greatly appreciated. you for your time and have a nice day. ==== The key combinations are listed in the CHAR browser. I'm not sure if the character resembling a phi is actually a phi, but anyway you can obtain it through ALPHA, o, ALPHA, RIGHT SHIFT, 9. John Branthoover schrieb in im Newsbeitrag: qXSa8.31897$d34.1991368@bin8.nnrp.aus1.giganews.com... symbol can <3908649c.0201241141.44a5ca2@posting.google.com> <3908649c.0202080758.2bc0cf7b@posting.google.com> <3C66BB93.F5AB79C0@math.fu-berlin.de> ==== Well, that sequence isn't so easy to accomplish. First, the empty name on the command line doesn't get put on the stack when I press ENTER. So '' '' STO on the command line results in STO Error: Too Few Arguments. That provides some protection against fooling around with the empty name. Ok, by putting the empty string on the stack and changing it to a name, I get '' (or NIBBHEX 84E2000 in the SysRPL display mode, which tells me that it's a global name with a byte count of 0), and TYPE returns 6. for a global name. That's using built-in tools, but not what I'd consider pure UserRPL, and the user ought to be aware that he's straying outside of the supposedly safe UserRPL territory when he does such things. But even with the empty name in levels 1 and 2, pressing STO results in Error: Circular Reference. Ok, I can store the null name in a variable, and then, as long as I'm not in the home directory or a subdirectory that already contains a null name, I can rename that variable to the null name, so the protection isn't fool-proof. But I don't think that anyone would do this by accident, and a user who does such a silly thing intentionally should expect that strange things may happen as the result. In this case, the indefinite recursion didn't start until I executed the null-named variable, and even then, the CANCEL key was enough to break out of it. In this case, not even the stack was lost. :-) All of this with ROM 1.19-6. James <3908649c.0201241141.44a5ca2@posting.google.com> <3908649c.0202080758.2bc0cf7b@posting.google.com> 03:18:41 -0600) ==== Just fiddling around: 'A+1' 'A' STO 'A' ->NUM 49G: Error: Circular Reference 48G: Error: Insufficient Memory [r->] [OFF] ==== That's how I got one, without paying even shipping costs, but I've read that HP doesn't do this in all countries. Also, it occurs to me that HP might not even be making these cables anymore, so maybe they're just sending out the ones they have left in stock. Try to get a free one while you can. The worst that could happen is that they might say no. How long is the cable? Does the AUG include the other commands? James ==== On Thu, 14 Feb 2002 05:50:52 -0500, James M. Prange Anyone have the phone number for HP Customer Support? I don't have my manual handy. ==== I have a SysRPL routine where I store some reals as LAMs which are then used to produce results. For final output, I'm trying to then tag the results with the LAMs. I know that that TAGOB requires a string, but I'm not sure how to do this within the routine. I tried the following: { LAM f1 ID>$ LAM f2 ID>$ } which does compile without a hitch, but when the routine is run you get an Insufficient Memory error. How do I convert the LAMs to strings and then put the strings into the { } construct? My Donnelly book is not available to me right now and Kalinowski's book does not elaborate. Your help and advice is appreciated. Simon ==== if you're searching for a cheaper way to get a hp49g. i bought my from the german ebay-site www.ebay.de for about EUR 99,-. There are still some to buy. They're all new (unused) but with an italian users guide (but they're sending a self-burned cdrom with the english guides). The advanced users guide is in english. BUT these calcs have a serial number below 93... They're these ones made in Indonesia (which some people have reported problems to get the hp connected with the PC [I've no problems]). After that i bought the original HP-Connectivity-Kit from abyte.de (a german e-store too) for EUR 9,-. And everything's just working fine. So I paid for a HP49G and the PC-Connectivity-Kit 109 Euros (ca. 90-95 Dollars) plus ca. EUR 25,- (22 Dollars) shipping. HP RULEZ! ==== My 49's keyboard is acting up again. Now it presses keys all on its own!! Luckily it it the leftshiftkey so you get a nice effect as well; the annunciator is turning on and of quite periodically (experimenting with ->KEYTIME has On a more serious matter: sometimes it will not turn on. I have to paperclip it before it does anything. I think this is also keyboard problem. Man! I think I will go back to the 48(or my sliderule)! Never had any trouble with them. -- This message was written with 100% recycled electrons Pivo ==== Are you sure it's a hardware problem? Did yu try it w/o any special add-on software installed? Raymond Peter Geelhoed schrieb im Newsbeitrag ==== maybe too much tetris? Peter Geelhoed schrieb in im Newsbeitrag: 3C6D2E98.94665486@student.tudelft.nl... ==== Raymond Hellstern schreef: I haven't installed any new software lately; the last one was TETRI 2.1 and it would be a very good virus that would emulate a hardware problem that well! -- This message was written with 100% recycled electrons Pivo ==== cybernesto schreef: Definitely! I think I have carpal tunnel syndrome -- This message was written with 100% recycled electrons Pivo ==== Sorry, just my poorly translated Spanish copy. But, don't be desperate! A Dutch folk called Tuesday is scanning and _transcribing_ (whoa!!) some current HP manuals: http://www.geocities.com/tuesdayspuzzle 20S and 12C so far. What's next? Bye. HPCC member #1046 ==== Check Calcpro website www.calcpro.com. I have bought from them several times and have had good dealings with its manager, Paul Nelson. He sells the HP38G calculator and may be able to sell you a manual for it separately. Calcpro's website even lists a separate book for the HP38G that makes a good substitute for its manual. ==== Does anyone know a way (without writing a new program) to convert something like y=2x -1 to its polar form? Also, is there a way to convert a specific point from polar to rect and vice versa? Micah ==== Converting to polar form is not a single key press. In general the polar form can be found by: Substituting r*sin(theta) for y, substituting r*cos(theta) for x, and solving for r. You enter Y=2*X-1, then Y=R*SIN(theta) SUBST , then X=R*COS(theta) SUBST, then R SOLVE. Result: R=1/(2*COS(theta)-SIN(theta)) A (hypothetical) program for this: << 'Y=R*SIN(theta)' SUBST 'X=R*COS(theta)' SUBST 'R' SOLVE If with specific point you mean a complex number like (2.,3.) then you can simply switch to polar or spherical mode. If you have the point as an algebraic object like 2+3*i, then you must explicitely find radius and angle. The functions for this are ABS and ARG. Hope it helped, . ==== Is there a way to use HPComm and the link cable to obtain a printout from the 49G?? Cam ==== You can use Capture Image from the menu Calculator. Is that what you mean? ==== On hpcalc.org there are a couple of print capture programs that emulate the HP IrDa Printer through the serial cable. Each had their own feature and the following: 1) Isn't limited to one PC screenfull, ie capable of continuos capture. 2) Can select between printing directly to windows printer, text file or graphics file. 3) Capture both text and graphics. 4) Idiot proof, so I might just be able to use it. ==== Well, preferably do 3. TRANSIO first, and set up for an ASCII transfer. If you want to print something that's stored as a variable, you could use HPComm to copy the variable to a PC file, and print that file out from your PC. If you want to print something that's on the stack, you can set up HyperTerminal or a similar application to capture text and save it to a file, and then use the 49G's print commands (CR, PRST, PRSTC, PRVAR, or PR1) to print to your PC, and then print the captured text file from your PC. The calculator also includes a PRLCD command, but that assumes that you're printing to an HP 82240 printer, and would just make a mess without a lot of processing for your particular printer or graphics application. Of course, if you happen to have a serial printer available, you should be able to come up with a way to print directly to that, and won't need to use the PC. James ==== How can I keep the HP49G from rationalizing the denominator? Rom 19.6 ==== If you have a ratio, then press the key arrow down to take the whole thing in the EQW, and then select the numerator and/or denominator before you use any commands, numerator and denominator are changed separately from each other. ==== Are you using Algebraic mode? You should use RPN mode. When in RPN mode, you decide by yourself when an expression is to be simplified. When using EVAL or something similar, the HP49G will simplify with radicals in the numerator only - this is the most simple in the eyes of the '49. To keep radicals in the denominator, do INV EVAL INV in RPN mode. INV doesn't force further simplification, but merely swaps a fraction if the expression is one. An example: '(X^2*SQRT(X)+3*X*SQRT(X)-10*SQRT(X))/(X^2-2*X)' EVAL -> '(X+5)*SQRT(X)/X' While '(X^2*SQRT(X)+3*X*SQRT(X)-10*SQRT(X))/(X^2-2*X)' INV EVAL INV -> '(X+5)/SQRT(X)' You can't use this trick in algebraic mode, as the calc will always perform a operation similar to EVAL when entering stuff from the command line. ==== I made a program for Optimizacion y Simulacion Numerica that included Simplex, Lemke (or Simplex Dual) and Charnes (or large M) methods. The syntaxis that we use in Spain is a bit different P) Min (-3X1-X2-3X3) with 2X1+X2+X3 =< 2 X1+2X2+3X3 =< 5 2X1+2X2+X3 =< 6 X1,X2,X3 >= 0 ==>Simplex X1 X2 X3 X4 X5 X6 0 -3 -1 -3 0 0 0 X4 2 2 1 1 1 0 0 X5 5 1 2 3 0 1 0 X6 6 2 2 1 0 0 1 i.e. -z | c ------- vb | A sol: v=-27/5 Net at home, so maybe it will be posted next week (And I don't have instructions yet). The program is oriented to help doing the exams, so it makes only one iteration, you copy the result, and then makes another iteration. It«s User-RPL. ==== so Note that there are some nice and fast SysRPL implementations of Simplex already available at www.hpcalc.org. But that's not the point, the original poster asked for *integer* programming, and you can't do that with standard LP methods. -- 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 ==== Yes, the program has a post-optimal analysis that includes Branch-and Bound Methods. (Well, only the Dakin variation). I did this with just User-RPL (btw, what«s LP method, Linear Programming?), but the program does only one iteration (so you can copy the result in the exam). Oh, we study 2 B&B methods: -Original B&B -Dakin variation I didn«t implemented Original because because Dakin is easier and I was short of time (finished a week before the exam). My teacher is considering to ban HP's in the exam because I showed her the program. I«ll try to post next week. NaN5K. ==== Oh, but those programs don«t implement Lemke nor Charnes ==== Can anyone suggest any where that would have one of these cards new or second hand to sell. I cannot reach da Vinci, who produced this card (DV13300), and CalcPro were unable to supply one also. Lloyd lecutler@bigpond.com ==== www.samsoncables.com/catalog/prodByCat.cfm?category=Expansion%20Cards James ==== Oh yes, it is! The non-consistent behavior of this machine can drive someone nuts! I do that a lot! Especially when I want to take a look at the region where some expression evaluates to true. And I often have to try it again and again until I see something. I think because when there are only integers, then the CAS is involved, but when there are reals, then the expression is simply evaluated to a number. For example X^2+1>0 is examined by the CAS, which it finds that this is always true, when X is real, but X^2.+1.>0. is simply put current number for X and evaluate. Yes, too much inconvenience and questinable things about plotting. But especially for plotting RAND, there is a good reason for this behavior. The HP49G has operations, commands, functions and analytic functions. Only functions and analytic functions are allowed to appear in algebraic objects. RAND is a command, and so it can't be a part of an algebraic object. That means that the only way to plot that is to make a program that uses it, and returns only one number on the stack. A bit different, but quite powerful, because you can use *any* command in such a program and plot the program. So you can plot things that otherwise would be difficult/impossible to plot. Es Gr237esli, . ==== I need help I try for month to learn system RPL with my HP48G but I dont have much success, please give me how should I make a good beginning. ==== Check out http://cgi-bin.spaceports.com/~hpkb/tutorials/sys.php , http://members.tripod.com/ekalin and http://hpcalc.org . -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel ==== www.hpcalc.org. In particular, Eduardo Kalinowski's book on SysRPL programming is the definitive source on the programming language. The PDF version is available here: http://www.hpcalc.org/details.php?id=1758 Eduardo's book on SysRPL programming with the 49G should get you started with programming on the calc: http://www.hpcalc.org/details.php?id=4443 ==== Please, see the next address : http://members.tripod.com/ekalin/tutorials/ The HP Knowledge Base from Eduardo M. Kalinowski. Too you can see the Eric Rechlin Web page (www.hpcalc.org). Type in the search window ... Mika Heiskanen or Matias Mutchi or Joseph Horn (Goodies discs. 4 & 10). Best from Neuqu216n - ARGENTINA.[:Ā) Miguel Angel CAPORALINI HERK **************************************************************************** **** ==== I *do* want packed arrays :-) Let me explain why: My program tests for primality, I first create a VERY long HEX data where the N'th bit is 1/0 depending on whether N is prime/composite. So my data looks like (in binary) 01101010001010001010... [1,(2),(3),4,(5),6,(7),8,9,10,(11),...] I intend to include the first 1000 or more of these.. This could be further optimised by excluding the even numbers and treating 2 as a special case, or even multiples of 3 and treating 3 as special case, etc. but I am not that good yet to worry about the complications that would result, keep it simple for the moment :) This seems very great, I'll try it out! I knew about this one, but I have to calculate n, which could be 0,1,2,3 depending on the input N. I think one can read that from a register, X say, using CON(1) X, correct? but how is that done using MASD on the HP49G? PS. The real goal of this code is to give the N'th prime by table lookup, and also counting the number of primes below a certain limit.. I may have to include an even longer table, 100,000 probably?? Very helpful indeed! __ Kamel, who's gonna have great fun with Assembly over the weekend.. ==== I'll try that as well!! I guess it should work, but the trouble is how many times should I use ASRB, remainder of divistion by 4 probably!? Let's see how it'll go :)) __ Kamel, who's get to check for ASM entry point for division.. ==== Okay. Since there are fewer than 170 primes in 1..1000, and that a number of that range may be coded with 2.5 nibbles (maybe 3, to avoid bit manipulation), you're probably better off making a table that contains a list of primes only. Such a table would take (* 170 2.5 4) which is 1700 bits, versus 1000 bits for the bit-array solution. This solution seems rather simple to me, as it allows retrieval of the nth prime with few sub-nibble manipulations, but that's probably nothing more than a question of style, for this range of numbers. Since you seem to really want to use an array of bits, I won't discuss your choice. Methods for compressing prime number databases certainly exist already. When the order of the bit isn't known at assembly time, you must use self modifying code to generate the proper CBIT/ABIT instruction. Just kidding! Assuming you have a value from 0..3 in register A.P, an array of bits in C.P, and you want C.P to become null exactly if the Ath bit of C.P is null, you can do something like (untested and deeply unoptimized, assumes P==0): B=0.B B+1.B ; Bb=1 { B+B.B ; shift left Bb A-1.P UPNC ; until Ap is null } BSRB.B ; remove the extra shift C&B.P ; perform the logical and As there are only four cases to consider, using a lookup table may be wiser. Fewer than 10000 primes in 1..100000, that's 100000 bits with the array of bits representation, which isn't all that big, especially if the array is compressed with BZ (please let us know how well it compresses when you do it, I expect a high redundancy ratio). -- David Haguenauer ==== Assumming your bit table is stored at address BitTable (or that you have the table address in Ca) You want to know is the Aa th bit in the table is true or false. Here is a piece of code... just insert the code you want after the ?CBIT=0.3... LC(5) BitTable B=A.A BSRB.A BSRB.A C+B.A CD0EX C=DAT0.1 % read the interresting nibble % shift the nibble to get the interresting bit on the 3rd bit in the register ?ABIT=1.1 { C+C.P C+C.P } ?ABIT=1.0 { C+C.P } % perform the test ?CBIT=0.3 { % bit is set to 1 } SKELSE { % bit is set to 0 } Have fun, Cyrille ==== :-) Im looking for one (Advanced Users Guide) if possible in pdf or any other format downloadable..... Dont look at me like that !! Im a student living in a third world country. ==== Go to www.hpcalc.or. There you will find as much documentation as you like. Wish you good readings, . ==== Go to www.hpcalc.org. There you'll find as much documentation as you like. Wish you good readings, . ==== I've been there, thank you, can you maybe give me a link to documentation for basic UserRPL programming. I don't understand why HP don't give you the programming doc's with your programmable HP 4849 calc, or with the $40 other country. ==== Visit the following links for excelent resources on UserRPL programming. The first is a great tutorial on pdf format by Eduardo Kalinowski and the second one covers some advanced topics like loop structures, branches, output methods, etc. The guide is from Justing Dossey and it is very well written. http://cgi-bin.spaceports.com/~hpkb/tutorials/usr.php http://cgi-bin.spaceports.com/~hpkb/tutorials/aurpl.php Enjoy! Diego Terzano ==== I heard of a Basic like interpreter - or compiler, note sure - for HP 49G. Was it in a dream or does it really exist ? Herve from Paris, France ==== Well, I wouldn't use the word dream as much as nightmare but the 49 has a programming language for algebraic mode that is called HPBasic I believe. Maybe that is what you heard? -Al -- Place your hand on a hot stove for a minute and it seems like an hour. Sit with a pretty girl for an hour and it seems like a minute. That's relativity. -Albert Einstein ==== It must have been a nightmare! Who would want to give up the RPL wonder? Anyway, AFAIK the Basic like interpreter refers to the 'algebraic' version of the UserRPL that is limited in scope and slooow! You wouldn't want to use it, believe me. Learn UserRPL along with a few SYSEVALS and there is no limit but the imagination! By the way is there going to be a new ROM in the near future? I am glad this newsgroup still insists on persisting! !Demeter! ==== Herve 871yf45t0e.fsf@austin.rr.com... 49G. ==== Herve 5bff2c2e.0203010531.66f790dd@posting.google.com... 49G. ==== Well, I think I have quite a good idea of how it works, what would you like to know exactly? Cyrille de Br216bisson ==== No longer published. :-( http://library.hp41.org/HP41/LibView.cfm?Command=List&Class=HP41&CategoryID= 7 You will probably find it more convenient to purchase a handy CD set with all that stuff and much more from Jake Schwartz: [BTW, did you know that Jake was mentioned in the Acknowledgments section of the January 1981 issue?] http://www.waterw.com/~jakes/ppccdrom.htm http://www.hpmuseum.org/cd/cddesc.htm http://www.hpcc.org/intro.html http://www.hpcc.org/membership.html Don't miss the student's rate and the Special Offer. The regular publications are Datafile (six times per year) and a valuable http://www.hpcc.org/datafile.html I hope to have been kind enough! Bye. HPCC member #1046 - ==== The HPCC still exists and is active. It is a Users club based out of London, England. But anyone worlwide may join. I believe you can sign up to get a sample journal for some nominal fee and then if you want to join just upgrade to a full year. The HPCC publishes their journal, called Datafile, 6 times per year. See their website for full details: www.hpcc.org Steve ==== I'm having a little trouble getting the 49G to display real answers (not imaginary) for these integrals. Can anyone show me how I could maybe manipulate the expression to make it work? int x^2*e^(2x) dx int x^3*sin(x) dx J. Yuan ==== J. Yuan schrieb im Newsbeitrag First make sure you are using the newest ROM. The current version is 1.19-6. Then configure your CAS by executing CASCFG. You can use the the entered expression (in EQW) and use the command INTVX to find the indefinite integral with respect to the current variable (should be X). Hope this helps Roman ==== That's exactly what I did, and the answer had numerous i in it. I believe this problem may be integration by parts. 1.19-6. the ==== My calc does it just fine. Make sure you're not in complex mode. If you use INTVX() you just want to enter the x^2*e^2*x without telling it what variable because it uses the default. I get: ((2*x^2-2*x+1)e^(2*x))/4 Micah ==== May I inquire on your flag setting? Mine still does not work. I a really complex answer with is in it. ==== Now that the library is at HOME go to the HOME directory , press the key VAR and then the menu key for the library. Note the number (ID) that the library has, because you will need it later. Enter the port number that you would like to have the library (0,1, or 2) and then press STO. Now enter the library ID and enter ATTACH. Hope it helped, . P.S. What do you mean with *.000 files I have never seen such a thing. ==== I don't like to start rumors.. but it seems to be true. I've read sometime ago (15 days more or less) about HP resuming calculators designing with a hp41 remake. I read that in: http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/forum.cgi?read=15299 It seems to be a 4th April's early fool's day..just it seems.. But today, our Spanish calculator provider confirmed me that new models are expected, and the info comes from European HP Manager.. it seems that these models are desinged from a new Corvallis team.. Have you heard something? J.Manrique CdU de la ETSIG #1077 HPCC Member ==== i hope NOT. i'd like to see an OLD calculator made, if you get the idea... -0600) ==== HP is always making new calculators; they are making the old ones with new colors (like TI :) and with cheaper assemblies and parts. When the 49G is long gone, with the 12C still be around? (kiddie tastes change rapidly, but boardroom taste is eternal :) [r->] [OFF] ==== Maybe it's a stupid question, but can anyone tell me how to change fo example 0.3333333 to 1/3. And how to get result like Sqrt(2) instead of 1.412... in eqation writer? ==== The answer you are looking for is XQ..... and please: read the manual. Dmitry escribi227 en el mensaje ==== You want to use XQ. Inside the EQW select what you want to apply XQ to, then press the CAT key and choose XQ in the list. If you prefer to type XQ, select the expression that you want to convert, use the |EDIT| key and append XQ to the command line. -- David Haguenauer ==== Or turn off approximate results in the options? Or is that something different? Roman ==== R Lion escribi227 en el mensaje Perhaps it sounds hard. I did not want be rude. ==== Can you please specify if you have the HP48 or the HP49? I could be able to help you better then. ==== Oups ! Sorry... HP49G, running V 1.18. Herve Karagiaouroglou a 216crit dans le message de WINDOW at copy function to 2. simul8.demon.co.uk:212.228.120.103 ==== I've reported this bug twice now on this newsgroup and once on the HP49 bugzilla site. The bug shows up if the xres value in ppar is set to something other than 0. The solution is to do 0. RES before starting up the plotting app. Search back on www.deja.com for previous mentions of this problem. James ==== Just bought a 48G+ a week ago and I am a little disappointed in the quality of today's HP's, painted keys, discoloration on the screen. Everything works great of course, but it just doesn't feel as sturdy as my previous calculators. It's made in Indonesia in 2000. Just wondering how long would the new keys last, I've owned a 28S for 12 years (of course it has molded keys) and a HP 15C for about 15 years both of them still work and the keyboard's like new. Any problems with your 48G? How long does the paint last, does it peel off after time? ==== Well, the painted HP48 keys have been tested for a million key pressed (using a rubber 'finger') and a million 'rub' using the rubber finger to rub the key... ie: the paint is supposed to hold tight, and if it's going away, you should send the calculator back for warranty... Cyrille de Brebisson both ==== My 48g+ is still painted well, and I've bought it 2 years ago... Sorry for my English,I'm Italian... 2c78c986.0203010310.1a950c2f@posting.google.com... quality works of ==== ==== Yes, you can find everything at www.hpcalc.org you should read the documentation accompaning the game -- This message was written with 100% recycled electrons Pivo ==== Of course! Go to www.hpcalc.org . Installation instructions are (nearly) always packed together with the files. Have fun, . ==== Is there a list of must have programs for 49G? The one one HPCalc.org (Calculus programs, integral database would be helpful) and science apps. J. Yuan ==== Did you use that name when registering your beautiful HP-45? Oh man, you would love to live where I am, in Barcelona: We've got a beautiful White Male Gorilla ... but we keep it in a cage ... :-) http://www.zoobarcelona.com/camara/English-sony.htm Bye. HPCC member #1046 - -- ==== Anybody know how many time the calculator is capable to be on or runing a program, i do this question because my hp49 have the problem that if i turn it off for a long time (an hour or more)she die and i lose my memory. So i want to mantein a program runing to prevent this ==== That shouldn't be happening. I didn't touch my 49 for 2 weeks over christmas and didn't lose a thing. Flat batteries could cause this but then calculator wouldn't turn on. I use Ni-Cad rechargables which give very little warning that they are going flat. So I've been caught a couple of times where they have gone flat and I've left them over the weekend so the Calc is dead on Monday. So far (fingers crossed) every time I've put charged batteries back in the HP49 has come back to life without losing a thing. Because I'm using Ni-cads I backup to flash memory regularly just in case. I find the memory pretty reliable. I've lost User memory a couple of times due to doing silly things, lost Port 1 once but so far flash memory has been bulletproof. Maybe you have a bad library, I suggest you make sure your using the latest ROM 1.19-6 and start the calc holding the backspace down so the libraries don't load. If it still happens it maybe hardware and you might have to return it to HP for repair. 19:27:42 -0600) ==== Good choice! Welcome to comp.sys.hp48 as well. The minimally sufficient editor could be Notepad, and newsgroup post composers); this is made possible by HP having embedded an escape sequence mechanism for representing character numbers 128-255 in ascii, by means of 3-character backslash sequences. For example, << and >> represent what HP calls Program delimiter characters, -> represents a right arrow, etc. There are only a few more very commonly used programming symbols within that range, all of which are listed here: http://groups.google.com/groups?selm=8013hf%24apm%241%40nnrp1.deja.com When you transfer ascii files to and from the calculator, using HP's built-in Kermit commands (not Xmodem), the required decompilation, re-compilation, and funny character translation of ascii text are all automatically and transparently invoked; type 3 TRANSIO [or TRANSIO(3) in Algebraic mode] to inform your calculator that you'd like all symbols to thus be translated, and henceforth your calc will confine itself to using only pure ascii in all of its external conversations. Fancier special-purpose PC development system programs are also available, which display and compose using all the identical 1-character symbols of the international character set ISO-8859-1, which the calculator uses itself, but since you may encounter many more pure ascii programs in newsgroup posts and elsewhere, a faint familiarity with the pure ascii system may also be helpful, and also is the only system totally portable across all computer platforms. As to what compiler, there is no better compiler than the one which is built into the calculator itself, and since that compiler is automatically invoked whenever you transfer ascii source files to the calculator, there is no need to have an external compiler on a PC. Likewise, when you export user programs from the calculator itself, if you have set the ascii mode flag (and use the built-in Kermit SEND command), the built-in decompiler automatically exports editable plain text, rather than a binary executable image. All the same, there is another PC development platform which can be used to compile, decompile, and to test all programs, just as if you were using your calculator itself. What's that? Why, it's the calculator emulator! (Emu48/49) General locations (do not omit www.): http://www.hpcalc.org/hp48/pc/emulators/ http://www.hpcalc.org/hp49/pc/emulators/ Basic files (current as of this posting): http://www.hpcalc.org/hp48/pc/emulators/emu48-1.25.zip http://www.hpcalc.org/hp48/pc/emulators/e48sp29.zip Additional files for HP49G: http://www.hpcalc.org/hp48/pc/emulators/hp49kml.zip http://www.hpcalc.org/hp49/pc/emulators/upd49rom.zip http://www.epita.fr/~avenar_j/hp/49.html [read all info, get all files, including save bank 0 program for updating older versions to 1.19-5 and later] Additional files for HP48GX: http://www.hpcalc.org/hp48/pc/emulators/gxrom-r.zip http://www.hpcalc.org/hp48/pc/emulators/mkshared.zip In the very first link that I included above (the symbol list) can be found a couple of one-line calc programs which can do the symbol translations independently of Kermit send/receive; I use these to exchange my plaintext user programs between my PC editor and my emulated calc, via the clipboard (and the Copy/Paste string functions of the emulator). Finally, the emulated calculator can actually send and receive (using a PC serial port and cable) to a real calculator, which not only takes care of getting programs finally into your calc, but can even do a ROM upgrade (ROMUPLOAD command) to your real calc! ------- BTW, here is the USA CIA's view of your country (I hope that they have an occasional fact straight :) http://www.cia.gov/cia/publications/factbook/geos/cg.html An excerpt: Infant mortality rate: 99.88 deaths/1,000 live births (2001 est.) Life expectancy at birth: total population: 48.94 years So if I lived there, I should, on average, be dead :) The means by which we live has outdistanced the ends for which we live. Our scientific power has outrun our spiritual power. We have guided missiles and misguided men. - Martin Luther King Jr Strength to Love, 1963 The world needs plenty more human consciousness to emerge into peace: http://www.mum.edu [r->] [OFF] ==== John ! thankyou !!! ==== keys. feeling. So, i can assume that *all* the GXs with black screen have painted keys, while *all* the GXs with blue screen have molded keys, regardless year and place of production. Am i right? -- Marco Polo marco.polo@katai.cina --- ==== Yes. No. AFAIK HP changed to painted keys for most models some years ago, while the new display seems to be used only for a year or alike. Raymond ==== As I said there are many ways to do things on the HP49. I find lists one of it's handiest features. I used to use a LASTX program for repeated operations like my HP41's LASTX command but its now dissapeared of my CST menu and been replaced with quick access to the List manipulation commands. I'm even thinking of changeing the default behaviour of the + key to add lists instead of concetating them. it ==== As I said there are many ways to do things on the HP49. I find lists one of it's handiest features. I used to use a LASTX program for repeated operations like my HP41's LASTX command but its now dissapeared of my CST menu and been replaced with quick access to the List manipulation commands. I'm even thinking of changeing the default behaviour of the + key to add lists instead of concetating them. it <22eb8.75071$3J1.2305675@news2.tin.it> 02:25:07 -0600) ==== How about: 'A+B' STEQ 30 MENU Note: *no* equals sign -- just an *expression* Now you see a label key for each variable; to store each value, type a value and press its label key. Finally, press EXPR= to calculate the entire expression. Note that this allows you to enter values in whatever order you like; it allows you to keep reusing the same values for A or B without being required to re-type the identical values every time; it doesn't waste your time forcing you to repeat wasted motions or duplicate entries. If you ever forget what values you had entered, and would like to *review* what is currently stored in each variable, just press the right-shift key, followed by the cursor-down key, and the screen displays all the values at once! Is this handy for you? Or would you rather use a full-screen Input form and navigate all around to the various fields using cursor keys, and have to press EDIT to view the complete values, one at a time, in fields that are too small to see? If you like the method above, please sign my help stamp out ridiculous GUIs petition, and forward it to folks like Microsoft, Symantec, ... If you think you might have trouble remembering the above few commands, and that mysterious right-shift, cursor-down key sequence for showing the values of all variables, here's a much more automated 49G program you can permanently store, to evaluate arbitrary expressions: << LNAME AXL { { Show << -35.3 KEYEVAL >> } } + 2 ->LIST STEQ 30 MENU >> 'MAKE' STO Now, just supply an expression, e.g. 'A+B' and then press MAKE; your menu now immediately appears, and within your menu is an extra Show key for reviewing variables. Owners of the good old 48G series can use the following instead, taking care to enter the hex values exactly as shown, not confusing SYSEVAL with LIBEVAL, and making a memory backup first anyway, of course: << DUP #353ABh SYSEVAL { { Show << #A3010h LIBEVAL >> } } + 2 ->LIST STEQ 30 MENU >> 'MAKE' STO [r->] [OFF] ==== WOW WONDERFUL!! It was what I was looking for: this way has all the features I need, it's very very fast, you can review the valued, etc. I agree with you :-) and, of course, I am ready to sign the petition... I have tried also this method, but it seems that I make some errors in writing it on the calculator. How can you make on the calculator the sign? And is the -35.3 a simple minus (like the one you use make simple operations) or I have to use the key '+/-' that produces 'NEG'? When I tried it without the sign the program worked, but the keys for the variables did not appear, just the 'SHOW' key and the 'EXPR=' key. What is wrong? I am pretty sure that I do not enter the code with the right sintax..... dix ps. I am starting now using the RPN notation, instead of the Algebric, is there any document that explains this mode of work? The manual I have from HP pays only attention to the algebric one..... ==== The fastest way to solve your data entry problem is with the use of lists. I'll assume you're in RPN mode, if not, I'll suggest you switch fow now, it will make things easier an faster. type this program and store it in a variable (ie. MYEQ) << -> A B << A B SQ ADD >> >> MYEQ STO where << >> = program delimiters (RightShift +) the forward arrow is RightShift 0 To use the program, simply put the data for A and B in the stack as lists and run the program. if A = { 1,2,3,...} and B= { 4,5,6,.. } you would enter on the stack: {1 2 3} {4 5 6} MYEQ the output will be in list form also, like {17 27 39} Hope this helps, Diego.- ==== You're NOT supposed to write the sign. It is used to represent ASCII characters not present on the PC. When you transfer programs from the PC to the calc those codes with the characters are correctly identified by the calculator. You can also use nnn, where nnn is the ASCII number in the calculator. If you do that, you must include the following header on the file: %%HP: T(3)A(R)F(.); Hope it helps, Diego.- ==== to the to understand the explanations: I am a real beginner :-(( without problems, I don't know how to enter the other program that John suggested: << LNAME AXL { { Show << -35.3 KEYEVAL >> } } + 2 ->LIST STEQ 30 MENU >> 'MAKE' STO I don't know how to type it and make it works on my HP 49. dix ==== Try StatPack49, a statistics program from Tim Wessman http://www.hpcalc.org/hp49/math/statistic/statpack49.zip it has some useful commands to deal with CDF and PDF's in several distributions, including the Normal. Hope it helps, Diego ==== Someone said get StatPack49, which you can do, but the HP49G has normal distribution statistics commands built in. Go to MTH -> PROB -> (NXT). The relevant command is UTPN (upper tail normal distribution). Look up NDIST also, that may be useful. -- Private First Class, United States Marine Corps ==== The UTPN command is a command, not a function. To make a function to do what you want, create the following program. << -> x << 0 1 x UTPN >> >> and store it in N. then N(x) will be the upper tail probablity of the random variable x on a normal curve, assuming a mean of 0 and a variance of 1. (Standard Table form) You can hard code any other values, or you can create global variables to contain these values for a series of evaluations. examples: 'UTPNF(0)' EVAL 0.5 'UTPNF(0.5) EVAL 0.1587 You can use this trick to create a function from any HP48 Command. Hope this helps, Bill wtstorey@ieee.org Bill alternate E-dress wtstorey@ieee.org.no.spam.please (Use the obvious) ==== R Lion escribi227 en el mensaje ==== Albert Graef a 216crit : First thanks for grw. Hope isn't too late to report this: I find a bug in the zoom feature, key +. Easy to reproduce: start grw with default configuration file (or not if first time) and empty stack move a little the cursor on the right ( 1cm ) now press + to zoom press + to zoom again a third time now.. no zoom anymore (first question, is that a limitation ?) BUT, if you move the cursor key down you will see many cross on the screen :(( note that If you have a graph the problem is the same, and sometime you can lose the cursor... - Denis P.S i use grw 1.2 and rom 1.19-6. ==== 1:{{0}{90}{50}} 2:{{'0.5+INV(S*1.8)+S*0.02' '-INV(S*1.8)-S*0.02' -1}{'-INV(S*1.8)-S*0.02' '+INV(S*1.8)+S*0.05' 0}{1 0 0}} ADIV Bad Argument Value! WHY???? :( Emanuele ==== Becuse Alg48 only uses rational numbers: you must write 1/2, not 0.5 Use ->Q or QPI with your arguments before an Alg48 command Emanuele Zattin escribi227 en el mensaje ==== Well, here I am answering my own question. Because the LAMs I wanted to tag to the results were reals, I ended up using the following code: result code LAM f1 %TAG This worked perfectly! Simon ==== What about connecting a hp48 gx with a notebook, via irda port? Is it possible? I have win xp... Ideas, thanks in advance ==== Is there any way to purge user created units on the 48G. I seem to have accidentally created some user defined units in my area directory and don't know how to get rid of them. I course I tried the obvious (to me anyway) ways to purge something analogous to a variable, but of course they didn't work. Any one else had this problem? I couldn't find the removal in ==== I know that 49G does not simplify expressions like the TI 89 does. And that's fine for most of the time (the simp command does not really simplify that well). Is there a program that can make the 49G have similar simplification methods as the 89? I'm in Calculus, and sometimes it just helps to check answer faster if the 49g simplifies. ==== Simplification is done differently on the 49G. It gives the user the power to manipulate the expression any way he likes. With the right combination of CAS commands you can reduce the expression to look exactly like the Ti. The following program tries to accomplish that. It's called AutoSimplification. http://www.hpcalc.org/hp49/math/misc/asv11.zip Hope it Helps. Diego ==== you're short on time and patience... I will try that program. Is there any where where I can get a printout of the main CAS commands? J. Yuan ==== any Look here: http://www.hpcalc.org/details.php?id=4562 (last time I checked, the pdf was not very readable, but you can pdflatex it yourself for a better result). And then there's the AUG, of course. -- 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 seem to have missplaced my manuals and can not remember how to do it. I guess 3 years without doing it is quite a long time.... Mark R. ==== you can download the manuals from www.hpcalc.org Short trip: Library on stack level 1, enter the port number, press STO, then do a warmstart (ON+C). After that, the library should show up in the library menu (rshift+LIBRARY) Raymond markvr schrieb im Newsbeitrag I ==== Mark R. * My 8 year old HP48GX is still a very useful tool (rshift+LIBRARY) <3053eba9.0202161800.7d4c45c1@posting.google.com> ==== Oh, ok. The B is for barn. 1_b=10^(-28)_m^2. It's built-in, not something that you've added. Wolfgang's UNITMAN is good for getting rid of these annoying metric units. See www.hpcalc.org/details.php?id=4318. James ==== James M. Prange schreef: Great unit, especially nuclear physicist puns: You can't even hit the side of a Barn Grrrrrrrrrrr........ -- This message was written with 100% recycled electrons Pivo ==== That's odd. I see that you are correct. I just don't ever recall seeing this ==== What is pdflatex? better ==== You know TeX, D.E. Knuths typesetting system, right? pdftex is a tex version with which you can create pdf documents directly from TeX documents; pdflatex is just the version to be used for LaTeX documents. (IIRC, I'm not complete TeX distribution for Windows, MIKTeX, but I haven't taken a look at -- 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 ==== Would you happen to know where I can find the software you mentioned for Windows, or a informational website on them? I am not very familiar with them, but need to look them up. version not at ==== Does anybody knows a source for programs for the FE/PE (engineer exams)? I'm specially interested in anything that has to do with the Chemical Engineering PE exam. Javier A. Gomez-Durand javier_gomez@engineer.com San Juan, Puerto Rico <2953da04.0201260148.4ccbb3d1@posting.google.com> <3C528F83.DAB18C1B@math.fu-berlin.de> <2953da04.0201262049.7c911ae4@posting.google.com> <3C53E07B.F1922230@math.fu-berlin.de> <3C55E2A5.CBEAA063@miu.edu> <3C63959B.E8741920@math.fu-berlin.de> <3C63E1A1.C56263FB@math.fu-berlin.de> ==== PS: I'm sure that Wolfgang's use of #DFFh DOERR (or the SysRPL equivalent) suits his purpose perfectly. For a more general solution, I came up with the following. It works with or without an existing STARTERR. If you don't want dire results, double-check that the value for SYSEVAL is identical (or use BYTES) before you run it. I suppose that if you double click the CANCEL key or do a warmstart, there's an extremely remote possibility that an extra variable named STARTERR could be left in the current directory; just PURGE it if you manage to make that happen. I think this is obvious, but I'll mention that it's written for a 49G; it should be harmless (except for creating STARTERR variables) but useless on a 48. Program name: QUIET Checksum: #62Bh Size (including name) 143.5 Size (not including name): 134 %%HP: T(3)A(R)F(.); << IFERR EVAL THEN ERRN B->R { ERRN DOERR } << << 'STARTERR' PURGE ERR0 >> 'STARTERR' # 8696h SYSEVAL @ CREATE 34. DOERR >> IFTE END >> To illustrate its use, first execute QUIET both with an empty stack and with a number on level one. Next execute the following two programs. << << 0. DOERR >> QUIET >> << << 1. DOERR >> QUIET >> Finally, execute the following program but press CANCEL while the loop is still running. << << 1. 1000. FOR n n NEXT >> QUIET >> I expect that you've gotten the idea by now. Of course, if someone thinks that it's worth translating it into SysRPL, and posting that, it won't bother me at all. Better yet would be a library that makes the CANCEL key and 0 DOERR behave just as they did in the 48 series. Is such a thing possible? James <2953da04.0201260148.4ccbb3d1@posting.google.com> <3C528F83.DAB18C1B@math.fu-berlin.de> <2953da04.0201262049.7c911ae4@posting.google.com> <3C53E07B.F1922230@math.fu-berlin.de> <3C55E2A5.CBEAA063@miu.edu> <3C63959B.E8741920@math.fu-berlin.de> <3C63E1A1.C56263FB@math.fu-berlin.de> ==== PPS: Of course, this accomplishes the same objective, and it's a lot shorter, simply, and presumably faster. Program name: Quiet Checksum: #700Ah Size (including name) 70 Size (not including name): 60.5 %%HP: T(3)A(R)F(.); << IFERR EVAL THEN ERRN B->R { ERRN DOERR } { 3583. DOERR } IFTE END >> James ==== BTW, the truth plot works fine on the 48G. Now I'm curious. I've got the printed manual and haven't found any typos on pp. 23-14 through 23-17. What is the typo? I'm curious if its in the printed manual as well. Dan ==== J.C. Randerson escribi227 en el mensaje Timit216 Hassan and you, have wasted many words with the TI fanboy.... a pity. Let's talk about interesting things: Can you explain of the above and perhaps share it? ==== pity. I would be interested in this too.. Could it be taken to another post so it will not get lost in this garbage? danny 89 owner, 49 user ==== It is the tenth and last part of the Trigonometry Marathon and we already have seen a lot of things. But there is still stuff waiting to be discovered. Would you ever think that the little HP49G is such a big place if you take a look from the inside? Until now we stayed in the real domain. Today we will dare a small jump into the complex. (As if it werenĪt complex enough already.. ) and here we go, our trip into the complex begins. You may already know that there are some relations between trigonometric functions and complex exponentials. Since the Complex Marathon starts right after the end of the Trigonometry Marathon, I think it is better to leave the derivation of these relations on the HP49G for the first part of the Complex Marathon. For now it is enough to show what can be done with these relations: 1) e^(i*x)=cos(x)+i*sin(x) 2) e^(-i*x)=cos(x)-i*sin(x) First of all, the HP49G can do this. The command SINCOS takes complex exponentials and returns them as trigonometric functions. It is the first command on the second page of the TRIG menu. We assume here that X and Y are real and also that Z is complex. Enter X ADDTOREAL then Y ADDTOREAL and then Z UNASSUME so that these assumptions are done. Also switch the HP49G to complex mode. Now, enter e^(i*x+Pi/2), press SINCOS, and you get e^(Pi/3)*(cos(x)+i*sin(x)). Enter e^(X+i*Y), press SINCOS, and you get e^(X)*(cos(Y)+i*sin(Y)). But enter e^(i*Z), press SINCOS, and the result is e^(-IM(Z))*(COS(RE(Z))+i*SIN(RE(Z))). Why the difference? Well, X and Y are assumed to be real, so the HP49G knows for example that the real part of X is X and the imaginary part of X is 0. But if Z is complex, and IM(Z) unevaluated, to denote the real and imaginary part of Z. If we add the relations (1) and (2) we get: 3) cos(x)=e^(i*X)/2+e^(-i*X)/2 If we subtract (2) from (1) we get: 4) sin(x)=i*e^(-i*X)/2-i*e^(i*X)/2 The command for converting trigonometric functions to complex exponentials is EXPLN. It is the first command on the menu EXP&LN. Enter SIN(X), press EXPLN, and you get the result: (exp(i*X)-1/(exp(i*X)))/(2*i) Though it is already readable enough on the HP49G, letĪs make it looking more familiar. Press LIN (third command on the menu EXP&LN) to get i*e^(-i*X)/2-i*e^(i*X)/2. LIN tries to make real or complex exponentials linear. ( EXPAND doesnĪt fit here, because it brings the two terms 1/e^(i*X) and e^(i*X) over a common denominator and so returns a more complex looking expression. ) Now, with i*e^(-i*X)/2-i*e^(i*X)/2 press DUPDUP to make two copies and then press RE to get the real part of the expression. Press EXPAND or COLLECT and you see SIN(X) again. This is correct, because we started with a real thing, that is SIN(X), and so even turning it to a complex exponential, it still remains real. YouĪll see how important this can be later on, in this part. Press now SWAP to bring one of the copies that you have made on stack level one and press IM to get the imaginary part of the expression. Press EXPAND and you see that the imaginary part is 0 as it must be. ( Since we started with the real SIN(X), we expect the HP49G to return 0. ) Instead of using RE and IM, you can also use SINCOS. Press ROT to bring the second copy on stack level 1, Press SINCOS and the EXPAND. The result is again SIN(X). If you enter SIN(Z) and press EXPLN, then you get (exp(i*Z)-1/(exp(i*Z)))/(2*i). Perhaps you wonder why there are no RE(Z) and IM(Z) in this case. Well, if there were such expressions, they would appear as (exp(i*(RE(Z)+i*IM(Z)))-1/(exp(i*(RE(Z)+i*IM(Z)))))/(2*i). That is, only in a form that is equivalent to Z itself, because every complex number Z is the same as RE(Z)+i*IM(Z). Not only the trigonometric functions can be converted to complex exponentials/logarithms but also the inverse trigonometric functions. Enter for example ACOS(X) and press EXPLN. The result is LN(e^(LN(X^2-1)/2+X))/i. If you donĪt like the representation e^LN(arg), then press EXP2POW to convert this to LN(X+SQRT(X^2-1))/i. There are some things that should be mentioned about this result. The first is, that the HP49G canĪt get REal and IMaginary parts of this expression. So if we want to do that, we must do something ourselves. Though this will be covered better at the complex marathon, let it be said here, that the argument of the LN can be written as a complex number of the form r*e^(i*theta), where r is the magnitute and theta the angle of the complex number. Thus we have LN(r*e^(i*theta))=LN(r)+i*theta. The REal and IMaginary parts of this are easy to calculate (if we assume that theta is the angle of the principal value). The second thing is that the HP49G doesnĪt consider assumptions about variables when it evaluates or expands LN(X+SQRT(X^2-1)). If you make the assumption X>=1, and then try to find the REal part of this expression, the HP49G doesnĪt return the expression itself, but simply doesnĪt return 0 but IM(LN(X+SQRT(X^2-1))). If you have only X+SQRT(X^2-1) and have made the assumption X>=1, then the HP49G returns X+SQRT(X^2-1) as the REal part and 0 as the IMaginary part. The third thing is that, if you start at complex mode you get LN(e^(LN(X^2-1)/2+X))/i no matter what assumptions you have made for X. But if you are at real mode, enter ASIN(X) and press EXPLN, then you get a huge expression containing REal and IMaginary parts, like for example (e^(RE(LN(X^2-1))/2))^2. If start at real mode and you have previously assumed that for example X>=1 then these assumptions *are* taken into consideration . I think that the HP49G in complex mode considers expressions like for example LN(X^2-1) to be general complex expressions and so it doesnĪt care to show explicitely what is the real or the imaginary part. But in real mode, it explicitely shows real and imaginary parts, as good as it CAS allows and tries to return results according to the assumptions you have made. A bit more on assumptions. Lets say that you make the assumpion X>=1. Then things like IM(SQRT(X^2-1)) are correctly evaluated to 0. Does this means that the HP49G considers this assumption? Well, it does in many other cases, but not for this one, though it looks like it did. Lets assume -<=X<=1 and find then the imaginary part of SQRT(X^2-1). How can we make the assumption X>=-1 AND X<=1? Entering this expression and then using ASSUME, results in an error. But you can enter X>=-1 ASSUME and then X<=1 ASSUME. If you now take a look at the list REALASSUME, you see that X>=-1 AND X<=1 is in the list. Quite hard to understand why the HP49G doesnĪt let you do it directly with X>=-1 AND X<=1 ASSUME and wants you to use X>=-1 ASSUME and then X<=1 ASSUME instead. But it has its reasons. If you do X<=2 ASSUME and then X<=1 ASSUME, then the finds out that X<=2 AND X<=1 is equivalent to X<=1! And this though X<=2 AND X<=1 *canĪt* be simplified with EXPAND or EVAL on the stack! Could it be that this is the built-in back door for simplifying logical expressions? (And also another marathon? ) Back to our imaginary part of SQRT(X^2-1). With the assumption X>=-1 AND X<=1 the expression IM(SQRT(X^2-1)) should be evaluated to SQRT(X^2-1) and RE(SQRT(X^2-1)) should be evaluated to 0. But it doesnĪt! If you make this assumption and enter SQRT(X^2-1) RE then the result is SQRT(X^2-1) and the result of SQRT(X^2-1) IM is 0, which is not correct, considering that -1<=X<=1. More about assumptions will be in the complex marathon. The forth thing is that the expression LN(X+SQRT(X^2-1))/i cannot be reconverted to ACOS(X) using SINCOS. So it looks like an one way ticket from inverse trigonometric functions to complex logarithms. LetĪs try to find if and how the conversion from LN to inverse trigonometric functions can be made. LetĪs say we have LN(Z) and want to convert it to ASIN(W), where W is some function of Y. We want to find what W looks like. So enter LN(Z)=ASIN(W) and solve this for W. The result is W=SIN(LN(Z)). Does this mean that whenever you have LN(Z) you can convert it to ASIN(SIN(LN(Z))? Well, unfortunately not exactly. The reason is the ASIN which can send one argument to more than one results. Because of this property, this function (and all other inverse trigonometric functions) are programmed so that they return the *principal* value of all different possible values, which goes from -Pi/2 to Pi/2. ASIN(SIN(X)) *not* necessarily equal to X. To understand this better, do the following: Define F(X)=LN(X)-ASIN(SIN(LN(X)). Then enter .5 and press F. The result is 0, which shows that LN(X)=ASIN(SIN(LN(X)). The function returns always 0 for arguments between 0.20787957635 and 4.81047738099. But try to calculate F(5.) and suddenly you get 0.07728317126. What is going on here? What are the strange numbers 0.20787957635 and 4.81047738099? I got a real headache thinking about the reason and was about to throw his HP49G away. ThatĪs why the 10th part of the trigonometry marathon had such a delay. But then Trabakoulas came and helped again. He told me to make plot of the SIN(LN(Z)) against LN(Z) to understand why the HP49G behaves this way. When LN(Z) is less than Pi/2 and greater than -Pi/2 , then a given value for the sine is sent by ASIN to the principal value Xo. (You start at the Y-Axis at Y, go horizontally until you meet the sine curve at point o, then go vertically to the X-Axis until you meet the point Xo.) SIN(LN(Z)) | Y +---o * | * | * |* | * --------------+------ * *| Xo=LN(Z) * * | * * | But when LN(Z) is for example greater than Pi/2 then the SIN function sends X to SIN(X)=Y,(from X go up until you meet the SIN curve and then to the left until you meet Y=SIN(X) at the Y-Axis ) but ASIN until you meet the curve SIN(X) and then down until you reach Xo.) We see that the arguments that 263belong to the principal values are those between -Pi/2 and Pi/2. But in this case the argument of SIN is not X but LN(X). That means that these arguments LN(X) go from -Pi/2 to Pi/2. And that means that X itself goes from e^(-Pi/2)=0.20787957635 to e^(Pi/2)=4.81047738099. SIN(LN(Z)) | | * * Y+-*-----o |*| |* ------------+-----+-- * *| Xo X=LN(Z) * * | * * | And it can get even crazier! See for example how an argument greater than Pi, can be sent to the other side (the spy who jumped over), to a negative Xo value. You start at X, go down until you cut the sine curve, then go to the left until you are on the Y-Axis at point Y. (The boarder to the other side ) Then ASIN (the master executor) sents you leftwards to the sine curve because that branch is nearest, and when you go up again, to meet the X-Axis, you realize that you are not where you started but at Xo<0. (The other spies cheated you, should we plan vengeance? SIN(LN(Z)) | | * * | * * Xo |* * X=LN(Z) -------------------Pi--+-- * | *| * | * | * | * | * o---Y------------o SIN(LN(Z)) And to make things 263better , the same happens at the other side, when LN(Z) is negative. The situation is mirrored there. I hope that the above was clear enough, but if it isnĪt wait until the PDF of this marathon is at hpcalc. The plots there are going to be much better. Anyway, we see that converting from LN to ASIN (or any other inverse trigonometric function) is not an easy thing to do. Now, the benefit of converting trigonometric functions to complex exponentials ( and/or logarithms ) is that many things bocome possible, which canĪt be done otherwise. Consider for example sum(n=0,N,SIN(n*X)). The HP49G canĪt return a result for this symbolic sum. And even if you have a numeric N, it takes a lot of time to return a result, when N goes to bigger values. But this sum can be calculated usinf conversion to complex exponentials. LetĪs do that. Enter sum(n=0,N,SIN(n*x)). Take the sum to the EQW and select SIN(n*X). Press now EXPLN. The term SIN(n*X) is converted to (e^(i*n*X)-1/(e^(i*n*X)))/(2*i). Press ENTER to take the sum to the stack. Now, before going any further, enter ĪNĪ ASSUME to tell the HP49G that the N of the sum is a real. Press DROP to remove ĪNĪ from the stack. Now press EXPAND and after some seconds you have the symbolic result for the symbolic sum! It is in complex form, but the imaginary part of it is 0, as it must be because we started from the real expression SIN(n*X). Press ENTER to make a copy of this result and then press IM. It takes a while, but then a result with trigonometric functions is returned. Press TCOLLECT to simplify this result to 0. Press DROP to get rid of the 0 and then RE to calculate the real part of the sum. Then press TCOLLECT EXPAND to get (SIN(X*N+X)-(SIN(X*N)+SIN(X)))/(2*COS(X)-2). This result is the sum of SIN(n*X) with n from 0 to N. It is valid for every N. We just have jumped to complex hyperspace (without a sheep on our back ), made things that are impossible in our real space, and then returned with the result. You can use this result to DEFINE user functions that calculate such a sum, with n from 0 to, say 1000, instead of waiting until the HP49G builds SIN(0*X)+SIN(1*X)+...+SIN(1000*X). The same way you can calculate sum(n=n0,N,SIN(n*X)) or sum(n=0,N,COS(n*X)) and so on. But wait a minute. The sum of SIN(n*X) with n from 0 to N is a finite quantity foe any X when N is a finite number. But our result for this sum contains 2*COS(X)-2 in the denominator. This is equal to 0, when X is 2*Pi*n1. Does this mean that then the sum is the infinite and we have made a mistake? No, because the numerator is then also 0, which tells us that we have 0/0 and so must work with limits. Because N in our result is an integer and because the HP49G still doesnĪt have INTEGERASSUME, lets put an integer value for N in our result. Press ENTER to make a copy if the result for the sum, and then enter ĪN=5Ī SUBST. Then enter ĪX=2*PiĪ lim. The HP49G returns 0 which is correct. That means that the expression (SIN(X*N+X)-(SIN(X*N)+SIN(X)))/(2*COS(X)-2) is 0 for N=5 and X=2*Pi. You can try also other combinations of values for N and X, like N=4 and X=2*Pi, N=3 and X=6*Pi and so on. This is also a nice demonstration. Because the sum SIN(o*X)+SIN(1*X)+SIN(2*X)+...+SIN(N*X) has no singularities when N is finite, so does its equivalent form (SIN(X*N+X)-(SIN(X*N)+SIN(X)))/(2*COS(X)-2). The real value of this form for X=2*n*Pi *is* the limit of of the form for X=2*n*Pi. (Rule of LĪHospital). It is not only that we can go infinitely near the point X=2*n*Pi to have a defined result for (SIN(X*N+X)-(SIN(X*N)+SIN(X)))/(2*COS(X)-2). This defined result exists also at the point X=2*n*Pi. because if it wouldnĪt, then also the sum SIN(o*X)+SIN(1*X)+SIN(2*X)+...+SIN(N*X) should have an undefined value for X=2*n*Pi, which is absurd! The same holds for every other form (like SIN(X)/X when X->0) if the limit exists. LetĪs move on to other conversions. The HP49G has also the hyperbolic functions SINH COSH TANH ASINH ACOSH AND ATANH built-in. The command EXPLN also converts such functions to complex exponentials. For example, enter COSH(X) and press EXPLN to convert this to (e^X+1/(e^X))/2. If you donĪt like this form (like I do) press LIN to get 1/2*e^X+1/2*e^-X. The opposite can be done with the command EXP2HYP, whic converts exponentials to hyperbolics. For example enter e^x and press EXP2HYP. The result is SINH(X)+COSH(X). The trigonometric functions SIN COS TAN ASIN ACOS ATAN can also be converted to hyperbolic functions with EXP2HYP. Enter TAN(X) and press EXP2HYP to get (SINH(2*i*X)+COSH(2*i*X)-1)/(i*SINH(2*i*X)+i*COSH(2*i*X)+i). And here is a table of all built-in conversions: EXPLN TRIG, INV TRIG ---------> EXP, LN EXP2HYP TRIG, INV TRIG ---------> HYP, LN SINCOS EXP ------------> TRIG EXP2HYP EXP -------------> HYP EXPLN HYP, INV HYP ------------> EXP LN EXPLN then SINCOS HYP, -------------------> TRIG LetĪs do some examples. ( Complex mode, X and Y are in REALASSUME ) 1) Show that COSH(X)^2-SINH(X)^2=1 Enter COSH(X)^2-SINH(X)^2. Press EXPLN and EXPAND to get a nice round 1. 2) Express e^(ACOS(X)) without using any EXP, TRIG or HYP functions. Enter e^(ACOS(X)), press EXPLN and EXPAND. Result: 1/(X+SQRT(X^2-1))^i 3)Express SIN(i*X) as a hyperbolic function. Enter SIN(i*X) and press EXP2HYP to get i*SINH(X) 4)Turn COS(X+i*Y) to an expression of functions with arguments X or Y. Enter COS(X+i*Y) since we want to have X or Y alone as arguments, press TEXPAND. You get COS(i*Y)*SIN(X)-SIN(i*Y)*SIN(X). Press the arrow down key to get this to the EQW. Select COS(i*Y) and press EXP2HYP. Now select SIN(i*Y) and press again EXP2HYP. Press ENTER to get the result to the stack. Result is: COSH(Y)*COS(X)-i*SINH(Y)*SIN(X). 5) Find the real and imaginary parts of SIN(X+i*Y). Enter SIN(x+i*Y) and press ENTER to make a copy of this at stack level 2. Press RE. You get (e^-Y*SIN(X)+e^Y*SIN(X))/2, whic is the real part. Is this equal to COSH(Y)*COS(X) from the last example? LetĪs see. Press COLLECT to get ((e^-Y+e^Y)*SIN(X))/2. Now take this to the EQW, select e^-Y+e^Y and press EXP2HYP. Result: 2*COSH(Y)*COS(X)/2. Press ENTER and then EXPAND to get COSH(Y)*COS(X). Now press SWAP to take SIN(x+i*Y) to stack level 1. Press IM to find the imaginary part. Press now COLLECT and take the resulting expression to the EQW. Select e^-Y-e^Y and press EXP2HYP. Press ENTER and then EXPAND to see that this is equal to the imaginary part of the previous example. 6)Show that if X,Y are real, then SIN(X+I*Y)*SIN(X-I*Y) is also real. Enter SIN(X+I*Y)*SIN(X-I*Y) and press TEXPAND. Take the result to the EQW. Select and apply EXP2HYP to all occurences of COS(i*Y) and SIN(i*Y). Press ENTER and EXPAND. The result is COSH(X)^2*SIN(X)^2+SINH(X)^2*COS(X)^2. 7) Find the symbolic sum sum(n=0,N,COSH(n*X)). Enter sum(n=0,N,COSH(n*X)). take this to the EQW, select COSH(n*X) and press EXPLN. Press ENTER and the EXPAND to find the symbolic sum. Yu can use PARTFRAC to split this a sum of smaller quotients, take this sum to the EQW and apply EXPAND to, each quitient. The result is then: e^(X*N+X)/(2*e^X-2)+e^X/((2*e^X-2)*e^(X*N+X))+1/2. 8) Convert e^(X+i*Y) to an expression with trigonometric functions. Enter e^(X+i*Y) and press SINCOS. 9)Solve COSH(X)+e^X=0 for X. If you try to solve this with the built-in SOLVE then you get the error 263Not reducible to a rational expression. But if you ENTER COSH(X)+e^X=0, press EXPLN and then solve this for X, you get a list with 2 solutions:X=2*i*n2+LN(i*SQRT(3)/3) X=2*i*n2+LN(-i*SQRT(3)/3). (I EXPANDed the arguments of LN in both solutions). Now, what do you think? Should TRISOL make its evolutionary way to HYPEXTRISOL? 10)Convert ((e^X)^6+(e^X)^5+e^X-1)/(2*(e^X)^3) to a more simple expression that contains hyperbolic functions of multiples of X. Enter ((e^X)^6+(e^X)^5+e^X-1)/(2*(e^X)^3). Since we want multiples of X like 2*X, 3*X and so on, as arguments of hyperbolic functions, it seems reasonable to use LIN, to turn things like (e^X)^n to things like e^(n*X). Press LIN. Now press EXP2HYP. The result is SINH(3*X)+COSH(2*X) for STARTEQW, modified to contain all commands like EXP2HYP, EXP2POW etc., Ending this last part of the Trigonometry Marathon, I want to say thanks to all people who commended, corrected and asked. This was one of the main powers that kept me on working. Trabakoulas also wants to thank you all, for helping find all his sheep. (Except the one at the ski jump, of course ) Before putting the COLLECTed PDF parts of this marathon to hpcalc, IĪll add a part with trigonometric/hyperbolic/exponential conversions on the HP48. But IĪll not post this part here. Also IĪll put the newest version of TRISOL to hpcalc. ), where Kojak will SOLVE complex cases with the joint forces of TRISOL and COMSOL. Or was it COMTRISOL? Or HYPEXCOMTRISOL? Well, we will see. Also Trabakoulas will be running on ice making complex jumps with sheep. Hyperbolic ==== ... You're right, Martin: I think its called a wedge product or some such. The problem is, in n-dimensional space, you need n-1 vectors as arguments for this *product*. In 3-dimensional space, you can take 2 vectors, say AxB. But in 4-dimensional space, you need 3 vectors, say AxBxC. Dan ==== With HP calcs we rarely find that a stable release is more stable than a development release which is newer. Somebody once said I was crazy when I was asking if a new rom would be released before I took a standardized test (it wasn't, I think it was -5 which I had asked about). The poster said, to a certain extent, that nobody should use development software in a critical situation (such a major test). When the next ROM revision came out, it indeed fixed the bugs which were causing me problems. I am quite sure that if it had been released in time for me to take my test I would have performed much better on the test (as it was, I ended up having to turn in the test with a lot of problems unworked because I had to do workarounds for the bugs which the developers had said they'd fixed but didn't release the ROM with the fix for about 6 months). ==== Most weaknesses of the TI89 that HP zealots point to are made up to give the TI89 a bad name. I just want to do the same STUPID THING the other way round. 1. Laplace: I bought the TI89 because of Lars Frederiksen's Laplace transform program. The HP49 could not satisfy me because it could not solve the following: LAP(SQR(x)), LAP(1/SQRT(x)),LAP(EXP(-A*X}/SQRT(pi*X)),LAP((1-cos(x))/t),LAP((1+2*a*X)/(SQ R T(pi*X)),LAP(2/x*(cos(b*X-cos(a*X)), LAP(2/X*SINH(A*X)). All these any MANY more can't be done by the HP49G, but can be done by Laplace89 on the TI89. 2. Integrate: The Risch algorithm in the HP49G just makes it MARGINALLY better. The Steb-by-Step for the INTVX command and other commands is useless because it's too crpytic and shows little or no steps to be useful in an examination. Can your HP49G integrate the following?: INTEGRATE(LN(X+SQRT(1+x^2))), INTEGRATE(SQRT(X)/SQRT(A^3-X^3)), INTEGRATE(SQRT(A-X)/SQRT(X)), INTEGRATE(COS(X)*SINH(X)), INTEGRATE(X*EXP(1/2*ABS(LN(X^2)))),INTEGRATE((X^3-X^2)^(1/3)), INTEGRATE(LN(1+X)/SQRT(X)), INTEGRATE(X^2/SQRT(X^6+25)). The TI89 can do all these quickly. The HP49G is faster only 20% of the time for all problems. 3. Summation: The HP49 can't evaluate the following summation: sum(1/n!,n,0,infinity) 4. Number of commands: It appears the HP49G has more command than the TI89. It appears so because many of the commands on the HP49G can be done by fewer commands on the TI89. a. Risch, Preval, Int, INTVX, INTV, IBP can all be done by INTEGRATE on the TI89 b. ABCUV, EGCD, IABCUV, IEGCD can all be done by EXPAND (if used well) on the 89 c. DIVPC, TAYLOR0 can all be done by TAYLOR on the TI89 d. LINSOLVE, SOLVEVX, SOLVE can all be done by one solve command on the TI89 e. BAR, SCATTER, DIFFEG, FUNCTION , etc can all be done by the SETMODE command on the TI89 f. LNCOLLECT, LIN, TRIG, TEXPAND, TSIMP, TLIN TCOLLECT are all done automatically be the TI89 g. Too many to mention 4. Built-in functions: It's true that HP49G has more built in function than the TI89. But it's a blessing that the TI89 lacks such built-in command, because the 3rd party programs on the TI89 that do those built-in command in the HP49G are the best you can ever find on any calculator. Examples of those wonder 3rd party programs are: Mathtools by B. Bhatt, DIFFEQ/LAPLACE/FOURIER by L. Frederiksen, SIGNAL89 by...can't remember, etc. What's worth doing is worth doing well. That's exactly what these third party programs do for the TI89. For example, the built-in EQW on the HP49G can't edit matrices, but the 3rd-party EQW on the TI89 can. 5. Spreadsheet App: The TI89's data editor is better than the HP49G's data editor. 6. Text editor: It's true that the HP49G has a better text edior. But I can't imagine somebody typing long texts on a calculor. Why go through such drudgery?. So, I feel those better features on the HP49G's text editor are unnecessary. Get A PC instead. 7. Directory Structure: The HP49G has a better directory structure. AND SO WHAT? Completely unnecessary to me. The TI89's directory structure still gets the job done. 8. More fonts: The HP49G has more fonts. AND SO WHAT? To me one font will suffice for a calculator. 9. Alarm Clock: I don't need this because it's not necessary, thank you! My alarm, multi-chronograph Citizen wrist watchcan buy two HP49Gs. Besides, most people own wristwatches with buzzers. 8. Connectivity Protocols: True , the HP49G has Kermit, Xmodem, Zmodem, etc. The TI89's protocol still gets the job done EQUALLY. The HP49G doesn't have something similar to TI CONNECT. 9. Programmability: The HP49G has no SDK. The C SDK for the TI89 gives the TI89 an edge over the HP49G's numerous programming languages 10. Display: The TI89 has a better display.-) 11. Keypad: There is a high probability that the key inscriptions on the HP49G will wear off quickly. 12. Processor: The TI89 has a better processor. There is more information for programming the MC68000 than there is for the Saturn. 13. Memory: The TI89's memory is more than enough for me. Conclusion: ANY product that fails to generate enough revenue for its maker is aa BIG FAILURE!! The HP49G (with all its 762 commands) is a PERFECT example of a failure!! No wonder they had to pull the plug on it. It's a shame BP's works was not seen as profitable to some CEOs--too bad!! ehehe!! ==== 12. I forgot to say the HP49G can't find LU's of symbolic matrices. ==== Oh! very well. We all can see how happy you are with your TI 89. And excuse me if I tell you that I'm very happy with my HP 48GX. Have a good day. ==== And I forgot to say that your post does not seem peace maker... R Lion escribi227 en el mensaje ==== In addition to the integrals I posted earlier, these are more integrals that the HP49G can't solve but the TI89 can solve: 1.INTVX(SQRT(1+cos(2x))) 2.INTVX(EXP(TAN(X))*(1/COS(X))^2) 3.INTVX((4X^2)/(SQRT(1-X^6))) 4.INTVX(COS(LOG(X))/X) 5.INTVX(SQRT(TAN(X))*(1/COS(X))^2) 6.INTVX((1/COS(X))*LOG(1/COS(X)+TAN(X))) 7.INTVX(((1/COS(X))*(1/SIN(X)))/LOG(TAN(X))) and others, in case you try to defend the HP49G, please don't get me started with you cowardly EXPLN command--let the INTVX take the bull by the horn. I have decided not to post a speed comparison chart for these too calculators. If I do that, the HP49G will look like an abacus. But, if I am provoked, I will surely post a speed comparison chart. There are many more problems the HP49G can't solve; i am just too tired to list them. OAA(TI DG name) ==== Wow!! A good way to start a war Peace Maker... J.Manrique CdU de la ETSIG #1077 HPCC Member ==== Peace Maker schrieb in im Newsbeitrag: 752cb1bc.0202170406.a06994f@posting.google.com... Give it to them OAA!!!!. ;-D I only have to add that the TI89 has Symbulator, while in the best case the 49 has CSIM entirely unable to solve simbolic circuits. Maybe you should do a doc with your posts. It gives an alternative to all of us who read Tim Wessmans comparison (and almost puked!). cheers cybernesto (here and in TI DGs) ==== Ok, I would like to know if Ti89 solves all these integrals without assuming anything for x. I mean, most of the integrals that you mention have some asintontic points for some x values (sqrt(x) is not valid for x<0, 1/(1+x) is infinity for x=-1,..) so you need to assume that things when you solve and integral with the hp49 to give the correct answer.. Probably the ti89 assumes that, probably not, I don't know, but giving an answer without taking care to this could be a bad answer.. Probably this is not the expected behaviour for a calculator, I mean, we are talking about machines that gives solutions to problems, but symbolics need thinkig a bit what are you doing.. I don't like many things from hp49, but software is great for the machine that it is.. Ok, I would like better processor, better screen.. but the same software with better machine is unbeatable.. J.Manrique CdU de la ETSIG PS: I hope some day any company will invest in handhelds calculators R&S departments, or this world won't be the same. TI is not a exception, its Voyager2000 machine is not more than a Ti92 plusplusplus.. Ok, it's a great machine, but for me, it's similar mistake than hp49.. just a hp48plusplusplus ==== I used the TI92 since the first release.My first TI92 even has a beta version of the first ROM(1.00).Thus i am much better placed to know TI89 weaknesses than the fanboys which used TI68k calculators since the TI89 release. And i stated that the TI89 have a lot of weaknesses. A lot of them have been lessered by the heroic work of some people such as David Ellsworth and many others. People that TI has pissed off and disapointed over the years and which have for most of them completely given up support of TI calculators. Without those people,fanboys like you wouldn't even dare to come here and post such messages. This function has been developped by a third party not by TI. A third party which has been very disapointed by TI,btw. Sure,but can your beloved TI89 finds the anti-derivative this function that anyone with even the lowest level in antiderivative computation can: (2*x+1)*exp(x^2+x) ? Each of this calculator fails on certain integrations. This is one of the rare case where the TI is better than the HP.Symbolic summation,eqyations sovling and autosimplifaction.For the other cases either the 2 calc are equals or the HP49 has the edge. And off course you think that it explain everything ? Pfft,what a fanboy. The TI doesn't have an equivalent for IBP function. Here we go again. Again third party programs versus built-in function of the HP. Really ? You will need to explain to me how you can do DIVPC with the TI taylor function. I know how but i doubt that a kid like you know that. Moreover as we are talking of taylor series,let me know how much time your TI89 needs to compute that: taylor(tanh(tan(x))-tan(tanh(x)),x,7). And has your TI89 an equivalent to the series command of the HP49 ? Sure,you can use the solve command with only one argument ? And can you use the solve command with matrix coefficients as arguments ? What a poor choice of example. Do you know that there is also the equivalent of the setmode command on the HP ? And that this command can do much more than the setmode of the TI ? Here i catch you. You don't even know that the TI89 has a Texpand function,do you ? And when it will be possible to control simplifications on the TI,let me know. Yeah,off course. To be serious,there are tons of functions that the HP can do without equivalent for the TI. I haven't time to list them but you should compare the 2 calc for polynomials functions,linar algebra,integer arithmetic and trigonometry. But this is the main problem with the TI89. For doing anything advanced you need third parties programs. Isn't that pathetic ? Moreover with TI attitude toward those third party developpers i wouldn't rely too much on them in the future. Heard of the TI-Cas affair ? I wonder why people like Bhuvanesh waste their time developping software for TI. You contradict yourself here. If you won't need a calculator for doing word processing,why would you need a spreadsheet on a calculator ? The difference of memory management between these 2 calculators is much bigger than just the directory structure. In fact the HP49 memory management is light years above the TI89's. Could you enter names with more than 8 characters in the TI ? Can your idea handle unlimited number of variables within availlable memory ? And if you say,yes then try this: For k,1,3000 0->#(k&string(k)) Endfor That would show you how pathetic the memory management of the TI is as this program just tries to create 3000 variables of a vey small size on the TI. And what about the TI89 24 KB limitations for ASM programs ? And that stupid 64 KB size limitation for variables ? Can you do backups of your RAM in the user Flash ? Can your TI89 even try to recover RAM after a crash ? This is called customability. To choose the font size or even edit fonts. To customize the usable part of the home screen. To choose between RPN and Algebraic. To choose between 4 built-in langages. To choose between pulldown menus and softkeys. To fully customize the keyboard,etc... Something you TI89 fanboys can only dream off. Funny to see that when you don't need a HP49 feature,you find it useless on the HP49. Guess what ? Apparently TI has finally found out that Time can be useful on a calculator as the V200 will have one as well as the TI89/TI92+ HW2 with the A.M.S 2.06. As well as the TI89 protocol get the job done,the HP49 software program get the job done. What a pity. Do you know that the HP49 has a built-in SDK ? And TI is famous for their negative attitude toward developpers. If not what are all those stupid limitations in the software and in the hardware ? As you haven't used the first TI92,you don't know how poor their screen was despite the high resolution. Not the new keyboards.Moreover the HP49 keyboard has a much better layout. This is certainly why there is much more software for the HP calculators than for the TI calculators besides games. And it is a shame that despite a 12 times faster C.P.U,the HP49 can outperform the TI89 for many things. The truth is that there is never enough memory. Unless of course with pathetic memory management such as the TI one. For gosh sake,i have rarely seen such a fanboy. You are happy with your TI89 ? Good for you. But don't come here insulting people with your name and your fanboyism. ==== Have you tried HP49 with neopolys? Marginally better doesn't mean better? If you don't want to use the bulit-in funcitions, don't use them, but it's better if they are available. I use very good software, probably the best available for my calc, but I love to program my own routines and those functions are wonderful for me. Haven`t you used the text editor of your calc? I use it often, and I want the best text editor possible. If you only perform sums on your calc, that's OK for you, you don't need any text editor. Better say that a text editor is unnecessary. But sometimes I do intensive use of it. Well, I think that you are conscient of the superiority of HP directory structure over TI.. It doesn't matter that TI do the job for you. Better is better. I don't use multiple fonts but better if I have it built-in. Maybe I will need them one day. Who knows. have a lot of symbolic mathmatics programs, a lot of science, chemistry, physics programs and thousands of utilities. Still do you think that HP's are less programmable (or harder programmable) than TI's? I agree with you here, but I coult argument like you saying some nonsenses like: I find my calc's display wonderful, so, why do I need more? I think better display is unnecesary bla bla bla bla... Faster doesn't mean better. And there is more info than you need to program the Saturn CPU in various languages. I know persons who fill completely 2 MB RAMcards in a HP48 GX expanded. For me it's wonderful if you have more than enough with TI89's memory ==== I also have a TI89, and use it almost never. Here's why: The HP49 shames the TI89 in quickly writing programs from the keyboard in UserRPL, which is extremely useful for me. The HP49's lack of automatic simplification makes the use of formulas very easy. As just one example, I have some 800 integral patterns that I solved with Maple, and entered into a directory in my HP49. This gives me much more control over how I want to integrate a function. I haved done the same for physics, difeqs, etc... You can't do this on a TI89! The HP49's file structure,with directories and subdirectories is completely superior to that of the TI89. The HP49's file structure is like a PC, well organized, usable, and an immense help to me. You can't use the file structure of the TI89, it's an featureless joke! The memory of the HP49 is more than double that of the TI89, and ALL of it is accessable to the user. In addition, the memory of the HP49 is protected with the ARCHIVE command, and the RESTORE command easily resets everything. If you crash your TI89, you're screwed until you can reload from your PC. I was trying to learn SysRPL, and regularly crashed my HP49. The ARCHIVE command saved my ass in the middle of a physics exam. I was able to reset everything, and finish the test. You can't do this on a TI89! Last but not least, almost every available program for the HP49 is free. Recently, TI had the nerve to advertize an Equation Writer program for sale, something that's been in the HP49 since the very first one came off the production line. I'm glad you're happy with your TI89. However, crowing about it here will gain you little. ==== Because two wrongs make a right? Are you becoming a troll now, OAA? Because if you are, you will have to shut up on the TI DGs about HP zealots. Deal? LAP(1/SQRT(x)),LAP(EXP(-A*X}/SQRT(pi*X)),LAP((1-cos(x))/t),LAP((1+2*a*X)/(SQ RT(pi*X)),LAP(2/x*(cos(b*X-cos(a*X)), Sure - LF is a good programmer, but it's nice to see that the TI89/92 needs third party programs to even have LaPlace included. It's a good feature, that it's not built-in, or do I misunderstand you? If I had the need for it, I could code a LaPlace program for the TI89, that is faster than Lars', and can do more transforms. It would probably even be faster on the '49. Don't kid yourself. I went to EE school with Lars here in Denmark - do you think he is God? How about inverse LaPlace? The TI89 isn't doing that well on those, is it? Why has he stopped working on TI programs? How about these?: INTEGRATE((2*X^2+1)*EXP(X^2)) INTEGRATE((6*X^5-24*X^2)*EXP(X^6-8*X^3+16)) And so on? But the 68k processor is also roughly 15 times faster than the Saturn. It has also got hardware multiplication and division - the Saturn has to do with software emulation. Can the TI89 do SUM(X*X!,X,Y)? It's '(Y+1)*Y!-1' if you're wondering. Ok. What about EULER, IBERNOULLI, ICHINREM, PA2B2, NEXTPRIME, PREVPRIME, No, the TI89 doesn't support the Risch algorithm for example. It can't do integration by parts either. What is INTV on the '49? Right, but not that easily, huh? You can program many things too. Ok, how about 'SQRT(X^2+X+1)' at infinity? Or 'X^X' at zero? Or 'ASIN(X)' at -1?. Or even 'SIN(X)/(EXP(X)-1)' at zero? True - that was three. SETMODE is really equivalent to SF and CF. The flexibility of the modes setting commands on the '49 is good in my opinion. Sure. An example: LIMIT((X^3-5*X^2-4*X+20)/(LN(X^3+6*X^2-27*X-140)-1),X,0) on the TI89 yields '20*(LN(140)-1)/(2*(LN(20)-1)*LN(7)+(LN(7))^2+(LN(5))^2+2*LN(5)*(2*LN(2)-1)+ 4*(LN(2))^2-4*LN(2)+pi^2+1)-(20*pi/(2*(LN(20)-1)*LN(7)+(LN(7))^2+(LN(5))^2+2 *LN(5)*(2*LN(2)-1)+4*(LN(2))^2-4*LN(2)+pi^2+1))*i'. The same operation on the HP49G yields '20/(LN(140)+(i*pi-1))'. Which do you find simpler? Hurray for autosimplification. How do you get to the HP49G result on the TI89? If a function is built-in, does that mean that it can't be programmed? The Matrix Writer on the '49 can edit Matrix elements in pretty print. That should be enough? Do you need a spreadsheet on a calculator... ...more than a text editor? The TI89 doesn't have a directory structure in my opinion. You can't let programs reach from one folder to another, hence it is a pain in the ass. And so what for anything? Are things you do not use superfluos? I do not own a wristwatch. Just because YOU don't need it, doesn't mean that other people won't. I still fail to see how you can complain about too many fatures? Not equally - The TI protocol is proprietary. Shesh - what about HPComm? Many people complain that TI Connect doesn't work on WinXP (it does with me though). HPComm works fine with XP. On the programming side of things, nothing beats the HP49G! You have ML and SysRPL built-in, and you have an SDK built-in. You can even download a syntax-helper and use autocompletion on the '49. It has higher resolution, but the dispay isn't better. You do not have significantly more workspace anyway. I seem to have escaped that high probability. I have used my HP49G intensively since it was released, and my keyprints are fine. The keys on my TI89 got very loose after one week of use, and I wonder daily when one will pop of the calc. And tat is supported by TI, right? You have zero support from TI, so you're basicaly f*****, when not using TI-BASIC. ==== Mine lists 982 commands by the way. ==== The cowardly EXPLN command? Is that because you don't understand how it works? If a carpenter doesn't know how to use a saw, is it the saws fault if the carpenter saws askew? But that's your intention, is it not? Why don't you post a speed comparison then? That'll show us all your understanding of math. I'm preparing one too, so we can have a contest? My comparison will be a bit bigger than usual, as I'm running quite a few tests (and not only math ones either). The TI89 is fast, yes, very fast - but it's pretty damn stupid if you ask me. We'll see. You do remember that the CAS of the TI89 is based on Derive, and not Mathematice? ==== What is Maple? What do you mean by entering integration patterns? Could you elaborate on that a little bit? I'm in Calculus right now and any good Calculus app is useful. LAP(1/SQRT(x)),LAP(EXP(-A*X}/SQRT(pi*X)),LAP((1-cos(x))/t),LAP((1+2*a*X)/(SQ RT(pi*X)),LAP(2/x*(cos(b*X-cos(a*X)), ==== J.C. Randerson skrev i en nyhedsmeddelelse:5ba17a08.0202170905.101cfa67@posting.google.com... I assume that you must have got your HP49 for free since you state that you haven't paid for its Equation Writer. Where can I get a HP49 for free, so I can have its Equation Writer for free too? What do think you were paying for when you bought your calculator? Do you think it was only for HW? ==== Integration is some sort of art. For me it is better to have tools on hand that allow me to somehow perform such integrations, than to have a command that perhaps integrate thisbut nor that. The offered commands of the HP49G when used wisely in combination can do wonders. (Well, not exactly but almost) One of the examples that you gave: INTEGRATE(COS(X)*SINH(X)) . Use EXPLN to turn trigonometric and hyperbolic funcions to exponentials and then you can integrate this. (I am totally trigo-expo-hyper-high these days ) Unfortunately! :-( Here I must disagree! A computer mathematics tool, must have a certain complexity (and versatility) that allows the formulation of many problems in many ways, and that allows the interested user to influence the way that some computation follows. If you simply want a black box, which gives you the answers that *it* wants and you just stupidly write them down, then the HP49G is not your tool. But if you are clever enough, you can use the many commands, to compose formulations that lead to the results tha *you* wish. The language for maths is a rich one. Don't forget this. A table is a couple of values that are violently ripped out of the place ehere they belong: The function. A good text editor is needed for programming, for formulating the problem and its solution. If you had made the effort to expore what the advantages (and sometimes disadvabtages) are, then you would love it. But no, without any imagination of what the directory structure allows, you simply say that it is not needed. Readability of programs in text form can be of vital value, when editing. Believe me, TRISOL would turn me mad otherwise. Yes, of course your alarm watch can wake up at certain time intervals, receive some dat from the pH-meter, Spectrometer or whatever-meter, do some calculations, put the data in some matrix and sleep again. When you use the calc as a stand alone data-collector, you will appreciate all these protocols, and XSEND SRECV even more. This is where the complex set of user commands does its job. Most of the time you don't need any additional tools to program something on the HP49G. Absolutely. And the TI92 even better. The speedy MC68000 is better when comparing brute force. But not for me! So you say that the best products are those which sell better? Well, Dolby S. And many many other examples. Revenue has to do with marketing, not necessarily with quality. I hardly can say that the HP49G is a failure because of shortcomings in software. If you give someone one fish to eat, he doesn't hunger one day. If you give him 100 fishes (TI), he doesn't hunger 100 days. But if you teach him how to fish (HP) he will never be hungry again. Except of course when all fishes will be fished and eaten, which is a very different problem that even the HP49G cannot solve. ==== .. LAP(1/SQRT(x)),LAP(EXP(-A*X}/SQRT(pi*X)),LAP((1-cos(x))/t),LAP((1+2*a*X)/(SQ RT(pi*X)),LAP(2/x*(cos(b*X-cos(a*X)), I hope for you that you always have a online connection to comp.sys.hp48 when you are taking a test else you could be in trouble if you had to solve this type of job with your calculator. ;) ==== solve If you're in trouble because your calculator can't do a LaPlace transform, then you should quit school. The idea is that you learn how to do it by hand, and then you don't need the calculator other than for checking maybe. Pretty sad when students rely so heavily on a calculator ==== do Does he incourage you to do it with the calculator? I'd like to smack him a couple upside the head! Incompetence in color for you. In which country is such practice followed? I'd never hire a guy who didn't know the reasons behind his work. ==== It is years since I was taking such tests, and I tell you we had *nothing else* than our pens and paper sheets. We had not even a slide rule and the exercises were real wierdos. But we succeeded not because we had learned to press keys, but to think. ==== For real world problems, the TI89 is often very slow. I have in my new still non-disclosed comparison numerous examples of the TI89 using more than 10 times the time the HP49G uses. The implemented algorithms on the TI89 are also limited in some ways. The TI89 is a fast school calculator, but lacks when the going gets tough (out of the common Schaum tables that is). I'm still surprised by your sudden complete lack of good manners, my dear OAA. You're not any better than the common troll now, are you? Wonder what Ray and the others would say about this. Hmm, Bhuvanesh reads this group too ==== I can never become a troll. You and co have just been annoying with the kind of attacks you make on TI. I know Lars is not a God, but he seems to understand Laplace transformations better than all HP49G programmers combined! Oops Your two examples are alike. Besides, you have been giving the same examples over and over again--get brand new ones. The TI89 does integration by parts automatically when necessary. I know the TI89 doesn't support Risch algorithm. If the TI89 can integrate wonderfully wihout the Risch algorithms, what's then the use of the Risch algorithm? Easy for me!!! Same old problems you have posted a trillion times. Do you need a clock on a calculator? Still ok for me. Because of too many unnecessary features the programmers of the HP49G almost encroached into the flash ROM space for user files At least, it can still upload and download files the way HP49G's standard protocols do. Can HPComm automatically search for program updates? So you do enjoy programming on your calculator with those tiny keys? I can't even imagine myself writing long programs on a calculator--I am too young to die. LOL I can always go to Motorola if I need support for programming the TI89 with assembly. Besides, TI support is getting better. Sorry, if I insulted anybody here. But, I said the truth, and that's why it hurts. I hope this war ends quickly. ==== I already know some about integration. U subsitution, etc. Where can I find some info on Maple? And can you share your integral database that you entered? Export from your calc so we can all benefit? LAP(1/SQRT(x)),LAP(EXP(-A*X}/SQRT(pi*X)),LAP((1-cos(x))/t),LAP((1+2*a*X)/(SQ ==== A troll is someone who starts fights without reason. You did - you're a troll. Erhhhh? No? I thank I have asked BP to include LaPLace transforms a zillion times before he did. At first the '49 didn't even support Delta and Heaviside functions, but it does now, after people have asked for it. BP didn't find LaPLace transforms that useful, and obviously the '49 users didn't eithre, or else it would have been much better. You want an example? How about this one: INTEGRATE((X^3*EXP(X^2))/(X^2+1)^2,X) And don't start with that RISCH crap - the above example is perfectly solved with integration by parts. On the HP49G, the INTVX command can even do it (you don't need IBP here). But you'll see.... 'ASIN(X)' Does that make them moot? But again - wait till you see the comparison. There are *many* more where they came from. Yes. I use the clock on the '49 alot. It's also good for data acquisition. But then again - you don't expect the TI89 to handle that, do you? ass. So the TI89 is best for you - we have established that already. If I were ok with a power cord instead of batteries, would that make a calculator using batteries more crappy than a calculator with a power cord? I guess not. We're not discussing opinions here, but hard facts. That's what you started, that's what you gotta stick to. Right. Let's get rid of pretty-print then - it doesn't help solving math problems, and uses up time for the programmers. How about the TI89 text editor? Let's complain to TI that they used up precious time on that one. And that's what you use protocols for, right? Come on - what about data transfers of other kinds? Have you ever thought of that? You can't fetch data with the TI89 - it's a closed black box system. No, but it does everything else. Is it hard for you to keep up with one AMS update every other year? Yes, as a matter of fact, I do. Every bit of code I have ever written for the '49 is written on the calc itself. There are also programming tools for the PC - did that escape your eyes? on my will Is that funny? It may be because it's not a prority to you, that the keys stay in pace, as long as the print does? And reinventing the wheel? Have you ever programmed assembly? I doubt it, since you take so lightly on code-reuse. Since when? Is that why they closed TI-CAS down (or Miclo at least)? Is that why they don't seem to keep their promise regarding the contest EW won? Is that why they are charging everyone for software that should have been included in the TI89 in the first place? Is that why they are maintaining the 24 kB ASM limit? Is that why *there is no programming support whatsoever*, outside of TI_BASIC? Is that why none of the 300+ requests in the wishlist have ever been fulfilled? AMS v2.06 better be good for your statement to be true. But, oh no, it has a clock, so it can't be good. Wonder what they didn't have time to include, by coding a clock? a SERIES command maybe? The truth? You told your opinion, and are trying to stuff that into our throats. Excuse me, but that's trolling for ya. You haven't insulted anyone personally, you're just making a fool out of yourself. I've certainly put any weight to what you're saying at the TI DGs for the last time. Then think twice before startig one. You may have some valid points in there, but you certainly make sure that they stay hidden by going on like that. ==== You are saying that you are happy with your calc, aren't you? Ok, then what is your problem with the hp49? You don't like it, do you? The.. live and let live If you would like a comparassion, here it is: -I have a linux ipaq3130 with 206MHz StrongARM procesor and an under-development computer algebra system on it that beats my hp49 in many ways..and I bought it for same money than a hp49... And I am still happy with my hp49, because it is great for the machine that it is.. To know more about Giac-Xcas check Bernard Parisse's site: http://www-fourier.ujf-grenoble.fr/~parisse/giacdoc.html I don't understand the reason for this TI-HP war, it would be nice if both could learn from the other's mistakes to develop better machines, but I think this is not the case. TI has the oportunity to make a great new machine (Voyage200), but they had decided to make a new ti92.. same mistake than hp49.. One finally and funny think, why Voyage200? It's similar to ==== At least it's the best Laplace transform program available for a calculator. hehehe! This same old problem has been posted 13, 872 times by HP fans just to give the TI89 a bad name. Get new problems, PLEASE!! I know. The IBP function is integrated into the INTEGRATE of the TI89 command. Check the algorithms used by the INTEGRATE command in the TI89 manual. Besides, on the HP49G, you still end up using the INTVX command after using the IBP command. Anybody can do all these with EXPAND command ONLY!! Hell YES!! The DIVPC command is used to find the Taylor series of a quotient. The TI TAYLOR command accepts quotient as well. At least there is still a high probability that the upcoming AMSs will contain additional function. For the HP49G side, there are no more chances for an updated ROM. I know! Damn it! I was just trying to say that the HP49G has fewer distinct and unique functions than it appears to have. The commands are repetitive! I know the TI89 has a TEXPAND function. I made a typo there, ok!!! Well the 3rd party programs end-up being better than several HP49G in-built commands!! The patient dog, my friend, eats the largest bone!! It's easier to type numbers than text with any calculator The differnce in processor speed between these 2 calculators is much bigger than any advantage the HP49G has over the TI89. About 127^8 types of allowable names is okay for me. Yeah, yeah, yeah! Upon all these they still had to plug the on the poor overdeveloped thing--too bad. Besides, the fact that TI calculator R & D is still alive raises hope for equipping the TI89 with all these things you have mentioned The TI89 rarely crashes. I never dream of that because they are not too important. All these things you have mentioned still ride on one SLOW processor. The TI89 has built in languages has well. It's really funny to know that all what the HP49G programmers got at the end for adding all these to the HP49G was a lay-off letter.:) I am aware of all that! I know that the HP49G has a built-in SDK. Goodluck in straining your fingers and your eyes with a calculator I know. The limitations are there to generate revenue for TI so that TI R&D (for calculators) will not end up like HP's. I prefer the TI89 layout because I hate having to press a second function key for a comma, '=','y','z','t',etc The 3rd party software available for the TI89 are usually better than their HP49G equivalent Not many at all!! Sorry to anybody I insulted, Ok? But I said the truth, right?!!!!! The truth hurt, right? The name, Peace Maker, I used is the name of a USAF MIRV-ICBM. It's the best ICBM the USA has for wagging a nuclear war. ==== That's a hot one. I know perfectly what the EXPLN does. I just did not want the HP49G owners to use it as alternative or a way of solving those integrals I listed. I didn't express myself properly there; so you misunderstood me. Yes. And it has always been the intention of you, Tim Wissman and Co to give the TI89 VERY little credit. You have never been fair. Did any TI calculator ever cause you to loose something dear? What has TI done to you? That's another hot one.-) Your calling the TI89 a stupid thing clearly show your bias nature. That just degrades the effort many smart people have put into it. I will not post anymore messages again about comparisons. This thing is getting out of hands. My aim was to compare the two calculators jokingly and not to compare our math skills. How would like it if someone calls your good efforts stupid? The TI89 is partially based on the Derive engine. ==== The processor speed has nothing to do with performance. I start wondering how many times you will correct yourself in this thread? Try this on your TI: 'factor(expand((x*y+2*x-z)*(x+y*z-3*y)*(6*x^3+x*y+x*z+1)))' The HP49G uses 18.36 seconds in comparison. Some time next week, if you ever get done, please post again. Here are a couple more: 'factor(expand((x-y)^4*(8*x+y+21)^4))' (HP49G: 14.03 seconds) 'factor(expand((x+y+z)^7-x^7-y^7-z^7))' (HP49G: 45.43 seconds) 'factor(expand((x^3+5*x^2-4*x-12)*(x^4+5*x^2+4*x-12)))' (HP49G: 4.64 seconds) 'factor(expand((x^3+x+1)*(x^3-x+1)))' (HP49G: 2.15 seconds) Of course, since names like xisdhuye & wzwrtyxy are easy to remeber and use? You have roughly 30000 times fewer names to use than 127^8, so you'll run dry before you know it.... That's a fat chance. Oh, no. But a flash upgrade whipes everything.... The TI92 is bad at crashing, but the newer AMS' aren't (v2.05 for example). The new '49 ROMs are very stable too. For you, that is. Won't you please stop making so many personal opinions ==== If you have to do the STUPID THING, then at least do it right. Many of the examples that you gave as, can be done with the HP49G. Because I don't want to call you a lier, I call you an ignorant. If SQR is square, then this can be done on the HP49G without any further work. And also this. Yes it can! Substitute X=SINH(Y), use EXPLN to turn hyperbolic functions to exponentials and EXPAND the integrant. Yes, without any additional work. Yes! Turn the integral to complex exponentials with EXPLN. Yes! In real mode EXPAND the integrand. Wait for my next attack F22raptor. And be afraid! Be very afraid, that your rockets against the HP49G will prove to be soap bubbles! ==== Some more greetings TEXPAND the integrand. EXPLN the trigonometric functions of the integrant. The one that is a coward here is you. Only cowards need lies to convince other people. If you choose lies as your weapons against the 49, then you must let us choose our weapons, which include EXPLN.  ==== The tricks you talk about are mathematical methods that obviously go beyond your skills. When someone masters this tricks, he/she won't have the problem Momi, my calc doesn't do this integral, I need a new one. What are you going to do when the TI doesn't integrate something? Despite your low understanding about tricks (or perhaps because of this) here you are: Substitution X=SINH(X)*a, EXPLN the integrand. Did you notice that your third example is the same with the first? Yes, sooo many problems of which you posted proved to be completely easily solveble with the HP49G. (See also my first answer) I can't imagine how you are going to do this, when your math knowledge is below zero. But anyway, go on, post your comparison, and then wait for mine. >:-D ==== I don't know if you noticed but this is an HP ng. What is your contribution? Your point is well taken but all the users here already own or owned at some point an HP calculator. I don't see your comments helping anyone. Or... maybe I'm wrong and users will create the programs to counteract the weaknesses you mention about the 49G using one of HP49G's numerous programming languages. In any case, it depends on the use you intend to give your calculator. You can't compare the wealth of advanced software already developed for the HP's. Civil and structural engineers will tend to agree that the HP's are unmatched by any other calc. As for the difference in the # of commands, when you learn how to use them you discover that the flexibility they provide is more than welcome compared to the more inclusive Ti commands. Arguing about calculator is really stupid. Use whatever feels best for your needs and go post constructive posts on the appropriate ng. ==== I don't know if you noticed but this is an HP ng. What is your contribution? Your point is well taken but all the users here already own or owned at some point an HP calculator. I don't see your comments helping anyone. Or... maybe I'm wrong and users will create the programs to counteract the weaknesses you mention about the 49G using one of HP49G's numerous programming languages. In any case, it depends on the use you intend to give your calculator. You can't compare the wealth of advanced software already developed for the HP's. Civil and structural engineers will tend to agree that the HP's are unmatched by any other calc. As for the difference in the # of commands, when you learn how to use them you discover that the flexibility they provide is more than welcome compared to the more inclusive Ti commands. Arguing about calculator is really stupid. Use whatever feels best for your needs and go post constructive posts on the appropriate ng. I know the HP49G has a higher success rate of integrating exponential expressions than trigonmetric ones. I apply tricks too in solving math. I posted those questions just compare the INTVX command with the INTEGRATE command--simple. You can prove and say whatever you want. I don't have time for trivial and unnecessary things like doing a time comparison and posting more messages to this thread. There are more important things to do with my time. Besides, I don't use calculators that much. ==== The correction to LAP(1-COS(X))/t) is LAP(1-COS(X))/X) also add LAP(1-COS(X))/X^2) INTVX(COS(LOG(X))/X) was not meant to be there, sorry! INTVX(X^3/(SQRT(X^8+1)))