A46 SuperMarioHP is a nice-looking Super Mario clone for the 49G, but it runs too fast to be playable. Has anyone managed to slow it down, or has ideas on how I could do so? Great! Now this last mystery seems solved for me. -- And that was my understanding and also why I tried several times to make the distinction between the two (field and *single nibble field*). A field is obviously extremely useful, but it also exists in the 68K (B,W and L) Jean-Yves Come on, it's not that bad. And I understand that if I've upset some people that they are angry (I get angry very quickly too). And in fact, it's really one person George Tsiros from whom I know nothing except that he's posted nothing useful since his first post earlier this month... Probably a very young person.. Jean-Yves Is the production of the HP calcs officially over by now? I thought that only further development has stopped, but that production of the calcs that exist today still would continue. Any information on this? Well, I'm glad I got my 49G last week :) This price is about right. We sell the 49G and the RRP in Australian dollars is $302.50. I looked in a couple of calculator stores in There is not a huge margin these calcs but I think $400+ dollars is over the top. Unless this was in a pack including cables and cards etc.? === Subject: Re: HP48GX Discontinued Here are the products that seem to be going end of line PRETTY SOON from a canadian supplier perspective. 20S ALGEBRAIC SCIENTIFICCALCULATOR *NRA 11/30/01* 48GX GRAPHIC EXPANDABLECALCULATOR *11/30/01* ENHANCED INFRARED PRINTER FORHP CALCULATORS *SUPL ISSUE* 17BII FINANCIAL BUSINESS ..CALCULATOR FR *NRA 11/30/01* 20S FR ALGEBRAIC SCIENTIFICCALCULATOR *NRA 11/30/01* The link : http://www.oxycom.com/canada/searchresult.asp?8249272xxqfg418=926341125 Alain iJVL7.3545$Kc2.319825@newsread1.prod.itd.earthlink.net... errors It seems that we have been lied to! . Oh, Come on... Don't you remember the statement posted last week from Ian Morris ? The business will continue just as before and as usual... How could you ever doubt that ?! Come on, Ok, I will just post an extract again to remove any doubts: We expect that the calculator line will continue as a good business for HP for a long time Jean-Yves a 11/30/01* 48GX press :-( (I hope I'm all wrong folks!) Mark from 11/30/01* 11/30/01* from Italy are NK Ahem, shouldn?t I be included in the Club as well? Thessaloniki, Greece, 40¡37.93'N - 22¡57.07'E Yup. I have a work issued 548. I love it. I could not live without its strengts and waknesses and it very much personal preference as to which brand you buy. The OS however is the same on all of them - Windows CE (version depends on the hardware/firmware.) First the weaknesses; - Not nearly enough mem (32k) but expandable. (I recently read about a 1G compact flash card) - Unstable. I am always resetting my 548. On average once a day which is woeful. Mst of my settings are saved though. What can you expect with MS windows? - Heavy. The 540 series are built like battle ships. - Expensive. Around the AU$1,000 which is about US$500. Now the strengths. - Tons of useful and brilliant software. Over and above the standard stuff such as MS Office, IE, media player, book reader etc. I usually add, - Bible reader - RPN calc - the emu 48 is not bad there are some other good calc progs as well - Acrobar reader - Avant Go - integrates with IE to download news pages etc. from the internet. Tons of good stuff. - Database tools to work with the Access databases. - Also a neat metronome with a insrtument tuner - Tons of games. - Extremely portable (not very comfortable in the pocket though!) - Synchronises with both my work and home PCs to get contacts, email (one PC only), calendar, tasks, files, and internet. I have a Jornada525 PocketPC.. a really fantastic machine. Having 256 colors is the cheapest PocketPC, and runs really good. I've only reseted my unit sometimes, because I've installed un-stable soft (some emulators and similar) Its 133MHz SH3 processor is not as fast as StrongARM at 206MHz, but it does its job, and I can takes some notes, make some docs for my classes, have all my exam dates in pocket, do some excell Jean Yves for the soft), and other nice soft like GP-Pari. I would like to see a Giac/Xcas version for this machine, but it's not posible :-( And after buying a 128MB CompactFlash card, I can edit tex docs, have all my favourites StarWars trailers, play some games, develop my new website, and read Lord of the Rings, or The Hobbit,... and it's not full, so I could listen some mp3 while I go to school in my car (with an adaptor I have)... Ok, it's not all good: - I couldn't connect it with my hp48 or my hp49 yet - I haven't found any linux that could boot from a CompactFlash for its SH3 processor - I can't use the rotate mode from HP, I should buy a soft :-( - I couldn't find any CAS developed for PocketPC yet ;-( I hope this answer your question... J.Manrique CdU de la ETSIG Samson Cables (www.samsoncables.com) is pretty good for connectivity cables. www.samsoncables.com is also good for calculator availability. I've dealt with them several times, and they are very speedy and reliable. -e In my opinon the TI is a superior machine hardware wise ie the chip inside and screen and keys are better, but the HP has better software, both on it out of the box and the user programs from websites. -- You should quit while I'm ahead. Remarking my posting, Werner Huysegoms said: I may have chosen the wrong word to say both calculators accomplish the same task using different procedures. Minor differences shouldn't be credited only to the algorithm they used (There is more than a way do do multiplication, but they surely chose the best ones). Microprocessor resources may help, but the ingenuity of programmers has the final word. Anyway, I didn't mean to say this or that calculator was better. I myself don't have a TI calculator, as I have always preferred HP. By the way, bovine manure is just as offensive as bull shit, if you tried to avoid this expression. Next time, please say it's foolishness, I am naive, G. W. Barbosa everybody! I was wandering why when u want a trig equation u cannot specify the interval (for example 0, 2PI), and why the HP49 doesn't give all the solutions? By the way can somebody help me to understand these equations for example: cos(2x+PI/3)=-(SQR2)/2 on -PI;PI That was not the initial subject. It is what you would like it to be. As I prooved floor works as nasty for BCD as for binary depending on the memory data size and of course carefully choosen example. I don't understand why, you are straying off from the point that binary is simply more accurate approximation of real number and usually faster on most CPU's ? When performing addition, subtraction and multiplication binary is exactly as good as decimal or as any other system you choose except in binary machine BCD will simply faster run out of precision due to the fact, that binary mantissa of the same size simply will hold larger number. It is division that makes the difference. And altough division by powers of 10 prefers BCD, division by powers of 2 prefer binary and divisions by prime nubers make binary way more accurate approximation than BCD for the same mantissa size. The only machine that maintains precison for all division is machine that uses fractions, not floating point approximation, and even then it is only good for rational numbers. As soo as you introduce e, pi or myriad of roots, logrithms, powers, trigonometric etc. ENGINEERING and SCIENTIFIC functions that make BCD calculator loose precision faster than binary. The fact that humans use legs for moving from place to place doesn't mean that cars with wheels are somehow inferior to the cars with legs :-) It is yet to be seen succesfull implementation of the car with legs :-) Trying to do with machine exactly as human do is not necessairly the most intelligent approach. For the division by 262144 Borland doesn't need 80 bit floating point. It keeps accuracy even for float size (32 bit) data and operation. It is not about rounding off. Division by 262144 is perfectly represented in binary using just 18 bits of mantissa. Unfortunately it needs more than 48 bits for BCD. No. Just move to higher binaries, like 4194304 or 8589934592. Sooner or later you will run out of BCD accuracy for relatively small binary numbers that easily fit into 64 bit double. True. What I'm trying to say is that binary just have more accuracy in general for the same data size. Just don't give examples based on chaotic divergence. You assuming that the person, you are disputing with will buy such a simple tricky proove. Jack charset=iso-8859-1 Division by 10 in binary (and BCD); You just subtract 1 from the exponent. Steen Until you ask for division on a CPU that doesn't do BCD divisions in hardware. Then we are talking about the world of difference. This by the away is what makes conversion so long :-) But I admit, that display conversion is one of the few advantages of BCD coding. On the other hand limited length of the floating point as it is implemented in TI 89/92 and HP48/49 has nothing to do with display time. For those data types where floating point is roughly 64 bit long, display time should be negilgble. I dare to say that it would take far more time for HP48 to respond to key pressed than to convert binary 64 bit FP into text string :-) I don't think so. Both TI and HP are so SLOOW for what we see today on a handheld market. HP because of use of ancient hardware, TI partially because of BCD implemented on CPU that was not intended for BCD by any stretch of imagination. None of those machines is intended for finacial market. As in HP at least Saturn is optimized for BCD and justifies somewhat choosen coding, but in TI case it is nonsense. PR is of course different matter, no doubt about that :-) HP48 fails to perform 1 3 / 3 *. Does it mean it is dumb and should be ashamed of itself ? ;-) TI for that matter returns perfect 1.0. Does it means it is a beeter calculator than HP ? There are many simple cases when BCD or binary floating point will return 0.999999999.... instead of perfect 1.0 even when we know it should be 1.0. It is just a nature of approximation. Only demagogue would use then chaotic recursion or floor function to proove that the other coding generates so much error. Let me ask you a question. If I wouldn't tell you what coding is used, but would give you a choice of coding that is generally 1000 times more accurate than the other which one would you choose ? What Microsoft has to do with it ? You meant all programming language manufacturers, and CPU manufacturers who generally either ignore BCD or implement only rudimentary support for it assuming, that if somebody needs special case financial accuracy, one will use proper tool for it like Mathematica for all rational numbers perfect coding or financial calculator or some obscure BCD library written by third party and bought per special request ? If we use analogy with car, do you really care what it does under the hood or you want just simply get from point A to point B reliably, comfortably and in an acceptable time for a acceptable price ? Is it really matter if it is steam, internal combustion, electric or nuclear engine assuming they all generate within 12 digits of accuracy the same amount of noise, pollution etc and give you similar parameters ? We don't, but if machines suppose to communicate with each other in some of the natural languages, it would be exactly the choice. Saves the bandwich. Why you think, distant planet probes code pictures and other scientific data in binary for each pixel instead of BCD ? Nobody forces the user of PC to read output directly from binary but in fact binary is the langauge of your own PC that is running this e-mail on your screen when you read it :-) IP, FP and FLOOR functions suppose to be used with understanding that floating point is simply an approximation of the real number. HP fails this simple case: 1 3 / 3 * FLOOR where my binary Borland C++ does not. You tell me if this means, BCD is just a piece of junk or the function is rather used with no deep understanding of floating point representation and really doesn't proove anything except that in some special cases one of the codings fails where the other does not ? If you do shave, round etc, than I dare to say BCD looses it's advantage over binary :-) It was always assumption that one uses only divisons by power of 10 so BCD is always perfect and binary suffers. As soon as you introduce divisiors that introduce numbers incapable to fit mantissa you are forced to do round off and shaveing. I would rather then have 3 or 4 extra decimal digits to start with. Sure. Jacek True binary uses exponent based on powers of 2 so simple subraction wouldn't do. It is possible to imagine binary mantissa married to exponent based on 10 but then you would have to be carefull about normalizing the number and generally it would loose some of the accuracy - speed advantages. Jack charset=iso-8859-1 No. A 10 bit binary exponent, where first bit is sign bit, could look like this: E0: 0000000000 (10 bits, designating exponent = +-512) E-3: 1000000011 E16: 0000010000 It's very easy to subtract 1 from the exponent; Steen There is nothing wrong with the coding you think about, but IT IS NOT a binary floating point as it is used by most CPU's, FPU's and programming languages. Traditional computer binary floating point has power of exponend based on 2 as it is described by IEEE standard. I agree, that it is VERY EASY to subtract 1 from exponent, but in binary floating point it means that you divide number by 2, NOT BY 10. Division by 10 was a subject of discussion. Dividing by 10 in BCD means only to subtract 1 from exponent, but in binary it means that you have to divide entire mantissa by 10 and adjust exponent accordingly to normalize the number. Of course the opposite is for division by 2 which favors binary. The issue disappears for the binary coding with exponenet based on power of 10, but it is not a binary coding used by today computers. Jack charset=iso-8859-1 I know, but I never stated that I used the IEEE standard? I merely code the mantissa and exponent binary instead of via BCD. Has nobody read my last 10 posts? IF one uses the IEEE standard 754, the point is moot. That standard merely has it better with division by 2, so you have just as many opposites :-) I know you know that of course. No, surely not in the FPUs. The coding I showed has many advantages over IEEE 754 - standardization is not one - but are the HP4x objects standard? :-) With one sign bit, 52 bit mantissa and 11 exponent bits (one for exponent sign), you can express numbers from 2^52*10^-1024 to 2^52*10^1024. This is a much larger range than IEEE 754, and it doesn't violate the decimal representation (hence keeping the one advantage of BCD). 2^52 yields at least 15 good significant mantissa digits (relative accuracy is linked with number of significant digits), while comparable BCD only yields 12. One of the reasons modern FPUs operate with 80 bit precission internally, is to compensate for the equivalent 64 bit loss when retuning from the FPU - when dealing with IEEE 754. Whatever - if we're not talking about the same thing, I might as well shut up :-) Steen ... I understand, that the normalization, you have choosen, is that there are no fractional digits in mantissa. Otherwise you need extra bits to store information about the decimal point location. With such coding the only problem I would see is implementing efficient addition and subtraction in order to match their exponents before addition or subtraction. In such case what you risk is the need to multiply mantissa by powers of 10 which is time consuming in comparison to bit shift. Another problem that requires division by 10 is to normalize the number after some operations. For example 200 can exist in such coding as 2E2, 20E1, 200E0 etc. Without normalization it is difficult to compare them directly. How would you implement comparison between 2E2 and 20E1 to find out that two those numbers are in fact equal ? I think, that you rather would like to made one FPU hardware that works for all kind of applications including one, that need and use Extended precision. For most programs double is of sufficient accuracy, then why waste memory for more precision especially when internally FPU helps with rounding off from 80 bit Extended. At least my Borland C+ Builder offers long double 80 bit extended precision to be used directly as data in programs to use entire accuracy of the FPU when needed. The software states, that using Extended yields more precision than double. == abc I found that in the new rom the larger your stack is the slower the memory access are. I'm guessing that this is a bug with the new rom, but it is easily avoidable by clearing the stack. Another option that I haven't tried yet is to store parts of the stack into memory and restore them when I need them again. when there are pieces of hardware that only work on Windows. Is it possible to flash the rom via kermit. I think I've seen a special linux flasher. Hi. I doubt it.... Jean-Yves Greek language again? And can calc-philia be cured by a calciater? ;-) Nick. Hey Nick, maybe it's just the coordinate system, you know, the angle of the POV or the azimut or something. The cortinate system I use is: 1.Take the cover from the calc 2.You can read the HP logo correctly? If so, turn it back! 3.Is the logo just like in a mirror? That's the angle! Well, there are 4 feet and a lot of keys... feet are easier to model 'cause there are just 4 of them and the case is considered rigid (Not so rigid for som guys out there, i mean, Julian's poppa?) lead, here's a simple program to implement the above: Would you be able to please supply a practical example of the above Jonathan. I have read through the docs, and they mean squat to me. Tim === Subject: BCD (was Re: Well, it's over then...Good Bye.) You undermine your own argument by your obvious cluelessness. The abbreviation BCD stands for Binary Coded Decimal. If you had a machine that used ten distinct signal levels to represent decimal digits, it would *NOT* be BCD, by any stretch of the imagination. You are right and I didn't mean it to be BCD. It would be native D - just decimal :-). Jack Is there a way to disable the Quit and Save dialog in emu48? Alternatively is there a patch that will update the emulators clock on a schedule? I am making critical time calculations and the very tardy emu48 clock is a problem. Please speak about the Emu48 version for the Mac. The dialog control and clock of the Win32 Emu48 version works quite perfect. The reason of the timer2 misbehavior in the Mac version is the optimization on Gray scale graphics on costs of the accuracy of timer2. Emu48 for Win32 has the opposite problem. The accurate timer2 here, prohibit a successful Gray scale emulation on the other side in the current implementation. Christoph Entity Christoph Giesselink spoke thus: Sorry, yes it is the Mac version I am using. I don't really understand, are you saying that the Quit and save dialog cannot be disabled AND/OR there is no way to update the clock? scale The Emu48 version for Win32 can permanently disable the Quit and save dialog and the end (Checkbox Automatically Save Files on Exit in the Settings... menu). Emu48 for Win32 has also a very precise timer2 emulation, so the (display) clock is precise too. The clock of the emulated calculator is synchronized with the PC clock at every startup of the emulator. So these problems, you're mention, are only valid for the Emu48 beta version for the Apple Mac. Christoph The version i have (0.6 beta) seems to synchronise the clock. What version are you running? No, since it doesn't have any way of saving preferences yet. You could always download the source and add that yourself. Or email the author and ask him to add that functionality. The file that contains the memory image of the calculator is the one you save when you select Save from the menu. Aliases to the file work fine for me. This newsgroup is the best place to talk about Emu48. Later... Jon jon AT purvis DOT co DOT nz That's *really* strange... I loaded Emu49Asc on my Emu49 emulator and stored it in a variable. Then loaded to the emulator an ascii program and converted it with success. OK. It worked. Then I played some more with other things. I also copied the flags settings from the real thing to the emulator. Then I loaded Emu49Asc and tried it again. It failed. The ascii program gets converted in an ascii text, with the << intact and no heading, and I find it on level *2*. On level 1 I find the the criptic x4d. What's going on? In the docimentation there's no mention of any special flag setting. Any idea? Paolo Cavallo I am a teacher at heart, and there are moments in the classroom when I can hardly hold the joy. P. J. Palmer, 1998 paolo.cavallo@iperbole.bologna.it http://www.alberghetti.it/paolo.cavallo/pc.htm EMERGENCY - UNO STRACCIO DI PACE - www.emergency.it settings What do you mean with program? The flag setting is a List of four hex secondary. I done a RCLF on Emu48 v1.29beta2 with HP49G ROM v1.19-6 and called the corresponding Emu49Asc version from the Emu48Asc v1.5 package. Everthing worked OK. So what do you try to convert? And what are your current flag settings? What emulator do you use? Emu49 is a DLL from the HP49 SDK. Do you really use this version? Christoph flag Carsten, 49 in a next step like Mika's entries list was for the 48! Luis. Undoubtly a very useful document for SysRPL programmers, although seemingly not yet in a final state. Chapter 9 contains a listing of undocumented pointers. A good idea would perhaps be to place it here in the NG so that more people can help to minimize its size. I'll just start with a few undocumented pointers concerning PPAR and graphics. Command Stack ---------------- Recalls XMAX from the PPAR list if existent. If not, the default PPAR is created in the current directory, with XMAX = 6.5 Puts the real argument as new value for XMAX in PPAR if existent. If not it is created as above. Similarily GETXMIN, GETYMIN, GETYMAX, PUTYMAX etc. By the way, there is a useful unsupported but stable pointer PTR 27AE9 which best is called 'PPAR because it quotes the name PPAR. All these commands are essentially used in the tool PAINT which sets PPAR in such a way, that menu option (X,Y) in PICURE shows the genuine pixel coordinates of the cursor in PICTURE, for any size of the PICT graphic, i.e., with (0,0) in the upper left corner. None of the builtin plot types can realize this. PAINT is very useful for programming games and animations. I guess that without this tool I would not have been able to program the complex screen of the ACC clock & calendar. Note thar PAINT will be assigned with ASMs from Keyman to leftshift-hold LeftArrow. Sets the scroll mode of PICTURE and is essentially the same as { } PVIEW, only that the latter cost 10.5 bytes, the SysRPL command only 2.5. Indeed, it was unclear to me already at the HP48 times why DOGRAPHIC didn't get an own name in UsrRPL. { } PVIEW has as much as nothing to do with the command {#x #y} PVIEW :-) Enough for today, Wolfgang === Subject: Re: help with 48 sx If you are in the US, you might want to call the nearest HP service centre and find out if the upgrade program is still in effect. If so, you can turn in your HP48SX for a GX at basically half price... and the difference is not much more than the minimum repair cost. If picked up a 48GX from HP Canada in Mississauga about a month ago for $130 Cdn. versus $270 Cdn. IIRC the minimum repair quotation was about $75Cdn... It's a great program. I have a sneaking suspicion that it is intended to clear out the stocks of 48GX and at the same time reduce their service requirements. They still make great machines.. My 48SX was 9 years old and going strong.. Geoff How do I install downloaded libraries ? I heard something about Port 0;1;2 .... Put the variable that contains the library on the stack ('mylibvar') Then do: DUP RCL SWAP PURGE P# STO where P# is the port in which you want to store the library (0, 1, 2) Then reboot with ON-C -Scott I have a *.lib-file and if I enter the filename (...lib) the result is just a string with some version information See http://www.hpcalc.org/install.php -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag Could be a bad or incompatible binary (e.g., if it shows something like HPHP48 on a 49G), wrong comm parameters, ...). How did you transfer the binary, and how exactly does the beginning of the string look like? -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag It looks the following: HPHP49..... then only ascii-art follows (it's a string) Hi Todd! Well, I guess I tend to assume RPN-Mode always. Must get used to the idea that this is not always true. ;-) But to tell you the truth, especially for such things like the programatical use of the solver and its menus, I really can't tell what the advantages of algebraic mode could be. Unfortunately the convenient shortcuts for storing values in vars are not available in menu 30 in algebraic mode. But you can type some value press the key STO and the menu key for the variable. I could't find any way for solving in this menu in algebraic mode, but since EQ doesn't have to contain an equation only, we can make the program a little better so that it also works in algebraic mode. First of all, the variable EQ can also have a menu definition, if we intent to solve an equation. Its contents can be: { equation { menu definition key F1 } {menu definiton key F2 } .... } The contents of the inner lists for tthe menu definitions are: { label { unshifted-action blue-shifted-action red-shifted-action } } So we can make a program that builds the menu and then takes us to menu 30. Unfortunately in algebraic mode the blue-shifted-action is always the one in blue letters above the soft keys (Y= , WIN etc), even if the menu definition says something different. So that in this case the soft keys are hard keys. (Which on my HP49G is always true, concerning the power that is needed to press them. ;-) ) The following program (for algebraic mode) does this. The character << STEQ(ANS(1)); FOR(I,1,NVARS) MLST+ NEXT; PURGE({'P3','P2','P1','MLBL','MLST','NVARS','VARLST'}); MENU(30) Unfortunately I couldn't reproduce the exact behavior of the solver but it is close. Perhaps someone with more knowledge of the algebraic mode can make it better. Store the program in SOLVEQ or some other variable. Enter your equation on stack, errh history level 1. It must be entered, because SOLVEQ(EQ) doesn't work. (Could anybody make it work?) Now press the menu key for SOLVEQ and wait until the solver menu appears. The variables are now in filled boxes, because we made the menu definition manually. Type a number and the ENTER. Press a variable menu key. The value is stored in this variable. Again, does anybody know how it could also work without the ENTER key. I mean simply type the value and press the menu key, without the need to enter? When you have stored values for all known variables, type some initial guess, and press ENTER. Then press red-shift and the menu key for the variable to solve for. After a while you have the solution on the bottom of history, and the solution is also stored in that variable. This menu definition mixes up the behavior of RPN and algebraic, but you get the idea. Even in EQ you can have your own specialized menu, tailored to best fit your needs. Note also that bottom of history for the HP49G is what has been entered most recently while past inputs and results are at the top of history. In a way this is inverse to our perception of time, isn't it? Perhaps this means that the HP49G developes from future to the past, with many unexpected features yet to be discovered. ;-) Or just use your HP49G with the screen downwards to get the right historical development? ;-) Or perhaps that's the ways LIFE is, errhh I meant of course LIFO. ;-) Nick. How to update the 'ROM' ? I downloaded a *.flash -file... I assume that you have a 49G. JYA's site has details: http://www.epita.fr/~avenar_j/hp/49.html -Scott http://www.hpcalc.org/hp49/pc/rom/ -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag Several months ago I built a data logger for HP48. Since there is no availability of A/D converter to RS232 I had to build a UART (a parallel to serial converter), and to use a parallel A/D. Today some of my friends (a TI89 user) showed me a very simple project (more so then mine) that does the same thing. The site address is http://www.dvz.fh-koeln.de/~b012414/osci89/osci.html I wonder how to control the recieve bit in the HP48 as an input bit without using the RS232 protocol, (the transmit bit controlled by the OPENIO and CLOSEIO comands. I would be very thankful if you could help me and I hope its not too difficult of a task for you. Tal. email: td@chem.ch.huji.ac.il Why can t my HP48G+ solve 2 _ | | ABS(x-1)^3 dx _ -1 ?? My peers' TI give the answer quite instantly Pied Is there a casio fx-9000G owner? i have a casio FX700P that i stol.. borrow from my brother (oops) he used to use it many years ago for engineering it does not use any power (magic)! i think the batteries have been changed only 3 times in 15 years with heavy use. it uses watch batteries. i used to play golf (golf program) with it. it has only 1430 bytes. it is a tough cookie Yes, I own an FX-9850G. However there are calculator reviews and comparisions on the site listed below. Excellent site. http://www.rskey.org/ I am not overly familiar with this calc, but will try to answer any additional questions you may have. Me got fx-9700GH. Replacing with a Hp49 now though. I would guess that there is, but I suspect that such a person would be more likely to read the alt.calculator.casio newsgroup than this one. I'm the owner of a 48GX SUPER. I mean MK, Erable, ALG48, Stat48Pro and other good programs... and best of all: my Personal Menu System ;-) that is the interface for access all above... Really, this machine makes my diary work more comfortable. Well, I don«t want to bore you, but there is something in my calc that I hate: the sloooooooow cam INFORM. Anybody konws a good INFORM replacer? (like the MK's CHOOSE2 replaces the original CHOOSE...) I want something easy to use in UserRPL. HP calculators feature an entirely compliant (and energy consuming) RS-232 interface which the TI89 lacks. Ignoring of course the obvious differences in connectors shape and pin count the TI89 needs a couple of passive components in order to make the voltages standard. You can however build a TI89 cable very easily. Look at http://www.ticalc.org/hardware/cables/ 9tgroa$49s$1@nntp9.atl.mindspring.net... Hi Todd! If the numeric solver for linear systems has to solve a system with more than one solutions, then it returns the minimum norm least square solution. (If I remember right.) It is numeric, so no general solutions. There is also the command LSQ, which can take the constans vector from stack level 2 and the coefficient matrix from stack level 1 and return this solution. If you want solutions in terms of some variable you can do that with LINSOLVE. With your example you could enter the vector [ x+y+z+w=0 x+w=0 x+2y+z=0 ] then enter the vector [ x y z ] (Note that variable w is not in this vector) and use LINSOLVE to get solutions in terms of w (which was not in the variables vector) Hope it helped a little. Nick. Here's what I did: First, I examined the library's visible commands and the x~ menu -- only difference is that the latter has BZC and BZCE. Since bringing up the library's menu is trivial, I assume that he wants BZC and BZCE: so I openned the library with OT49 and saw neither. I found them in bz after converting that to a string ( +) and using BZC POS, so I used Emacs to edit bz and didn't see BZC; it was in a NIBHEX, so I deleted everything about the NIBHEX and compiled the result. It's a compressed string :-/ Decompressing that and editing it gives you the code of BZC and BZCE, along with the code that makes a menu with them -- I extracted BZC and BZCE, defined them as variables, added them to $VISIBLE, and then replaced my copy of bzman still in flash with the new one. Then I ran for the bzman documentation, because I've no clue what these programs do =) They seem to work fine as I've extracted them. Having used the modified library some, I think now I'd rather remove the whole ~x library thing, since I never use it anyway. I always have 262MENU. A person knowing Emacs and some SysRPL could probably do all this in about two minutes. The ~ command in OT49 is self-decompressing. The NIBHEX is in fact a normal string, but MASD doesn't like it and returns the hex form. The rest is a decompressor (it uses =aBZU in ROM, so it's small enough). You can decompress the string, somewhere down towards the end there are a few commands which build the menu. If you separate these from the rest and execute them, you get a normal menu list. No magic at all =) Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman I embarassed myself recently in maths class when I, by compounding interest by-the-second, I got a number that beat a continual compounding. After speaking with the teacher (whose TI89 had no trouble), and evaluating the expression in pieces, we discovered a rounding error in one of the HP49G's results. So... how can I deal with this? I'm looking at LongFloat now on hpcalc.org, but other ideas would be nice: I don't want to have a similar error on a test. The TI 89 and 92+ calculate with 14 digit precision. The HP 48 and 49 calculate with 12 digit precision, but it internally uses 15 digits for the calculations and then rounds them. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman The TI-89 calculates with 16digit precision internally and rounds to 14. 3BFC06BD.76C3D040@iname.com... interest the HP49G's Can you show us this rounding error? Scott -- Scott Hemphill hemphill@alumni.caltech.edu This isn't flying. This is falling, with style. -- Buzz Lightyear m3wv0j4xoj.fsf@pearl.local... interest the HP49G's For example: 1 999999999 sqrt 999999998 sqrt - / in numeric mode: hp49g : 63291.1392405 hp49g (extended real): 63245.5064067697 ti89 : 63247.1064449 win98 calc : 63245.5531559334217176879455116388 (32 digit precision) in exact mode (hp49): 1: '1/(sqrt(999999999)-sqrt(999999998)) EVAL 1: 9*sqrt(12345679)+sqrt(999999998) 1: 63245.5531559 But this is not a rounding error, simply a rounding. This nice example posted by Aquilino Nicol207s shows very good some aspects inherent in Numerical Analysis. Literally every computing device comes into trouble with this kind of problem, where two large, neighboring reals are subracted one from the other, which leads to cancellation of digits. Long time ago there was a HP-15C Advanced Functions manual with an instructional appendix Accuracy of Numerical Calculations, a very good reading in this context. Both roots above are of magnitude 10^5 and differ only at the fifth place after decimal dot, i.e. 9 (nine) digits are cancelled when the roots are subtracted without care. This matches with the observation above: HP49G 12 digits precision - 9 digits cancellation = 3 trustworthy digits in answer, and so on. The only way is to avoid cancellation, which both TI89, and HP49 do, if they are given the chance to compute at first symbolically the result, which is on both calcs 9*sqrt(12345679)+sqrt(999999998). This result will be numerically computed to full machine accuracy. ------------------------------------------------------------------------- Ralf Fritzsch Bundesanstalt fuer Wasserbau Federal Waterways Engineering and Research Dienststelle Kueste Institute - Department Hamburg ------------------------------------------------------------------------- Unix _IS_ user friendly - it's just selective about who its friends are. ------------------------------------------------------------------------- Would the HP49G's cable used to connect the calculator to a PC work with TI-89? Did anybody try it? everybody, I'v just found this incredible emulators that I never thouhgt possible, but I'm convinced that those are the best. I've tried to install some games found in www.area48.com and It was imposible, I've also tried to Install some Libraries like Bible and Economics, found in Maragnon's Website, I've followed the instructions step by step but when I'd recalled the library using Right Arrow an 2 I've selected port 1 and press 0 then STO nothing happens even turning off and on the emulator. I'm new using HP49, not the same with HP48, please help. Oops, sorry, how dare I question that great JYA - the oracle from which all HP calculator knowledge floweth - for this can never come of anything : he is never wrong. Geez, take an arrogance pill and call me in the morning... ---------------------------------------------------------------------------- --- before replying. Of course :-) Just like Y2K created demand for retired programmers familiar with old mainframes and ancient programming languages :-) The only advantage of BCD (if we don't count speed of display) has over binary is, when you use strictly addition, subtraction, multiplication and division by carefully choosen divisors including 100 used in banking interest calculations :-) Why don't you try to divide BCD by 3, 7 or 11 where there is immediate loss of accuracy thousenfold bigger then in same size 64 bit binary ? If you really need a financial calculator that mainatins precision for strictly financial calculations which exclude division by most of numbers, but significantly LESS ACCRATE for scintific calculations why did you bought scientific calculator ? Really ? Why modern programming languages barely support it or not at all ? According to the Motorola documentation all formats (integer, single precision, BCD, etc.) are converted for actual operations to Extended Precision Real format which happens to be binary :-) I wouldn't call it strictly support unless Motorola is not precise with it's documentation. The most used C++ language is more or less standard. Borland C++ is considered one of the most ANSI compliant C++ implementations, also does not have native BCD format. As for BCD libraries, there are also Roman Numeral libraries :-) The fact, that the format is not built in simply states that it is rarely needed. Financial calculations require financial calculator. What I don't get, is why SCIENTIFIC, ENGINEERING calcuator is crippled in performance and has less accuracy to guarantee, where the only advantage of BCD is that after running financial calculations in a loop that lasts couple of months there will be inaccuracy at the single last of the 12 decimal digit of the outcome and that only if you avoid divisions by 3,7,11 etc. because in such case financial calculations would be immidietly ruined about 1000 times less accurate for BCD than the same size binary :-) You mean floor() ? The following Borland C++ 5.0 program: double a,b,c; a = 0.1; b = 0.9; c = floor(a+b); displayed as expected 1 Borland C++ uses binary math for double :-) It also does work OK on my VC++. I guess you meant power by symbol _. Sorry, but power is implemented in caluclators via logarithms, so BCD will be 1000 times less accurate than binary for 64 bit manitissa :-) If you will use multiplication on the other hand in a loop, again my old trusted Borland C++ 5.0 executes program as follows: double a,b,c; a = 0.1; c = a*a*a*a*a*a; and returns 1E-6. Where is the error ? Jack What you call cumbersome is on the other hand universal for shifts from 1 to 31 bits at the time :-) 64 bits prefers Saturn, but M68 can easily be extended from its native 32 bit to 64 or 96 or 128 bit :-) Saturn would die on such extensions. Jack And the mentioned web site also gives an example where 9 is divided by 10 in a loop. It suppose to expose binary error. The example states as: Decimal Binary 0.9 0.9 0.09 0.089999996 0.009 0.0090 0.0009 9.0E-4 0.00009 9.0E-5 0.000009 9.0E-6 9E-7 9.0000003E-7 9E-8 9.0E-8 9E-9 9.0E-9 9E-10 8.9999996E-10 My trusted old Borland C++ v 5.0 :-) using binary math clearly doesn't want to follow the example :-) The program as follows: double a = 9.0; int i; for(i=0; i<50; ++i) { a = a/10.0; } Generated output as: 0.9 0.09 0.009 0.0009 9E-5 9E-6 9E-7 9E-8 9E-9 9E-10 9E-11 9E-12 9E-13 9E-14 9E-15 9E-16 9E-17 9E-18 9E-19 9E-20 9E-21 9E-22 9E-23 9E-24 9E-25 9E-26 9E-27 9E-28 9E-29 9E-30 9E-31 9E-32 9E-33 9E-34 9E-35 9E-36 9E-37 9E-38 9E-39 9E-40 9E-41 9E-42 9E-43 9E-44 9E-45 9E-46 9E-47 9E-48 9E-49 9E-50 Hmmmm. Author was biased ? :-) Jack Hi Jean-Yves, hope all is well with you. I felt that that was 'stretching' it a bit. The operations that always work in hex mode are the exceptions, not the other way round. At least, that's the way I see it. Just a detail. On the BCD vs. Hex discussion: since with BCD you treat only the binary equivalent of 53 bits at a time, one can expect a %CH(53.,64.) = 20% slowdown, in the best of cases. Since some operations are far easier in hex (division by 2, for instance), it will be more than that. But I agree a calculator must work in BCD.. calculators, and computers in general, are there to cater to *our* needs, and not the other way round. Who cares about a bit of slowdown? We humans work with the decimal system, and it is only fitting that the tools we use mimick our behaviour - or try to, at least. After all, why would we need GUI's? Give us back the terminal interface! It's sooo much faster and less memory-hungry! The same discussion, albeit on a somewhat larger scale... Werner Huysegoms xwerner_huysegoms@freegates.bex (delete the leading and trailing x) that HP no longer rules? Are you sure you don't mean ACO? ... Bye. Jordi Hidalgo johil@tv3mail.com PS: I don't know why TI is so frequently referred here. I thought that everybody knew it sucks. Like Casio. C'mon folks, what's so wrong with you? Repeat after me: TI sucks, TI sucks, TI sucks. Every night before sleeping. And you'll join us. Then, we'll teach you RPL. ;-) I did thank you. On the other side, I hope your sleep will teach you something useful, it's been lacking lately Jean-Yves The condition is worse than I expected... ---------------------------------------------------------------------------- --- before replying. Well, the software of it is issue to personal prefs, but the hardware leaves the HP49G in the dust. So I prefere to repeat Both suck, both suck, both suck. But not before sleeping, 'cause I get nightmares. ;-) Nick. A bit ? Have you tried TI89/92 ? When you have CPU that can only do 8 bit addition and 8 bit subtraction on BCD, then clearly choosing BCD for that 32 bit binary CPU is not a smart choice. You are probably right in theory, but in practice how much are you willing to pay for BCD ? The reason I'm asking is because modern CPU's barely support BCD or not at all. If on the other hand you want to stick with outdated Saturn because it is so powerfull in BCD then your future might be as good as for HP49 :-( Wrong argument. The interface with the human is decimal and the internal format is generaly invisible to the user. In most cases you woudln't be able to tell the difference between binary and BCD. What I want from calculator is engineering and scientific accuracy and speed in programs. If I need accuracy for finacial calculations I will buy appropriate financial tool for it - FINANCIAL calculator :-) If you have color display it doesn't matter for you if it is multiple bit plane or multiple bit per pixel type of display. The machine takes care for you and display everything in the same color no matter what internal format is used for the color coding. To use effective BCD you would need also semiconductors that recognize at least 10 voltage levels. Why emulate BCD on a system that is by nature binary ? Terminal was still there underlying at least in home version of Windows until XP. It was called DOS :-) By the way. Your argument should be the same for computers. Why calculators suppose to work in BCD emulating the way humans think but computers are somehow excluded from that argument. What is exactly the difference here ? Jack No, _ = index, he gives a recursive definition of a sequence here. The example shows that small initial rounding errors can get worse to the point where the result is total garbage (a quick calculation shows that it takes some 40 iterations here until the 49G and IEEE 64 bit fp disagree on the 1st decimal digit). Of course, after a while *both* calculations give you total garbage here, but at least the 49G gives you the same nonsense you would get if you carried out the calculation on paper. :) -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag Piffle. All you need is a CPU with reasonable BCD instructions. Such as a vax. [ducks for cover] James Sorry. I misunderstood. Let's try. I don't have exactly 64 bit BCD. I'm going to use HP48 and TI89 instead for whatever internal precision it has. Step HP48 BCD Borland C++ binary TI 89 BCD 1 0.1 0.1 0.1 ... 10 0.961563495092 0.961563495113824 0.961563495114 ... 20 0.712123922 0.712123859197354 0.712123858479 ... 30 0.91224689536 0.912206704125676 0.912206244544 ... 40 0.007451550857 0.0251163632065463 0.02537723934 Well. TI89 has more digits, so it seems that Borland C++ binary is closer to actuall value value than HP48 :-) But I see your point. If wee choose point that is close to the fractal edge like Mandelbrot set for example, the sequence will diverge depending on the miniscule differences between codings to the completely different values very fast. The problem is, that in the above example as you mentioned all machines failed. You would need unlimited fractions like TI89 or HP49 to actually calculate it properly, and then it doesn't matter what coding you use for integer. Binary is probably faster to calculate but pain to display. Unfotunately fractions here are purely theoretical. Those fractions grow so fast, that none of those machines will calulate first 40 steps in a reasonable time. TI89/92 actually will run out of integer size somwhere around 10 step (they are not unlimited). HP49 might run out of memory and battery life. Jack It was just sarcastic ;-) BCD is microcoded software overlay on natively binary semiconductor :-) Anything emulated by definition is slower than native ;-) [Hunting season ?] Jack Hmm, the 49G gives me better results than the 48 (.02646... after 39th iteration which is still pretty good, given that the TI 89 has more digits). Agreed, in this example. But you can never be sure. Next example please. :) Yes, the number of decimal digits roughly doubles with each iteration. upto the 19th iteration (524288 digits ;-). That already took almost a iterations. Such are the joys of exponential complexity. :) Albert -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag On Thu, 22 Nov 2001 02:52:42 +0100, Albert Graef Well, it seems both of my examples were crap ;). The first one is just a blatant error (forgot about rounding). In the second one I assumed it would produce a considerable roundoff error since the recurrence relation I was using is very sensitive to initial conditions but alas not true in this case. Anyway, here are some examples that produce real results. Every one of the following Java code fragments produces erroneous output. Let's start with an extension of my first example: double x=0; for(int i=0; i<8; i++) x += 0.1; System.out.println(Math.floor(x*10)); This should produce 8 as output right? Wrong ;). It spits out a 7. Here is the simple HP48 program that produces the correct result : %%HP: T(3)A(R)F(.); << 0 1 8 START .1 + NEXT 10 * IP The next example is very dramatic in terms of the error: double x=0, y=0; for(int i=1;i<=20;i++) { y += 0.10; if( i % 10 == 0) x += 1/(1e-10 + y - Math.floor(y)); } System.out.println(x); The output is 9.999954764879936E9 which is in stark contrast to the correct output of 2E10 which is produced by the following analogous HP48 program : %%HP: T(3)A(R)F(.); << 0 0 1 20 FOR I .1 + IF I 10 MOD NOT THEN DUP FP .0000000001 + INV ROT + SWAP END NEXT DROP My final example uses only additions and multiplications ( no Math.floor() trickery ) : double x=1, y=0; for(int i=0; i<55; i++) { y += 0.1; x *= y/(i+1)*10*x; } System.out.println(x); By the time i hits 55 x is already equal to 2.122280026585516 which is in stark contrast to the correct value of 1.0. You can guess what happens to the value as the number of iterations increases past 55. :) Here is the HP48 program that produces the correct result : %%HP: T(3)A(R)F(.); << 1 0 1 55 FOR I .1 + DUP I / 10 * ROT DUP * * SWAP NEXT DROP These are only a sampling of an inexhaustible variety of examples where binary floating point fails on data with decimal fractions. Of course, each case could be tweaked in order to produce correct results but when you're working in an interactive environment such as a calculator where the user expects quick, correct answers, seamless interaction between everything and overall uniform behavior that's kind of hard. :) ---------------------------------------------------------------------------- --- before replying. Because it will be the second calculator regarding power, the first being the golden BORG-1000 by Rcobo, Steen and VPN. ;-) Talking about colors, I played with my (black) HP95LX after a long time (shame on me!!) and I saw again that it has a numerical equation solver and an RPN calculator implemented. The solver is much faster than the one of the 49. So why isn't it possible to implement the rest of the software of the HP49G for the HP95LX or similar hardware? Nick. *********************** TI has a Silver Edition for TI-83 :) its sound as a copy!! I Think! The TI has only diferent case! here the system is the same! Arghh... And best would be silver and grey legends, to stay compatible with the color scheme;-) I imagine it would look very cool since you have to wear sunglasses to withstand the reflections... Raymond 4 weeks ago I picked up a 49G from office depot. great deal at 135 bux. With the news that hp calcs are finished, today I decided to go back and pick up an extra, and perhaps a couple of other hp scientifics. guess what no more hp sci-calcs. just those damn biz calcs for biz types that are destroying all rationality for the sake of paper bux! be sure to see www.anti-matrix.net I want to but 3 to 6 HP cases for the Infrared printer HP 92169 F, G or H (Burgundy, Brown,Black). Prefer Burgundy and Brown colors. I really want them and will pay up to $46.00 each (last educalc price was $22.95) but they must be in like new condition. Jim Klein You could try the FormMaster. It's the fastest input form engine for the HP-48. But: it's not a direct replacement. You have to design most forms by yourself. A few forms are included as examples, amongst many other tools. The second fastest thing for the HP-48 is in the UISTUFF package. This one is a *real* replacement for the built-in input form engine. It snaps into place, beeing fully compatible to the built-in one. Of course it's not as fast as the FormMaster, but it's fast as can be while staying compatible. Check out the SpeedBrowser, *the* replacement for the built-in fullscreen browser. It offers many features like switchable font and choice between direction arrows and sliders, while staying compatible to the built-in browser. And if you want to design your own forms in the easiest possible way, you can try the graphical input form builder for the HP-48. All those packages can be downloaded from www.hpcalc.org Just search for 'Hellstern' as keyword;-) == abc to who X Xcatly !!! [VPN] (snip) (snip) Thanx! Just hope to helped some other users. I didn't know I can also post to http://move.to/hpkb Anyway you can download from www.hpcalc.org from section hp49docsmisc Kickaha from EQW having pretty print lim... on the screen. ENTER and up arrow EDIT ENTER You will get an error message Invalid Syntax. Would you please confirm this, too? ..Heiko Oh my God ! It full of BUGS !!! Veli-Pekka haven't charset=iso-8859-1 Confirmed. It's because of the equal sign in the command line. I was the first, again a long time ago, to report this too. It's very difficult to remedy, since it's embedded in the RPN parser, so it won't be fixed. Summation is a special case handled, and that won't happen for other commands using the equal sign. Steen other It is not difficult, it is a challenge ;-) difficulties. What about approx 0.-limits?: lim(sin(x)/x, x=0.) (does hang up) and lim(1/x, x=0.) (works fine) Who need to be challengend for this? I have posted it meanwhile ago, somebody confirmed... Should I add this to Erics bugzilla? ..Heiko PS: Limes is one of the very usefull differnces to HP48 :-) i«m sorry if this has been aswered before but does anyone know how to calculate a limit with two variables in the 49??? exhibits the same behavior. It doesn't respond to hitting the ON button intially, I have to hit it many many times (30+) before I can even get a response. It then reboots, and if I turn it off and then on again quickly it works fine, however, if I let it sit for a moment, or don't hit the ON button exactly right, I have to go through the whole process again. I'm assuming I damaged it somehow. Is there anything else I should/could try? The calc is no longer under warranty (though I don't think the warranty would cover accidental damage on my part anyways) - is it still worth sending it to HP to get it fixed? No to all of the above. The only thing I had in memory was the libraries. The calc crashed before and I lost all data I had on it. Chad Johnson ! In spite of the bad news, I'll keep on creating progs for my HP :). It's very useful, even at school !! Who would be interested in coding a program (ML and SysRPL) in order to reduce the size of the grobs by compressing them ? (B&W or greys) Something like HPJPEG :) HPThifu I can't load no sh.it into yorkeM. i do use Load in port 1 and on+c the calcm but the port is [sh.it]ed with garbage. load in stack does not work too. sorry for the bad words. -- ICQ: 127036017 && 126735906 mIRC on #Win9x na Brasnet e #Novocanal da Dalnet, al216m de irc.via-rs.com.br no canal #chatcity. AIm: Renan Birck ComVC:1127268 OK, I start scratching again. ;-) Yep, null, nema, nada, nichts, tipota! Here you got me! ;-) Message sent by SATURN ANIMAL with talking brain ? ;-) 4 { 2. 2. } The error message is ok, but the result is not: 1 2 Û 0h 4 { 2. 2. } This only happens with 2-dim arrays A bug ?? Veli-Pekka I have been searching through the www, a few docs and this discussion area however I can not find to much info. on the use of ALLOC. Does anyone where I can get some documentation for the use? also can you use it under HPTools? Timothy Ney What is 'ALLOC' ?? Maybe you mean GETTEMP or CREATETEMP ? Raymond On 15 Nov 2001 15:20:37 -0800, NeyT@dnr.qld.gov.au (Timothy Ney) ALLOC and its accompanying ABASE are pseudo-ops that help when you want to build data structures within assembly. ABASE expr sets the allocation counter to the value of the expression expr and each subsequent symbol ALLOC expr assigns the current value of this counter to the symbol symbol and then increments the counter by the value of expr. For example, consider the following data structure: ABASE 0 A ALLOC 1 B ALLOC 2 C ALLOC 3 D ALLOC 4 SIZE ALLOC 0 * Size of data structure Assuming the start address of an instance of this is in D0 then here is how you might access the variable C : AD0EX LC(5) C A=A+C A AD0EX or just D0=D0+ C if the value for the symbol C doesn't exceed 16. Note that neither changing the order of the variables or changing their size would affect the first of the above code snippets. The official documentation for these pseudo-ops is packaged with the latest version of the HP-Tools source code under the DOCS directory in the files DOCS and SASM.DOC. This can be found at : http://www.hpcalc.org/hp48/pc/programming/hptools-src-3.0.6.zip There really isn't anything more than what I said in the first paragraph though. Hope this helps... ---------------------------------------------------------------------------- --- before replying. With the HPTools v3.0, if C can be resolved locally, it can be any value (poisitive and negative). The HPTools will automatically split this instruction into several D0+16 , D0+16 etc... Jean-Yves HP's future is of no concern. ACO's is hardly questionable anymore. Yes, if your reasons for switching ever held water at all. The only thing that has changed is this: ACO will no longer be producing updates to the ROM. *Some* of us have done the intelligent thing in wishing well to good people in ACO, and then pondering the new possibilities (like a public ROM; a community effort to fix it and extend it; more sophisticated ROM delivery (would you like the Normal Version or the Stripped Version or the Extended Version or the...?)). There is nothing close. A laptop with a lot of software and work won't be as portable or as cheap or as easy to use; a palm pilot won't be as easy to use and likely not as fast at certain things -- also not as inexpensive. Some here are contemplating The Next Great Calculator and if you want to wait for that, you should just keep your HP48GX -- it's still good game; some people use it instead of the 49G; it's the next-best calculator in terms of handling its own feature set. Hi Julian! There is no (User RPL) method to do this. Perhaps the Sys RPL and ML gurus out there? Nick. P.S. Isn't it funny to see that we expect that the whole calculator working must be configurable? I mean even the last detail of operation. Funny the way we expect it to be able to add 1 to 1 =) Though I don't assume that pure UserRPL may be needed for every change. is there an easier way to enter complex expression in polar form or we have to enter each time : (z,/_r) by using characters for , and /_? === Subject: Re: Complex entry have Well, you can use a space character in place of the , character (saves you a keystroke). I'm not sure what you mean by /_ but you can enter the angle character (128d) by pressing RightShift SPC on the 48 series, or ALPHA RightShift 6 on the 49G. In a PC file, if you're using Kermit translate code 2 or 3, then character 128d can be entered as <) or 128. I really don't think that it's all that difficult. James ok, thank you! Does the () are realy needed? === Subject: Re: Complex entry You're welcome! Hmmm, for entering in polar form, I don't, offhand, see any easier way. As far as I know, the angle symbol is valid only in complex numbers, vectors, and character strings. values are in rectangular form and displays them in either polar or rectangular form depending on your mode setting. In any case, I don't see that as being easier than using the LeftShift () key. James James there is a easier way to enter expression in polar form or we have to enter each time : (z,/_r) by using characters for , and /_? I am in the polynomail solve matrix w/ polar coords turned on, coefficients: [1 0 0 -8] roots: [(2.,<0) (2.,<2.09) (2.,<-2.09)] The < symbol is the polar symbol I believe. I have my FIX set to 2 Rom 1.19-6 I want to be able to recieve an answer of in this form -1 +- radical 3 Yes, but only for numeric complex quantities, I think. Or am I wrong here? Greekkings, Nick. | Is there any way to create a custom menu that contains submenus? i think, my program is not the best one, but it works - i hope it helps: //////////////////////////////////////////// { DET { VEKTR Ç 'CST' RCL 'CSTT' STO { EGVL RREF TRAN { } { } { UP[CapitalCCedilla] CST MENU È } } MENU 'CSTT' RCL 'CST' STO È } { } CST } store this in 'CST' and type LS CUSTOM on the keyboard. //////////////////////////////////////////// in the main menu is DET, VEKTR and CST (for recall) if you press VEKTR it calls a program, which stores the acutal cst-menu in a temp-variable and creates a new menu (line 4, 5) and restores the initial CST variable. line 5 is a sub-program where you can go back into the main cst-menu. i'm sure, someone has a more elegant idea! greetings, harald I am not shure, but it should be possible with Wolfgangs OT49 from Eric's site. ..Heiko Instead of the saving/restoring, you can use TMENU which doesn't overwrite your CST variable. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman I have heard a lot about people decompiling sysRPL commands. IS there an easy way do to this? I believe it would be an excellent way of learning ML. On Tue, 6 Nov 2001 19:48:59 +0200, Veli-Pekka Nousiainen I wanna play too! ;) PS: How about a meeting on IRC? ---------------------------------------------------------------------------- ------ Remove the random permutation of HPRULES! from my e-mail address before replying. Yes, especially at http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/forum.cgi?read=12638 Need I say more? Nick. Carsten Dominik escribi227: You know how VPN's mind works... I guess he just saw L1791, which reminded him of another WR's library: APPSman, which used number 1791 in its first releases. Or maybe OT49 can give the same functionality as L1791? Running MacOS 9.1 in a G4 Cube: 1) The emulator wakes up at the clock time it was when it went to sleep. Can I make it update the clock without quitting and restarting the app? 2) When quitting there are 3 dialog boxes asking me stupid questions, can I disable this and set it to autosave? 3) Where is the actual file kept that contains the memory? Can I move it with an alias? 4) This emulator talk may not be appreciated by the hardware oriented types here, is there a NG, bulletin board or mailing list for this emulator? http://www.gnarlodious.com/ Poor calc, bring it into active life again, you hear? ;-) It could be done in such a way, that those things are not necessary and additional key presses are avoidable. But such things as unpacking and packing again in a different data structure are not what I meant. One example would be the COMB and PERM functions that didn't work properly in summs. This is corrected in 1.19-6 but it shows what I mean. COMB and PERM couldn't be properly used in summations. Other example, which gave reason to a long discussion: Store some value in your current variable VX and in the same directory try to solve some equation which doesn't need to contain the current variable. Or consider m rt%3D10%26hl%3Den%26group%3Dcomp.sys.hp48%26scoring%3Dd%26rnum%3D16%26selm%3 D There are many many such examples, which show that a complex system like a CAS in combination with a powerful OS has of course to provide a rich set of commands, but it is exactly this what makes it necessary to spend a long time in conceptual work, in order to remove such diffioculties right from the start. Of course it is still possible to do anything (inside the unavoidable frame of limitations), but for a beginner it is frustrating not to be able to achieve some solution only because of an unknown necessary step between the calculations. Add to this the incomplete documentation and you get the picture. If you already had a 48, then you can somehow manage to get an answer. If you start with the HP49G, then good luck. Well, one good reason for using a calc is that it should free me from all these things. If calculator usage demands such a knowledge, then I should rather take a UNIX box. (I don't mean math knowledge here) Oh no, no, no, dear, that's money making policy. I didn't buy the calc in order to go into the necessity of buying additional books. The documentation which enables us get the advantages of the calc, should be in the box. It is good to know that there are books about it, but it is a shame that a beginner *has* to buy them in order to get the power of the calc for his/her needs. And for such a complex system the docs should be at least a small encyclopedia. After all this, I should mention that I find the HP49G the best calculator regarding problem solving power. I only think that it is a mistake to keep its power hidden through a combination of not needed complications in its CAS-OS and very insufficient (and boring) documentation that comes with it. I think that this is a mistake, because it keeps potential new buyers away. Nick. X would X They are shit, sorry, sith lords now... VPN Just couldn't keepmy mouth shut... Or perhaps spies still working for the other side? ;-) Nick This says it all : http://home.velocitus.net/tandl/fiorinian_rhapsody.htm I'll try. :) I still have a SysRPL project lying around on my harddrive which waits to be completed. Yes, that's right, such things should work out of the box. But any system of such complexity will have rough edges, they get ironed out over time. Well, at least if development goes on, which is questionable right now, to say the least. :( Fortunately, Bernard Parisse has promised that he will continue fixing bugs in the CAS. Agreed, of course the manuals should have been at least as good as those for the 48 series. But this horse has been beaten to death so many times... Let's be glad that there are at least some useful books which you *can* buy. :) Albert -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag === Subject: Re: Evaluating ACO Poor Carly! Just imagine here crying all the way to the bank when she finds herself among the ranks of the unemployed. James comp.sys.hp48:142993 comp.sys.hp.misc:29132 comp.sys.handhelds:80361 misc.forsale.non-computer:108883 I've received some inquiries asking for more information on the condition of the unit. I've posted hi-res, closeup pictures of the unit at: http://homepage.mac.com/collin_ong/PhotoAlbum2.html The serial number is 3135S00289. The picture posted at the auction site is the actual unit and everything that is shown with it. The only thing not showing is the plastic packing container that goes inside the cardboard box. It holds the calculator, cover, manual, and pamphlets. This unit was my own and I was the only owner. I purchased it through the HP internal employee sales department when I worked there for a summer. I used the calculator when I was going through the EE program at UC Berkeley for about 2.5 years (during school months only). I always kept it in the case, but was still somewhat surprised to see how good the condition was when I inspected it. Although the push action is not as crisp as when new, none of the buttons are at all mushy or sticky and they all work well with good tactile feedback. If you are interested, please see the auction page: http://cgi.ebay.com/aw-cgi/eBayISAPI.dll?ViewItem&item=1296629239 Collin The site was closed a few weeks ago by multimania. I must find time to open another site. fxmit works on HP48 / HP49. midiplay crashs on HP49 since it does not include the last version of fxmit. You can mail for details. Sylvain Gamot charset=iso-8859-1 Try the Wayback Machine: http://web.archive.org/web/20010816011932/http://www.multimania.com/sgamot/h p48/ -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel Reason behind this: I'm currently teaching a course on MIDI programming, and I thought that I could as well try some of this stuff on my beloved 49G -- it's sad that it has been gathering a little dust lately. Even if the calc can only transmit at 9600 baud, that should be plenty for a single MIDI track without too many controller messages, useful enough for simple algorithmic composition and remote synth controlling. One problem is of course that you somehow have to convert those 9600 up to 31250 baud (MIDI standard). I'm currently investigating one of those least in theory, Midiman didn't reply to my questions yet.) The 2nd problem will be exact timing, without multithreading and timer signals this will be a little awkward. And the third problem is that I know nothing about accessing the serial interface, can this be done in SysRPL? IIRC I haven't seen this in Eduardo Kalinowski's books. kidding, all I'd actually need is a fast serial interface or USB. :) Albert -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag charset=iso-8859-1 See: gyrscebivqr.f1klwu0.pminews%40news.hamburg.pop.de Bye, Detlef -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel OK, I installed the patch and is IS working! THANKS! Also, is there a way to save object 'HOME'? Entity Jean-Yves Avenard spoke thus: Well, as an aspiring Immortal I would take cause with this statement. Exactly how much time is #FFFFFFFFFFFFh ??? Gnarlodious http://www.gnarlodious.com/ === Subject: Re: Getting emu48 running calculator happens as I'm in no hurry to go either, but I doubt that the universe allows immortality. Good luck, and at least Live long and prosper!. That's in ticks, of which there are 8192 per second, so my calc tells me 34359738367 seconds, or about 1089 years. I wonder what sort of calculator will be available then. But see my other post before you try #FFFFFFFFFFFFh James Wow! 1089 years! Wow! HP999999999999998GX: 1 TB Ram, 3000000 functions, 20000 graph types, 1000 ports, 300 programing languages. :) I find three are enough to keep me busy. (Actually, even only ML would suffice =) Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman And no manual. Sorry, couldn't resist. ;-) -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag 203-167-156-88.dialup.clear.net.nz If i've stopped one person from using PPC2002, then it was all worthwhile. ;) Would you be interested in a floating-point RPN calculator in that layout? I was thinking of making an HP-45 simulator for PalmOS (since i can't find an RPN calculator i really like), but i can do one in the Tofu48 layout if people think that would be more useful/fun? I might even do Tofu38 when i feel like writing an algebraic parser. :) Later... Jon jon AT purvis DOT co DOT nz problem with the printing on the IR printer. Is there some possibility how to pring plot od some function (i.e. SIN(X)) with switched axis? I mean this - clasically you have horizontal axis X, vertical axis Y. The sinus plot is displayed OK. BUT when I want to print it with Y as horizontal and X as vertical (so the picture rotated for 90 degrees clockwise), is it possible? John === Subject: Re: HP48GX and 82240B It seems to me that I saw a program for that at http://www.hpcalc.org/. James http://www.hpcalc.org/hp48/utils/misc/grobprnt.zip this is your file! Acrux, Italy No, I think the 49 doesn't do weighted stats. It must be programed for it should be very easy.) Nick. I'll just add the situation in OZ. TI's are offered to the students at a discount and recommended by teachers etc. But the schools also get credits for each calc sold that they can put towards either free calculators or additional teacher resources to support the TI calcs. What is really sad is that here we are, all agreeing that HP has been taken over by the Marketers and the engineering has been lost. What marketing has HP been doing? What is their target market? Students, No sign of marketing in this segment. Engineers/Professionals, they're lossing this market with the current offerings. world than by better organized(just that HP You are right saying that it doesn't make any sence. But the world is full of such examples that don't make sence. Unfortunately...:-( Nick. Now that is fanaticism! ;-) Well, the inferior product also takes emphasis over the excellent TI-89/92+, which are TI products. It might change, though... Bhuvanesh. News Flash, November 2095, Idaho wilderness. It was reported today that a grave was robbed last night. The ghoul admitted he was after an HP-48GX, which had been buried with its former owner. The calculator was in perfect condition, having been buried in a sealed plastic bag with the batteries removed. The ghoul was said to have had a large collection of HP calculators, K&E slide rules and Commodore 64 computers in his one room shack. :-) This kind of thing happens all the time in the supermarket business. I live in Cleveland TN now, but was born and raised and lived in Baltimore Maryland until I was thirty. Those folks from that area of the country probably all know about TastyKake pies,cakes, etc.( MY Favorite is Butterscotch Krimpets) Several years ago I saw them on my local grocery store shelves. I was really happy.( In heaven actually. Now if I could get a really good Philly Steak sub here or a good corned beef sandwich on Jewish Rye Bread The store carried them for a while and then they disappeared. I asked the store manager why they did not carry them any more, and was told that they had been told by the Little Debbie Company (Really big in this part of the country)that if they continued to carry TastyKake products, Little Debbie would no longer sell them their products. Sounds like extortion to me. I still see TastyKakes stuff from time to time,and I have lookouts that report to me if they see them in the local store.( We are all from DC,Maryland or Pennsylvania) Harold A. Climer Physics/Astronomy Lab Instructor U. Tennessee At Chattanooga === Subject: Re: HP calculators On Thu, 15 Nov 2001 17:37:31, Harold A. Climer A virtually perfect example of an extortive and monoply supporting 'refusal to supply'. Maybe you should complain to the Anti-Trust division at the local Justice Department Office... Of course, since they just got raped in public by Microsoft, they might not be too interested in doing anything.. Geoff ... OK, while I agree with most of your points, I think you're using a wierd kind of tribalism (TI vs HP) to obscure the very real and negative changes in HP is down... the only silver lining that I see here is that some in this newsgroup are taking the changes as stimulus toward the development of better calculators -- by themselves, and they certainly have more of an interest than HP seems to about quality. See Steen's recent postings. Oh, how utterly distusting! This is a very disturbing post, and I hope that, if it was made in humor, that its maker should suffer some kind of minor calamity in punishment! How dare someone ever suggest, for any reason, that a perfectly functional HP-48GX should be doomed to be put forever out of use; operationless until horribly obsolete, when nobody will ever play games, investigate maths, or The creator of this post should be ashamed! that he could even produce such words without a disclaimer or obcenity warning points to serious flaws in his own morality and personality. Like, go see a shrink -- and give your calculators to someone who can better care for them! As for me, I must spare a moment to weep for all the homeless and broken calculators, and the desolate never used; their purpose never met by interaction with an inquisitive human mind. The greatest joy to a calculator must be to be placed in the hands of an attentive hacker or a young child (who may be influenced into the former); the greatest pain... not destruction, but to gather dust -- to be ever ... I'm sorry, I cannot finish. I can nolongre see thescreen ROFL :) -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag Well, HP position in Australia is pretty-strong. The HP39G and the HP38G are the calculators recommended by the WA education. Jean-Yves put taken has could I'll been Thats good to hear. I have a 12 yo who will be needing a calculator next year. I've suggested getting him a HP but the wife, who works at the school, says No, the school is only recommending TI##. He'll be confused if he has a different calculator to everyone else. I've also demonstrated my HP49 to my Sister in law who's head of a Highschool Maths Dept and she was also firmly in the TI's camp. What would you recommend? I'm sure if I get him a HP he'll be the only kid in school with one. education. teachers support the ago bought it. that general the Years ago, HP meant High Price. However, while paying their High Price for a product, one could be assured that one was buying quality. Unfortunately, HP has lost their vision. Like many companies, they are on a quest for the Holy Grail, which is profits, no matter what the cost. Although many of us have denigrated TI's products, website, etc., at least their calculator division is not in the process of disbanding and throwing there highly skilled programmers and engineers out into the street. I, for one, am more than disenchanted with HP's cavalier attitude relative to their formally valued customers. Because of this, I will most likely not purchase another HP product. For those of you who are predisposed to masochism, continue purchasing HP products and you will not be disappointed. To those programmers and engineers who have endeavored to make the HP well-known Bronx cheer. Just my $0.02 Johnet123 I purchased a HP48G about 6 years ago. What a beautiful machine. The build quality was fantastic. Anyway, it's sat in the drawer for the past few years, and I pulled it out to find the serial port doesn't seem to work anymore - it fails a loopback test. Oh well, I had pretty good use from it, and everything else still works fine. As the HP48G+ is still available, and is pretty cheap compared to what I paid for the HP48G, I bought another one. Boy, has the build quality gone down! My original HP48G was made in Singapore. The new one is made in Indonesia. The plastic case is thinner, the battery cover is loose, and almost falls off. The paint on the keys has bubbles, and the right arrow key seems to have lost some paint in one corner. There's really no comparison between the two. I'm very disappointed in HP. They really aren't the company they once were.. :( Don The only difference between the HP48G made in Singapore and the one made in Indonesia is the keyboard.and screen. The keys are now painted rather than molded and the screen is now black instead of blue. However, everything else, it's the same production line, the same plastic, the same parts. So the plastic case is not thinner (exactly the same) neither the calculator. If you have bubbles on the pain (which I doubt) it's cover by the warranty. For your HP48GX, I doubt the serial port is broken after few unused years. Have you tried connecting the pine 2 and 3 together and then run the RS232 test ? Have you tried connecting it to your computer, I suspect more the cable than the calculator.. These beasts (HP4xG) never die Jean=Yves years, it everything aren't The 4X series isn't unique in never dying. I have a five HP calculators, dating back to a HP-65 (1975). They all still work fine. -Dave Senzig made in than Does this apply to all Indonesian units? My 48GX is an ID044 and has painted keys and a blue screen. My screen is blue - identical to my Singapore made 48G. And the plastic _seems_ thinner, certainly the battery cover. But I guess I could be imagining it, the paint issue certainly got me off on the wrong foot with my new purchase. :( There are bubbles - and the paint is thin on many keys. It's dead Jim. I shorted pins 2 & 3 together when I ran the loopback test. I think the fail code was 00020 (something like that.) And I can sit them side by side, plug the new one into my computer and download stuff to it. Then plug into the old one, and it doesn't work. But now that I have two, I can simply IR stuff from the new one, to the old. Don Hi Clemme! (Du sitzt in der Klemme ? ;-) ) limit value from positive or negative. You can calculate such limits using the function lim with the expression and an equation of the form 'x=value+0' or 'x=value-0' Example: To find the limit of exp(-inv(x-1)) for x approaching 1 from the positive direction, enter exp(-inv(x-1)) , then enter 'x=1+0' , then press lim. Example: To find the limit of exp(-inv(x-1)) for x approaching 1 from the negative direction, enter exp(-inv(x-1)) , then enter 'x=1-0' , then press lim. Hope it helped, Nick. BUT you just don't have to program it !!! I would use my HP 41CX... Veli-Pekka To use recursive descent compilers and EBNF syntax try a compiler- version is 1.51 ???. Excellent IMHO. Maybe the simplest way to go if you want to built your own translator. There are examples in using Coco/R for building various tools including translators and assemblers. Do that, you'll sxee what I mean then ;-) Well, I can operate it somewhat faster with my thumbs pressing the keys. It's the painted labels. It is OK if regarding proper workings of the calculator, but still I see the perfectly readable keys of the 41 and 48, which have some more years behind them, and wonder how the 49 will look like in, say, 5 years. Will there be any key with labels at all? X Nothing whatsoever on my HP 49G's the keys are in perfect condition I have tried several models and the very first batch seemed to have softer rubber and wearable paint. there are even more new chinese made models with less keyforce needed. After two years of usage I resemble now, when look at my arms, Robin Williams at Popeya. ;-) VPN, the keys are not so stiff, stop whining your wimps!! When I had downloaded the last file of hp49119-6.flash, a message had appeared in LCD, something like that: The system may not be installed. Please go to 'Download Pack' But when I had choosed [2]Download Pack, a message had displayed: ... fail. File is missing or bad CRC. What it means!? After that, my HP49 don't work fine. Stranges characters had displayed. in order to fix it, I had erased the ROM and the 0 User Bank. So, now the menu No system or +/ENTER pressed is displayed when I turn on the calculator. PLease, help me! What can i do to fix it!? Make sure you follow the instructions for flashing the rom exactly. It's different for the 1.19-6 rom than the earlier versions. There are also instructions in the same page on how to fix things if something screws up. http://www.hpcalc.org/viewzip.php?id=3240&file=49.html The calculator is still capable of downloading rom images even if you mess up, so don't panic :) And of course, make sure the cable is firmly connected the whole time you are transferring the rom. don't want random bits creeping in. I got the same message when downloading the ROM. Ignore it and reboot the calculator, in my case everything seems to work fine. In fact I did not even lose the calc's memory during the ROM upgrade. (I did check the ROM version on the calc to make sure it actually did download the ROM) Andre === Subject: Re: Problem when upgrading HP49 ROM It looks to me as if it didn't download correctly. Perhaps low batteries, or a bad connection or cable, or something got zapped by a cosmic ray.... Make sure you have fresh batteries in the calculator (but don't discard the partially used ones; save them for less critical usage), make sure the cable connections are good, and proceed to the If it fails: section of the instructions. If it fails again, then ask for help again. You're welcome, and your English is good enough to get the message through. James He's kidding! Th reason for giving such false statements must be too much of the ALG-Mode usage which is very similat to the TI 89 behaviour AND the 49G has even the ENTER key moved to the same position as it is on a TI 89. Veli-Pekka PS: Please stop teasing the newcomers!! Doh! My signature line was too far down on the message. It says in parentheses to question anything you hear from me. Sorry if I caused anyone to go out and buy a 49. I'll be more discreet in the future. -Al -- -Al Arduengo ------------ If it weren't for lack of context, there would be no news. ~Scott Adams -- -Al Arduengo ------------ A closed mouth gathers no feet. -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag I was thinking of writing some programs for Hydrogeollogy for the HP48 or HP49. I have already typed in the formulas for many equations from Fetters Hydroge9ology as well as from other texts and stored them as variables. However I do not think this is the best way to this. I would appreciat some suggestions for the HP community as to what would be the best format. Equation library format or programs(I am not much of an RPL programmer). What would be some of the tools I could use to make this job easier? I would like to sart this after the senester is over,during the Christmas/New Years break. Harold A. Climer Physics/Astronomy Lab Instructor U. Tennessee At Chattanooga There's eql49, which seems to provide a modular approach, i.e. it checks all libraries for additional equations (I've never ever checked, but it must be that way - you can choose which library you want). Maybe you could make such a library. eql49 would the provide browsing, MSLV access etc. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman I was wondering if anyone here has ported QPI from the 48 to the 49? I'm doing. Kevin Frandsen frandse@enel.ucalgary.ca charset=iso-8859-1 Have you checked out the command XQ? What is it that QPI does, and you can't do on the '49? Steen for the 49 either, so I had no hope. Kevin Frandsen enel.ucalgary.ca SSchmidt@nospam.dk says... I found that if I turned off many of the default simplication flag settings, symbolic matrix operations like inverse were significantly increased in speed. Sorry I don't have the HP49G handy or I would tell you the settings I am using. Maybe someone could do a little experiment to see which flag settings are optimal for matrix inverse, etc. On Tue, 6 Nov 2001 19:41:03 +0200, Veli-Pekka Nousiainen Ignore the GX in that thread. It would be compatible with the 49. :) I never said that an FPGA would be more efficient or consume less power than an ASIC. Anyone knows that's not true. :D I'm just saying that it wouldn't be inefficient or cost loads of money. By the way, I think the Xilinx people could give you a good talking to. ;) ---------------------------------------------------------------------------- ------ Remove the random permutation of HPRULES! from my e-mail address before replying. ---------------------------------------------------------------------------- ------ Remove the UH!ESPRL from my e-mail address before replying. I agree. Either hpcalc.org and NG Kickaha The commands are the same as with the HP49 Jean-Yves === Subject: 'TOFF' problem (Was: Re: Getting emu48 running) can't find comp.sys.hp48. 40ne ROM available in I agree that it might not make sense to patch the 49G ROM, but the original post mentioned the 48G. calculator happens as I was thinking that the value for 'TOFF' was limited to #FFFFFFFFh as the maximum, and I really do hope to survive that long. I tried #100000000h, and at first it looked like it would stay on indefinitely, but when it did turn off, trying to turn it back on resulted in the old instant OFF behavior, so I did a memory clear to get rid of 'TOFF'. Thinking that perhaps it was some sort of interaction with the system time, I checked the date and time, and found that it was about 25 days fast. Well, I don't usually have the time display on, so maybe it was just set wrong? I set the time and date correctly, and then stored #100000000h in 'TOFF' again; no problem, until I did a RightShift TIME, which brought back the instant OFF behavior. After another memory clear, I found that the clock was fast again. I cleared the memory and set the time and date again, and then stored #FFFFFFFFFFFFh in 'TOFF', and tried RightShift TIME; this didn't turn it off, but did set the time about 89 years fast. It seems that 'TOFF' still has some limitations, and I wouldn't advise using more than 32 bits for its value. Having written that, I would mention that #FFFFFFFFh ticks works out to over 145 hours, which seems plenty long enough for the calculator (would the batteries last that long?), and I don't think it's too bad even for an emulator. All of this was done using ROM 1.19-6. James I will hjave a look at this. Can you put it in bugs.hpcaalc.org ?? I thought I had fixed this bug in the previous ROM (it was reported by Joe Horn) Jean-Yves === Subject: Re: 'TOFF' problem (Was: Re: Getting emu48 running) Hi Jean-Yves, Done. Joe I don't know that Joe reported any problem with 'TOFF', but I reported (at 1.19-4?) a problem that when an object of the wrong type, 'TOFF' then the instant OFF behavior occurred. I don't know whether it corrupted the time system like this problem does. In that discussion, you mentioned that storing a value greater than 2^32 in TOFF could defeat the 5-second protection, which is why I questioned the #FFFFFFFFFFFFh value. Apparently you did indeed fix that bug, as storing that program in 'TOFF' with 1.19-6 results in the normal 5-minute automatic shutdown, as if 'TOFF' didn't exist. James of M68K. division commands. The register size is 64-bits, which translates directly to 16 nibbles or a sign bit and 15 digits of internal accuracy in DEC Mode In HEX Mode the registers translates to 64-bits of raw power !! very exotic Not so, when you're mainly dealing with numbers (even CAS uses long integers a lot, I think) Not much of help in symbolics, though. not as good, in my opinion. Not at all. It has main accumulater and some lesser ones plus a few registers. Nothing to compare to the 68K This makes it very hard to program while maintaining OS compatibility. The ACO team members were (are still) Jedi Programmers while I think myself as a Padawan. Veli-Pekka The software written for the Saturn line I mean the implemented algorithms are a master piece of art and use effectively the BCD arithmetics. The series chosen for trancendental functions (among others) have compensating error terms. That is, the erros don't accumulate so that a much better accuracy is maintained throughtout the calculations. Programming the XScale for an Xpander like product is very difficult because you have to write again from scratch all the mathematical algorithms involved. Naturally they are now binary based and you have to convert the answers to decimal and vice verse. Ofcourse the possibility to use LongFloats and the raw power of even a 50MHz XScale running native machine code will more than compensate but still doing for ecample matrix calculations accurately is difficult (at least to me). The computer algorithms are seldom direct translations of the math involved. That is, to maintain accuracy in a limited digit amchine you need to be careful in your choises of algorithms and implementation. Veli-Pekka A very worried XScale fellow. faster Ofcourse I can see your smiley there, but I think that Jean-Yves has had it! I can almost here him crying in the emptyness of the internet space-time continuum: Enought with the calclutators and the Company bean-counters :-) Veli-nesh Hmm... how did you know about my physics background? :-) Bhuvanesh. So ? Where is the problem except extra timing required ? The binary machine will always be more accurate than BCD for the same size of data. You seem to forget, that 64 bit binary number is about 2000 times more accurate than 64 bit BCD number. Any typical floating point calculations introduce round-off errors no matter what base is chosen to represent the number. It is ridiculous to give carefully chosen examples when BCD divides equally by the other number nut generates infinite fraction in binary. Those examples are nothing more, that propaganda :-) There are many examples where the opposite is true, and in general no matter what notation is chosen, there is always round-off error simply because as Cantor proved set of real number in infinitely times larger that set of integers, and computer floating point is in general integer coding of real number approximation :-) , but binary is simply couple of thousand times more accurate at the same data size, so round off error stack up way slower than for BCD of the same data size. You can compensate BCD by extending data size which of course is a waste of memory and CPU cycles to move and calculate data, but in effect equal binary accuracy. For mentione 64 bit binary mantissa BCD would require roughly 19 digits, that is 76 bits of data. That is why BCD is a dinosaur of the calculator era, when early calculator CPU's were so primitive, and memory so costly, that even conversion from binary to decimal and vice versa was an too expensive for the market. Jack I have one built in my Windows, made by Microsoft. It represents 0.1 with no problem and probably outsold all HP calculators altogether :-) I can demonstrate to you, that Visual C++ or Borland C++ using binary math have no problem with displaying 0.1 when they actually represent it internally as IEEE binary. It is a matter of carefully choosing round off when converting to and from binary to display the number. Binary is about 2000 times more accurate for double size data so round off during conversion can give up one decimal digit for display roundoff and be still 100 times more accurate. Jack with respect to the last two lines: May I suggest a brand name for the possible startup with the RealEngineeringAttitude[TM]: BD (or DB this time, as last time they flipped a coin, Bill got named first and Dave second). Could you, please, give some examples how many clock cycles it takes for Saturn to execute typical commands binary commands and the size of the operand ? Jack Interesting idea. Buy the basic machine and then shop for 'accessories'. I like it. If you do this and you're thinking of even making it accessible at the high school level and even in exams then make sure that you include a quick, easy and foolproof way to demo what modules are installed for an inspection. Otherwise there's no chance of getting it ok'd for use it in exams. Even with that it would be viewed with suspicion because of the power you're talking of giving it. (Not arguing against - just commenting). I agree that is what has happened over the last two years, but at least HP were still contributing by employing the folks at ACO to to a lot of the development. At this stage, HP are not prepared to do anything, and I suspect, bargaining on making money out of other people's efforts. Andre Claassens developing developing having A Very Good Idea !!! Agreed ! Will it have Complex Units then? (1.5_V, <) 45_o) + (-200_mV, <) 0.5_rad) Veli-Pekka PS: How about a new name: VP-58GX ;-) Can we Xpect a TI-guy like Bhuvanesh (reincarnation of Shiva?;) to join in? How about some 100,000.00 USD or 100.000,00 Euro Veli-Pekka Reday join in for some euros & a full time manager on tech design & programming & marketing 15 years of pure programming & analyses &cetera 5 years of marketing (SW+HW) + consulting (including some coding) Ready to start a neww calculator line charset=iso-8859-1 new Was that full time, for a hardware platform that you knew like your own back pocket, becasue you designed it yourself? If we should be realistic, let's be realistic. I'm still hoping a little that Mr. Parisse would consider his CAS for the new device. It could be fairly easy to port. Again; I believe you were very limited by the hardware, and you needed alot of guesswork, to use the existing ('48) software. HP didn't help you much, did they? How would you have fared if you could just do DrawScreen() in your code, and the complete display would be redrawn, automatically taking care of all global dependencies? I don't think it was ever that easy developing for the '49. Yes, but at that time much of the work done was pioneer work. Now, we just pick all the goods from the shelves, and implement well-known algorithms. I have implemented an embedded (i960/87C51) multitasking OS from scratch in three weeks. It's not that hard, when you just seperate OS tasks from application tasks. This need to be done anyway, since we may want to go the OO way, and we want a very modular design. Very. It may still just run out in the sand - nothing is decided yet, I'm just throwing ideas around...... I'm glad, and would be proud to receive your help Jean-Yves. Any opinion is appreciated. Steen charset=iso-8859-1 Hehe ;-) Isn't period used as thousand seperator in America? I'd write US$ 100000, or US$ 100k or... Where? No point really in prototyping 100 devices, right? I'll very soon have a price for the hardware minus housing & keyboard. I'll also have to sign an NDA with Intel :-( You must remember that what I am talking about is me starting a business of my own. Me and a couple of my friends will become partners, and leave our day-time jobs for this - we will then be on the handheld device production market, and not necessarily doing this only as a hobby. A contract between partners will solve such issues - I have had one business go down because of difference of opinions. Absolutely. Yes. I'd be glad if I could use an existing one, as my primary idea is a platform. If you extract the essence of my posts, you'll see that the device will have performance to boot, be portable (durable batteries, weight around 2-300 g), and have fast interfaces to the outside world. I'd then produce measurement & automation add-on tools for the FireWire port. As an example, one could imagine a 100 MHz AD-converter, pumping data through the FireWire connection to the device (with FireWire you can reserve bandwidth, to optimize for data collection for example). Then you could use the 240x128 or 320x240 display as an oscilloscope. If the XScale should take care of the entire bandwidth of one FireWire channel, it'd still have more than 4 Gb/s internal bandwith to serve a display and other controllers. Theoretically, you could order an expanded version with 8 FireWire channels, and still have enoght time to show all 8 channels on the display. How about that? With the built-in RTC, you could leave the device for a couple of days, letting it take care of sampling temperature or other things that rarely needs sampling. I'm talking about much more than a '49 on steroids. Steen Hmm... how did you guess that ;-) Why wouldn't I join? Sorry, I don't have $100,000 yet :-) Bhuvanesh. EXLR function returns a list of the L/R-sides. Just my 3._cEuro [VPN] all. Why don't you simply go and write to Carly ? I'm sure she's listening... carly_fiorina@hp.com or Iain Morris (EPS manager, who took the decision about ACO): iain_morris@hp.com Jean-Yves you would be surprise. A study was made a while ago about how much an unhappy customer cost (or better say, make a company lose) during his lifetime. With how he will influence others etc... it's huge ! (ten of thousands of dollar). So one calculator may be nothing, but the rest and all adding up, it is Jean-Yves I mailed them yesterday and got a reply from Iain Morris: --Begin Message-- We have not exited the calculator business! We are closing our faclity in Melbourne, Australia and will be transfering our support of the calculator line to other HP locations. We expect that the calculator line will interest in our products, we appreciate your business. Iain Morris President, EPS. --End Message-- -Scott So why shut down ACO? -Mike -- http://www.mschaef.com X touchscreen, With a keyboard overlay and a KEYMAN58 I'm sure that (almost) everyone could be pleased. I'd personally prefer to have a choise between both a vertical and a horizontal model AND both with 12-function keys for soft menus that means two rows for the vertical model. Even four colours (Br.) or colors (US) would be Xtremely helpful while plotting PLEASE someone: give us a plotting MARK in the 49G to help to see which curve is which?!! Veli-Pekka X The earlier mentioned capability to access a text file (eg. a string) as records eg. a NL of CR-NL might be a record separator then accessing a line directly anywhere in the string with a single pointer would make the life of a beginner programmer a whole lot easier. Look at the 41CX instruction set working on the port 1 Veli-Pekka habit. I understand now. This could be achieved by enhancing the string manipulating functions. Though it can be programed, it would be a nice thing to have. But so is life. We never have everything we want, because at the moment we have it, we start thinking about 10 other things that we would also like to have. Nick. As there is so much talk at the moment about new directions I'd like to ask: Just what is the perfect calculator each of us expects? If somebody uses this for market research then great I'll have a replacement for my 49 in a few years. If you assume it's going to be 2-3 years before a new calculator gets to the market we can assume that the PDA's will by then be fully functional windows devices. You'll able to run apps lick Mathcad, Maple etc with the only restrictions being screen real estate and hardware interface (touch screen limited keys. Any new calculator would have to compete with these and be sufficiently different in functionality to compete. So here is my stab at the basics: Size: Pocketable, smaller than HP49, size of HP42 if possible. The 48/49's are to big to be carried anywhere in your pocket. Whats acchievable will depend on interface and screen. One option could be pull out or flip out screen that retracted displays single line of result like a cheap calculator. Power: Uses standard batteries say 2xAAA. Batteries life of 2 months minimum. Screen: HP49 seems reasonable but higher resolution. 2 colour to keep cost and power down. User Interface: Has to use a keyboard for rapid input of calcs. RPN of course but Algebraic will be needed since RPN will have even less of a profile than now. 48 and 49 seem to have it about right for an advanced calculator. You can't get away from the need for a numeric keypad plus extra's for command functions, advanced features and menu. Touch screen just won't cut it as I don't want to stop to pull out a stylus in the middle of a calc. Memory: 1.5 Mb in the 49 seems to be heaps. With memory so cheap just give it access to memory flash cards or a heap of internal memory. CPU and OS Can't comment except it should be capable of emulating the HP48/49 functionality with improved speed. A dream would be to have the capability to run the same apps as your Desktop unit but the size and display of a calculator unit that I'm thinking of would make interacting with an App thats designed for a 15 and a mouse very arkward. Cost of calculator would be much greater. External I/O USB would seem to be the most common cable connection with plenty of speed for a calculator. Some kind of cable-less connection would be nice. abc Tried it. Seems the 39G program is the one to use. -- Thierry Morissette tmorissette@hotmail.com pMcG7.26384$gR5.1851456@weber.videotron.net... program the ... redesigned some powerful commands in the new LIBman on ftp://ftp.math.fu-berlin.de/pub/usr/raut/HP49/tools/ As is well-known, paging the (rightshift) LIB key is very slow if more than 15 libs live in the ports. Solving this and other problems is what the revised LIBman is aiming at. One can now sort the LIB menu and insert spaces in the menu (following a proposal of Kickaha), to make one page for game libs, one for math libs, etc. All attached libs are shown in a choose box with a rich 2-page application menu. One of its options is to list all commands of the selected library in a satelite choose box. If then the scanned command has a HELP option, one may consult the text like in the CAT box. Unfortunately, at present only Emacs and OT49 seem to provide HELPs on some of their commands although this can easily be should be provided to such lib-commands whose functionality is hard to remember and applicable in special context only, say. Have fun, Wolfgang REMARK. LIBman is very small (1.5 KB). But small does not mean poor. LIBman uses the system CHOOSE engine and the Virtual stack. Although only poorly documented, these tools belong, IMHO, to the most ingenious inventions of ACO. It will probably be very hard or even impossible to port LIBman to the 48. X X How about providing a (very) small example on coding a HELP ? :-D Veli-Pekka Veli, being too lazy to read docs may cause problems :-) If you want introduce HELPs in an existing lib, split it stack) and follow the instructions on the dialog box. However, that doesn't release you from the pain first to read carefully - Wolfgang Wolfgang Rautenberg schreef: Obviously you did not download LTools. Perhaps it is too large? -- This message was written with 100% recycled electrons Pivo Veli, on your request, I added a concrete example in OT49.txt on my site: Making HELP on the Keyman option IfE?P. In view of fact that this functionaliry was only very way to teach - theoria docent, exempla trahunt :-) - Wolfgang Veli-Pekka Pressing ALPHA followed by LS and letter does work here. Greets, Marco Thomas, Marco, I can confirm both methods: after pressing ALPHA ALPHA: LS hold letter or LS letter Locking small letters, with ALPHA ALPHA LS ALPHA is not working in CAT, but I think it is not needed. Just another improvement for me, in handling [49] ..Heiko RplExt 2.0 is a library on hpcalc that adds two perfectly useless (IMHO) and one interesting program: BigChoose. BigChoose is supposed to and does act exactly like CHOOSE (but the delay on accepting subsequent characters (to jump to a selection) appears to be less), except that it doesn't accept a number for where the starting cursor is, and it's bigger. So you can see more selections on the screen and more of them in BigChoose, so it does what it claims to pretty well. The only problem is that it's noticably (read: annoyingly) slower than CHOOSE, and it gets VERY (read: excruciatingly) slow on large inputs (57-sized list is pretty bad, but 1200+-sized is horrible) -- which may not be a problem. I don't want to use BigChoose in my interfaces because of its speed (and I don't yet need the 'Big' part of it), but since it's only 2KB I'm keeping it around for future use. Cyrille has provided me with this information, so I'd like to share it. I will post this message to hpcalc.org, too. If you put the !DBGINF directive into a MASD source, the assembler does not only generate your compiled object, but it also returns a string (on level 1) full of debug information. The structure of this string is as follows: 5 DOCSTR 5 Length 5 Number of links (source files) n*[ 2 Number of characters .. Name of link file ] 5 Number of symbols (labels and constants) n*[ 2 Number of characters .. Name of symbol 1 Type: 9=Label 2=Constant for labels: 5 Address of label for constants: 16 Value of constant ] n*[ 5 Offset in code (this list is sorted by offset) 2 Number of link this instruction comes from 5 Character offset in link where this instruction starts ] Notes: - If the source string is unnamed, i.e. in TEMPOB, the number of links is 00001 and the number of characters is 00, immediately followed by the symbol table. - The label symbol table is, according to Cyrille, supposed to be an *offset* table. However it seems that the current (1.19-6) MASD has got a bug which makes it put *addresses* into this table. The associations table correctly contains offsets. Any mistakes here are mine, not Cyrille's. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Salut, Jean-Yves! But the last vertical pixel line is truncated ! +oo or oo or -oo becomes v This certainly is a feature in the display routine, right? Try it out and you'll see. While waiting for the 1.19-7 (c) 2002 I'll try to edit the minifont as suggested. Veli-Pekka of X In a math class we did some statistics work related to exponential functions -- the teacher told everyone to get TI-83s (and I got one, but I used my 49G instead) and follow some rediculus instructions that let to each of us with points plotted (but the TI-83 people had little squares instead of their dots, and could see them on the plot better than I) and an exponential fit to the plot using some FlashApp stat program (but I used the 49G's stat functions). Actually, I think the TI-83 people had to go to a computer and do something that looked tedious to get their fit. When all was done, I had a function of the form A*EXP(B) and they all functions of the form D(E)^F -- which is both what the teacher wanted and what would be more useful for the question at hand. So I've two questions: 1) How can I convert A*EXP(B) to D(E)^F -- where D, E, and F are real numbers? 2) How can I otherwise produce a function of the latter form from a stat program on the 49G? TIA Hmmm... The 40G which I happen to have in my possession seem to have something similar (MARK in plots) I's prefer the Color CASIO solution EXP2POW ?? VPN use ScatterPlot, it's a great addition to the calc. I might be wrong, but it sounds to me like they did nothing more a power fitting and discarded lower order terms. You cannot fit the exponential with only one polynomial on all R, but you can do it with a certain degree of precision at a specified point, that's all. Julian and Marco: The expression on the right isn't a polynomial, its an exponential. When you have D * E^F, E is the base. I don't know about the HP-49 but the HP-48 readily fits statistical data to an exponential. For positive values of E and E' (another base), exponentials can readily be converted: they really all look the same regardless of base. They all have the same general shape. In general, we want A * E'^B = D * E^F. The more frequent problem is, given the function of F, D * E^F, to express it in terms of EXP; i.e. E'=e. However, your problem: given the function A*EXP(B), to express it in terms of E^F, where E is given, is tackled as follows: A = D = the value at B = 0. Dividing both sides by A=D and taking the natural logarithm of both sides, we get B*ln(E') = F*ln(E). For E'=e, ln(E')=ln(e)=1. So B=F*ln(E) giving us F = B/ln(E). Another way of expressing this is: A*EXP(B) = A*E^(B/ln(E)). Done. You see, the coefficient doesn't change but the power does. Hope that this helps. Dan automatic Just my 3._cEuro Veli-Pekka X-EXP32-SerialNo: 50000000 Make, A=D B=F, and E=2.7182... Sorry the joke. Pio ------------------------------------------------------------ Get your FREE web-based e-mail and newsgroup access at: http://MailAndNews.com Create a new mailbox, or access your existing IMAP4 or POP3 mailbox from anywhere with just a web browser. ------------------------------------------------------------ http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 80,000 Newsgroups - 16 Different Servers! =----- X-EXP32-SerialNo: 50000000 Make, A=D B=F, and E=2.7182... Sorry the joke. Pio ------------------------------------------------------------ Get your FREE web-based e-mail and newsgroup access at: http://MailAndNews.com Create a new mailbox, or access your existing IMAP4 or POP3 mailbox from anywhere with just a web browser. ------------------------------------------------------------ http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 80,000 Newsgroups - 16 Different Servers! =----- } } Lin Yes, that is the right name. SOS for the HP calcs. ;-) Only joking (but with a bitter taste, regarding that HP seems to abandon tha calculators :-( ) Nick. No, the XScale is build around an ARM core (ARM v.5TE instruction set). A Pentium III core would always consume MORE than 1.7W @ 1GHz... J232rn X Sure! When I was in the university I used some T or C shell for scripting and it was a lot of fun. Then I added some grep commands and finally went for awk I used to have a DOS-based awk when I owned a 200-LX I used ARexx a lot in my Amiga 3000 (and A1000) due to programs directly supporting it. Hah! The system was multitasking way ahead of OS/2 or NT. No need for total reboot and seldom even a windowing system reboot. New object-based data-types could be loaded from WWW and then you do need to restart a painting program in order to read in a new type of picture (PNG, better JPG datatype, etc.) I have totally forgotten the HyperScript of my emulated Mac. I wish JYA had introduced macros in the Filer. At least the MetaKernel FILER.PAR could be introduced... Veli-Pekka, slightly off-topic :) PS: The awk-programmable command into the 49G ??? I settle for mawk, but I still wait for emacs regexp VPN X X VPN Quotes this from the HP Museum Forum http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/forum.cgi?read=12666 Re: Message from Iain Morris, President, EPS - Hewlett-Packard, posted by me on 16 Nov 2001, 1:55 a.m. I cannot help myself being in silence. First, Doug fiercely exposes a lot of truth. Now, you (me?) extends this to fine questions: Have you actually used something like the HP48, HP49? I mean, really _used_ it?, and affirmatives like They (the calculators) were designed by people who used the product, who enjoyed using the product. They were designed by people who (I believe) looked like they enjoyed going to work everyday, who saw their job as _fun_ and not simply something that paid the bills! This, I believe was _once_ the culture within HP. . I know Iâm nothing but another user, but as a teacher, I would like very much my students having RPN calculators (if hewpack and reliable, better) so it would easy math understanding. At least the way I believe is best: the funny way. The tricky way. And showing them how values can be hidden in the stack for later access is amazing. Iâm still clapping form here. Veli-Pekka Nousiainen schreef: Do you need to use regexps in a program? I figure that it is used mostly with user interaction, and therefore in the CMD. -- This message was written with 100% recycled electrons Pivo the Haa! Pivo himself! Don't ask me why, just deliver and I'm happy!! Releave every feasible (to your opinion :)¬ as a UserRPL Command to us die-hards Veli-Pekka PS: I still use DOS & do batch programming (.BAT) Try *nix shell programming (bash, csh etc.). You'll like it. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman A TTRM usually doesn't clear port 1. You may have to pull the batteries you can also short the battery contacts to discharge the built-in capacitor). Then boot it again and see. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman These files are explained in detail far better than I could in the documents that come with HPTOOLS. But why not use MASD? It is much simpler for smaller programs, and the new version that comes with 1.19-6 version of the ROM is very good. If you hook this up with EMACS, you have an outstanding development tool on your HP49g. Dennis The most useful thing I found is that more complex DEFINE statements are allowed. For example the old DEFINE statement required: but Jazz syntax: (no quotes). This meant keeping 2 radically different sets of DEFINE statements for JAZZ and MASD when compiling versions of code for both 48 and 49. Now that syntax is simlar, you can more easily modify source code for use in both calcs. MASD also accepts more than one argument in the DEFINE statements: etc. MASD added support for EXTERNAL and ROMPTR2, plus other changes. Dennis I_hate_spam@dont.even.try a 216crit le 16/11/01 22:43[CapitalEHat]: I also have an HP40g: do you know where I can find the same help for that calculator? 1)100-120 KB 2)80-99 KB 3)60-79 KB 4)40-59 KB 5)20-39 KB 6)< than 19 KB Reply if u want to answer :p. mine is 2(80-99 kb). i have algb and erable(sometimes mierdable, FYI: [mierda]=[shit]). ps: do u have a good ti92 emulator? please mail me renanbirc(@)brasnet.org [remove the ()]. -- ICQ: 127036017 && 126735906 mIRC on #Win9x na Brasnet e #Novocanal da Dalnet, al216m de irc.via-rs.com.br canal #chatcity. AIm: Renan Birck ComVC:1127268 renanbirc[@]brasnet.org renanbrazil[@]yahoo.com remover os colchetes([]) Usually 6. Mine is a G+. I use Jazz (light) + Hptables + ...... xDD Search for vti (windows). Try www.ticalc.org Ricardo -- Linux Registered User: 202 170 Kernel 2.2.19 http://xie121.infovia.xtec.es/~rblasco/ Un222os, hermanos linuxeros 640 Kb de memoria son m207s que suficientes (IBM AT Designers - 1982) Bueno, espero que esto no aparezca en la versi227n final (Bill Gates en la presentaci227n de Windows Crash Debug '98) You may want to try: http://www.hpcalc.org/hp48/graphics/misc/grobflip.zip Robert Tiismus : : Well, to defend myself, what amount of use I have made of the CAS has : shown to me that I can perform most of the algebra I use faster than : keying it into the calculator and then waiting for the output. : However, if I were tackling advanced mathematics, I would probably : have to rely on the 49G. As an EE, I occasionally delve into a little : math, but most of my work is just brute force number crunching. And : as such, the 48G is easier for me to work with (old habit and large : enter key in the RIGHT place). For General Math, 49G is good, but for specialized math 48GX is best. 48GX does allow upgrad to 2 MB ram! it's neas a palm pc! This is an interesting distinction, but could you please define it? I've heard 4MB too. And it will survive crashes, battery failures & shake-outs ? I prefer my professional HP 49G with flash !! Veli-Pekka I can contribute to the debate if chines or indonesian calcs are different: I've had three ID933.. calculators. One had a horrible keyboard which was extremly hard to press. The others (all second hand) where so so. Then I had an ID946... (or something in that range) which had a better keyboard then the one with the horrible one, but maybe not better than the two used ID933... calcs. Now I've got a chinese CN130.... and I _love_ the keyboard -- at least compared to the ID machines. I still like the tactile feedback of the HP48 a bit better, but the chinese calc is fine to use. I think the main reason now that I still feel a tad more secure on the HP48 is that the keys are much higher and are tilted while the HP49 keys are flat and are not as high (I hope I use the right words in this context). To the display cover: the ID models all had the same (rainbow effect) display cover. The cover seems a bit different on the CN model, and I haven't seen the rainbow effect till now, but I didn't viewed the calc in polarized (like viewing it outside when the sun's in the right angle) light till now. So I'm not sure if that effect is cured. Another difference between the ID and CN models is the blue color. The color on the ID model is shiny, while the blue on the CN models is much more structured and the printings beside the keys are somewhat harder to read. But that's not too bad. I still wonder why the design wasn't made optimized to ease the reading of the imprint. But I'm finally happy that with the CN keyboard the calc is usable for me. I think otherwise I'd have sold my ID calculator and lost interest in the HP49. Now I guess you'll be reading the one or the other question here when I explore the (for me) new possibilities ;-) Peter -- Great HP48/49 links: http://www.hpcalc.org http://move.to/hpkb to find *old* postings search: X X What about the scratches in the new display cover? Do you need a light saber or a diamond drill? Try with your finger nail, please?! Veli-Pekka Judging from the number of posts asking about nth-roots on the TI-89/92+ discussion group, I believe you :) It's very surprising that so many students don't know the meaning of nth-roots. Antilogarithms, too... I don't know why they are called that (it just makes things unnecessarily confusing). And you have to know the math, of course. I got my TI-92 only after finishing calculus and introductory physics, and I think that was a good thing. I'm not that good at math, but at least I got a good conceptual understanding of the math I needed before I started using the calculator. Bhuvanesh. I missed out on them while I was there, I guess :( rcobo, are you familiar with them? Bhuvanesh. While it is true that the HP position in Australia is very strong (30-40% in W.A. for example) JYA is not quite correct to say that the HP39G & HP38G are recommended by the W.A. curriculum council. In fact the council simple has a list of permitted calculators of which they are part. What happened here was that when graphical calculators were suddenly permitted in 1996 the four companies HP, TI, Casio and Sharp all entered on a level playing field. Casio, at the time, were the only ones who had a low cost alternative (~AUD$70) with less abilities whereas the others were all fairly similar in price at around AUD$150. Most private schools, for whom expense was not quite so critical, saw that the HP was vastly superior because of the aplets available to enhance its ability. The government schools, where cost was the over-riding factor, went for the low cost Casio and hence automatically graduated to the higher cost CFX-9850GB PLUS when they needed more capabilities. In W.A. TI and Sharp are currently only small players in a market that is dominated by HP and Casio. This is not necessarily true in other states. For example in Victoria the HP38G was initially not approved because of fears about its ability to do algebraic manipulations. A year later they realised how stupid they'd been and approved it but in the meantime TI had largely established a stranglehold. In most cases what made the difference was the ability of the sales team during the initial period when schools were making a choice. In W.A. we had a guy called Bob Parker who was incredibly able and made all the right moves. In retrospect it is a pity that the HP39G was not available back in '96. If HP had been able to offer a high/low cost pairing in the HP39G/HP38G as Casio did then they would have captured the entire market without a doubt. Do more? Like what? Especially with C programming, the TI-89/92+ can now do anything the HP49G can (you have to download the programs, of course :) Bhuvanesh. Hmm... well I don't think I lack the understanding of math 8-) but IMO it is a good thing to already have the calculator, be it a HP or a TI. Solving equations by hand, for example, is only a waste of time - unless you need to, say, solve a logarithmic equation *while studying them*. In my experience, the better students will both understand the math and know how to let the calculator do all the hard work. The less advanced ones will get stuck with the math and not even *consider* using the calc. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman IMHO, We must teach math for brains not for fingers, and I guess that calculators could be useful just for math investigations. I can do limit, derivate or integrate in seconds using just a pencil, many others need to open the pouch turn on the HP/TI and type and analyse the results, that sometimes are one line confusing expressions. Marx Pio ------------------------------------------------------------ Get your FREE web-based e-mail and newsgroup access at: http://MailAndNews.com Create a new mailbox, or access your existing IMAP4 or POP3 mailbox from anywhere with just a web browser. ------------------------------------------------------------ Bhuvanesh a 216crit : I'm not sure every math command of the 49 has an equivalent on the 89/92. Of course it is virtually possible to program everything on the TI89/92 if it does not require too much memory but that does not mean someone did it. polynomial factorization for example. Do you know if there a program that can find Jordan cycles (generalized eigenvectors if you prefer) for non diagonalizable matrices? (that could interest some of my students who do not have an HP...) computers (calculators with CAS) can be excellent teaching tools. for example the book advanced engineering mathematics by Robert J. Lopez is excellent. it loads on the sofware Maple and you can follow all the steps from your computer. CAS allows students that are not very good with math techniques understand concepts that were prohibit to them (like magic skates). IMHO memorizing tables is a waste of time (learn just the few basic ones). one day, I will be able to put a chip in my head with all the tables and formulas he he (BORG). BTW last september somebody made a working interface to living tissue ;) yeap, i know them very well. my brain feeds on dark chocolate, but when none is around, i have to go to the machine (with tastykates pies: apple, lemom, etc...) i've never seen a butterscoth krimpet yet. anyway, it is all molecularly decomposed in energy that fuels (burns) in my body to support the momemtum of life. i like the fuel sweet ;) Yes, of course someone has to do it. Not yet... :) Generalized eigenvectors? Yes, I have a function called GenEigVc in my MathTools package: I really wish we could have long symbol names... Bhuvanesh. It actually doesn't seem so hard (well, compared to Risch-type integration). We already have the polynomial GCD. If someone could generate the Q-matrix, have a nullspace-finding program, and implement some modular arithmetic, I think it could be done. I actually haven't tried it yet. Bhuvanesh. There are some additionnal steps: * Hensel lift * recombination of modular factors to true factors I don't think it is easier than the Risch algorithm implementation. You can look e.g. at modfactor.cc/modpoly.cc in my giac package (I use Cantor-Zassenhauss instead of Berlekamp but that's not much harder). And fast recombination would require LLL... have now the same functionnalities as the 49 in linear algebra. Oh, is there a program that lets you do ML on these without restriction or license? X X What? Is there no way to have long symbolic names in the TI 89 ? I got rid of my 89 in the spring 2001 and remember this one. Sounds more like the HP 38/39/40 variables. Yack ! :-P Veli-Pekka Well, at least it is much better than the TI-83 series, where you are limited to single-character variables! Actually, you can return limited to 8 characters?) Bhuvanesh. Well, I program in C, not ML, but the license is needed only for developing Flash applications using the TI SDK. Bhuvanesh. Eight chars? What a relief - it's still usable! The TI-83 limit is horrendous! You mean that there is not even a number after the letter? Veli-Pekka Can you do it directly on your TI 89 ?? I only did some simple TI-BASIC programming... Veli-Pekka I would have to disagree. I have 4 HP48XX calcs. Two from Singapore and two from Indonesia. The plastic is a different color, the fit/finish of the case castings is lower on the Indonesia units. It is pretty easy to tell the difference when you put the units side by side. The battery covers on the Indonesia units don't fit as well: one is too loose and the other has to be forced on. Charles Perry P.E. in warranty. build anymore - gone As I was a TI83 user before coming to the hp49 (what a change!) I played several times a game named Plain Jump which is the conversion of babal for the 83 (also ported on the 86 but dunnot know about the 89) Bye, Vincent -- ----------------------------------------------- | Vincent Rubiolo | | 2nd year student at the school of engineering | | Ecole des Mines de Nantes (France) | ----------------------------------------------- I mean Have you ever had your HP understood by a Casio user (or whatever else calculator user)? Moreover a friend of mine has a TI89 and a so called babal on it...That s definitely the crappiest one i ve ever seen ! :) Pied 3BF378BC.34E5556@eleve.emn.fr... the I need help to crack the case on my 48. I susect a contact problem w/ the on it, but am about to use force just to know how. Everything else on the damn thing works. I need to try this before buying a new one. Please respond to hippyjoes@aol.com www.hpcalc.org Ricardo -- Linux Registered User: 202 170 Kernel 2.2.19 http://xie121.infovia.xtec.es/~rblasco/ Un222os, hermanos linuxeros 640 Kb de memoria son m207s que suficientes (IBM AT Designers - 1982) Bueno, espero que esto no aparezca en la versi227n final (Bill Gates en la presentaci227n de Windows Crash Debug '98) For Newbies a more direct URL: http://ca-on.hpcalc.org/hp48/docs/opening/ VPN X Hi people, I need to access the submenu that is presented after running ~. How do I do that? I'd like to be able to access the menu directly (that is, without being forced to compress a dummy string) TIA! Marco { :Laenge =: :Hoehendiff =: :Anzahl Pars =: { 1 0 } V } INPUT UPDIR It should display me Laenge, Hoehendiff, and Anzahl Pars and I should possible to enter a number. the problem is that the program always skips the second line (Hoehendiff). What is the proplem? === Subject: 'sin' function problem When I try to enter on my 48G+ : 'sin(90)', it tells me that sin is Invalid Syntax. But the sin key works well. How is that possible? Remember that virtually all command and function names are upper case. So it's 'SIN(90)'. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman I can write it without any problems, ande after pressing Eval key I get 1 which is correct (DEG MODE) Ricardo -- Linux Registered User: 202 170 Kernel 2.2.19 http://xie121.infovia.xtec.es/~rblasco/ Un222os, hermanos linuxeros 640 Kb de memoria son m207s que suficientes (IBM AT Designers - 1982) Bueno, espero que esto no aparezca en la versi227n final (Bill Gates en la presentaci227n de Windows Crash Debug '98) estoy testando la escrita de news. -- quit: `pirch`ato (pirch@bcduiyg.poa.terra.com.br) (dead socket) ICQ: 127036017 && 126735906 mIRC on #Win9x na Brasnet e #Novocanal da Dalnet, al216m de irc.via-rs.com.br no canal #chatcity. AIm: Renan Birck ComVC:1127268 G'day folks I have some small (~2-5k) gifs that I would like to transfer across to my 39g and be able to view. Can anybody suggest a program that will convert them or similar. I don't feel like typing them into notepad/sketch so I am hoping for an easier solution. Any hints/tips? cheers /Victor http://ca-on.hpcalc.org/details.php?id=3902 VPN So after I have converted my .gif to a .gro, I can just transfer it across and view it? cheers /Victor my AND with HP49G 1.19-6 you can display 4-level grays as well. Veli-Pekka X Hi JR, There are several options for programming in all the available languages and you should be able to find them at www.hpcalc.org in the PC tools and/or programming sections. Heck, you can use any text editor to program in USER RPL. As for what the languages are: 1. User RPL is the highest level language available on the calc. It provides necessary argument checking anf error detection to prevent (in most cases) crashes and data loss. User RPL is just the entire collection of all the built in commands on the calc. 2. SysRPL is System RPL. It is the set of commands that the user does not have direct access to using the calculator in a normal mode. There are many more SysRPL commands available than UserRPL but most of these commands require careful use as they do not provide automatic argument checks and error handling. You can speed up programs significantly using SysRPL and in many cases it will result in smaller code size than an equivalent UserRPL code block. SysRPL requires a compiler (available in PC executable form or several on board versions like Jazz) and a table of entry points. The table merely equates the tokens used in the source code like DUP#1+ or XYGROBDISP to their address locations in the ROM. In my opinion SysRPL is the most popular of the languages used by the regulars on this group. 3. There is no Flash Rom Assembly Language as you call it but there is Assembly Language. This is the lower level of the three we are talking about. It provides virtually direct control of every aspect of the hardware. With this control, however, comes the increased risk of crashes, data corruption etc since NO error handling o or argument checking of any kind is provided. The programmer must handle all aspects of the code insuring that possible errors are handled, argments are present when needed and memory is controlled and manipulated correctly. It is not nearly as readable as higher level languages and involves direct manipulation of the chips registers, memory contents, IO buffers etc. It also requires an assembler and linker which, as with SysRPL, is also available on the computer or the calc. Other options less popular are Machine code (hex code), an old language targetted at game design (forgot the name), and probably a few other unknowns. MASD, the programming environment found in the HP49 and the HP48 version of MetaKernel provides some slick high level methods of programming in assembly. Of course the UserRPL compiler is built in and requires no third party tools to use. You can debug UserRPL on the calc or you can use Emu48 to do it all then fetch the program from the emulator to your actual calc. The possibilities are numerous. Hang around here and see. Make sure you read everything you can in the programming section of hpcalc.org. Rgds, -Al -- -Al Arduengo ------------ The only `intuitive' interface is the nipple. After that, it's all learned. -- -Al Arduengo ------------ Real Programmers think better when playing Adventure or Rogue. Pardon? Since finite precision 2-ary cannot represent some finite decimal fractions, it cannot be always more accurate. If you mean something else then please explain. This issue pops up on usenet every so often. As long as ppl don't switch to thinking in 2-ary, calculators which don't perform proper decimal rounding soon end on the scrap heap. Of course bcd arithmetic is costly, but so it is if you have to perform decimal rounding after each elementary floating point operation. -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag 64 bit BCD is capable to represent 16 digit decimal number. 64 bit binary is capable to represent 19 digit decimal number. Therefore 64 bit binary is about 1000 times more accurate than the same size BCD coding. 64 bit is typical floating point size in computer science. This is well known fact in computer science. Of course, one can always find handful of biased examples where one coding is more accurate than the other. 64 BCD coding for example is inaccurate to represent 1/65536, where binary represent such precisely well within it's 64 bit range. For the fact any fraction, where denominator is power of 2 is more exactly represented by the binary coding (requires less bits) than by the BCD coding. But when you compare sqrt(2), 1/3, golden ratio, PI, e and similar everyday mathematical and scientific coefficients you will find out, that all those are 1000 times more accurate in binary that in BCD when one uses typical 64 bit data size for mantissa. The only way to achieve accurate answer is independent from coding and requires user to use infinite integers and fractions. But whenever one has to deal with irrational numbers binary is superior to BCD. This is simply nonsense. All calculators made by Visual C++ or Borland C++ compilers use IEEE binary floating point and I haven't seen a single complaint about incomplete representation. It is just a matter of proper programming. When the calculation easily leads to 0.1 after just a few simple calculations, binary calculator will always display 0.1 as good as binary, because internal representation can be easily rounded (and generally is) off for display purposes using last one or two digits not for display but for display round off after conversion. The answer still will be 10 times more accurate than BCD on the display due to the fact that using two digits from 19 for round off still leaves user with 17 digits for display, not 16. User will not even see, what coding was used. The chain of floating point operations that would eventually loose full two decimal digits of accuracy by round off error (before it is visible on the screen) would probably require loop of several thousand of multiplication's and divisions and you would have to be really biased to choose such, that BCD would preserve accuracy in the same time. In general such loop using non biased operations such as division by 5, multiplication by 7 etc. would quickly lead to loss of BCD accuracy much faster than binary. When chain of calculations is very long, and involves infinite decimal representations for intermediate calculations, BCD will end up with way less accurate answer than binary. How bad BCD can be you can easily find out on your HP48 or HP49: Execute 1 ENTER 3. / 3 * and you will find out, that your accurate BCD gave you answer in the form of 0.99999999... If you do exactly the same operations on Windows Microsoft binary calculator (the one installed by Windows), the same simple operation: 1 / 3. = * 3. gives you 1.0 as expected by first grade student in the primary US school. You tell me, what should be an answer, that even a child knows ? For me it seems that using some of the extra internal representation digits for display rounding is far superior to falsely assumed exact BCD. Really ? I just had to do one on BCD calc and it was one of the simplest and often used operations of dividing something equally between 3 people and then reassembling it back into original one piece. What can be simpler than that ? :-) Jack On the other hand, 64-bit floating-point binary cannot represent 0.1 exactly. So if you want to do exact arithmetic on decimal fractions, you need to do clever things on a binary machine that you don't need to do on a decimal machine. -- Andrew Koenig, ark@research.att.com, http://www.research.att.com/info/ark What is ACO? Australian Calculator Operation What is CAS? Computer Algebra System(am i correct :o?) : Is the 49 being discontinued? I think that the ansewr is a giga YES. : How about the HP 48? I just bought one from Cynox with 512 K of memory and love it. I thing that the 48 *is* discontinued :(. But it will have long live coz the programers and good programs. Any answers are : appreciated. : Alan There is nothing especially clever or mystic in using some of the digits to round off outcome for display purposes. Especially when you are 1000 times accurate to start with, you do have extra digits to spare. What is so special about 1/10 that makes it so different from 1/3 ? or from PI, e sqrt(2) etc. for that matter ? If you are really so attached to exact representation of as many as possible rational numbers, that you should use not modern computer binary and not european culture decimal but actually babilonian system based on 12 that is far superior in that sensed to decimal, simply because 12 divides without remainder by 2,3,4 and 6, where 10 is divided without reminder only by 2 and 5. Divider of 3 and 4 is far more often used in real life, than 5. FLOATING POINT IS AN APPROXIMATION of real number and that how it is defined even in your calculator manual !!!!! Use symbolic math if you want to preserve accuracy. Cardinality of the rational numbers set is denumerable, where rational numbers that cannot be expressed neither using binary nor decimal in finite form are non-denumerable - far more exceeding number of rational numbers. Therefore, binary by being about 1000 times more accurate is much better average approxiamtion of real numbers than BCD coding of the same memory footprint. But even for the rational numbers when denominator and numerator are not choosen by biased example binary will give you usually better approxiamtion in most cases simply because in general case even rational numbers cannot be expressed using limited number of digits of computer representation. Actually by having more digits to start with there will be entire universe of rational numbers that can be represented accurately with 19 digits of binary but cannot with 16 digits of BCD in 64 bit mantissa representation. Check you Windows calculator and you will find out that 0.1 is displayed as 0.1 despite being binary floating point internally, which you as an average user wouldn't even know, if I wouldn't tell you :-) You are welcome to find out example, that exposes supposed flaw in the binary representation within Microsoft Windows calculator. To the contrary, your HP48 or HP49 fails miserably to perform 1 3. / 3 * where Windows calculator gives you the proper answer. Jack This is amusing. Mathematically, *any* choice of number base is going to provide exact positional fractions only as far as the number's factors, and since there is an infinite number of prime numbers this choice will always be insufficient for full expression. Then again, *any* choice of number base is sufficient for all rational numbers, when you use two numbers in that base per number =) I think the Babylonian system was based on 60, anyway (hence the instances of 60 in your own culture). A duodecimal system is good for other reasons, and would've made arithmetic slightly easier; there are some websites devoted to it. You say 'therefore' here as if your first sentence leads logically to your second, but I do not understand how it does so. The non-denumerable numbers are not merely 'far more' than the denumerable numbers -- for a given base and floating point system they are infinite whereas the denumerable numbers are finite. The difference between .9999... and any given number of '9's as in .999 :-/ I don't think that this is a miserable failure. That depends on who you are talking to. :) High-precision decimal calculations are vital for a wide range of applications. Why do you think there are ANSI/IEEE standards for decimal floating point arithmetic? It's because even if your calculator/program/whatever can correctly *display* 0.1, the internal 2-ary representation is still *not* accurate. Such errors propagate and can lead to weird results during lengthy calculations. 'Nuff said. Take a look at, e.g., http://www2.hursley.ibm.com/decimal/ for more details. -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag what about? I think the 48GX might already have on in MetaKernal or SS. I and others are hoping that the 49G ROM will be GPL'd, and JYA said something promising about this, but I don't know how it's going. Or using a little intuition (with your excellent insight about the existence of inverses for certain classes of Diffy) and less than 1 minute of calculator processing time to create, try this if you dare (like I have nothing better to do :-) { 0 93,696,464,117,719,556,736,504,464,493,337,370,164,620,001,317,888,324,328,8 31,426,389,280,359,805,576,090,715,432,322,401,032,293,433,784,888,862,269,6 8 3,183,146,356,189,041,191,099,815,475,797,458,795,246,958,595,863,619,591,03 7 ,326 266,031,129,579,840,039,117,982,986,585,836,782,427,099,195,767,546,522,208, 751,439,008,958,272,281,968,974,886,776,191,752,249,840,279,043,533,267,279, 6 79,871,204,851,696,913,034,621,798,918,811,448,377,949,882,161,698,355,475,3 5 4,525 583,003,997,228,581,583,709,759,642,328,686,693,185,901,013,979,242,341,670, 425,054,866,261,259,414,975,114,576,055,707,329,801,134,879,806,327,863,506, 7 68,126,257,186,740,269,392,298,183,750,136,952,417,973,079,709,879,684,880,7 3 9,567 } It's easy to come up with much bigger ones, but for posting purposes I chose this one (big enough step size and easy enough to type on a computer). You mean that the 48/49 don't have ASCII files capabilities? Is it better for you to manually set pointers and get records instead of just using a string? (Which can be used as an ascii file much easier?) HP-IL is OK, but I don't mind if it is USB or something else. I think it is the availability of many devices that makes HP-IL so attractive. What if we had USB or simple serial devices? Nop! Graph many curves with different colors. Very good when graphing 10 curves simoultaneously. He he, so we would be asked if we have a driving licence for a track, when we want to buy the next calc. ;-) == abc Well, this is just for fun... I could tell you wich of the four feet of your calc' cover will fall apart first. Well now I cn't becuse the 'A' key of my keybord hs fllen prt just now (this is not a joke) Oh! it works again! Well (I swear that was not a joke), if you are left handed, the first one to fall apart will be the lower rigth, and if you are left handed, the lower left. Reasons? Nick, VPN, start posting! All 4 of mine fell off when I put it in my backpack for 1 week and moving around with it. Aaron I purchased my 49g 3 weeks ago, 2 rubber pads are missing from the cover. found one in my pocket, and was mystified till I realized what it was. Replaced it, and it is now missing, along with its brother. Well, my 49G cover has a missing foot on the top left -- but this foot was lost when my dad had the 49G. Maybe he's underhanded? Confirmed. Yet another one :-( Nick. try: while stack is in minifont and you will get something after the opening << ALSO try: +oo or -oo and it looks like v AND there is more... JYA promised to fix this, but forgot :-( VPN If you don't like how it looks, just modify the minifont. There's plenty of tools to do so Jean-Yves I love RPN and quality HP calculators. I've never owned a TI that lasted more than a year before it went from an expensive toy to a cheap piece of shitty plastic with the worlds poorest keypads (not counting the HP-31/32/33/34 switches which were on a par with TI's). I've been an HP user since 1977. The raw computing power in a Pocket PC (even my $199.00 Jornada 525) so overwhelms the processor speed and memory capability of an HP-48GX or an HP-49G that I don't see a real need to have a dedicated piece of hardware called a calculator in the future, though for sentimental reasons I'll continue to keep a 48GX at home, one in the office and back them up with my HP-15C, HP-42S, two HP-28Ss and when the enemy is finally at the gate and I'm too old to read the LCD screens, my HP-67 with its non-continuous memory and its bright little LEDs. After that, I'll fall back to my Post Versalog and my K&E Log Log Decitrig. :-) I expect that it will be a lot easier to have a killer calculator living next to WORD and EXCELL in a fast Pocket PC than in any other hardware form. This is not what I want, but it seems to make sense that it is the direction things are heading. This disassociation of hardware and calculator may make the calculator world a frightenly powerful place to live. Just my $0.02 Jim Klein X Only the Saturn CPU chip Yorke (along with the display pixel count) It's the software that counts. A new ROM 1.19-7 will appear AND in the meanwhile this group will find an answer withe or without HP's help !!! There will be new HP calculators from the new division in USA and the TI/HP are both good calculators. Suggestion: Kepp your HP 48GX (if you already own one) AND buy the 49G and you will have it all !!! (like I do) Just a follow up. I had a little trouble a few weeks ago accessing the Calcpro web site because the Calcpro web site had a file access problem. It is fixed now, I placed my order and recieved everything in a professional and timely manner. Jim Klein last night, while browsing through catalogs at work (i work at night) i found this deal: http://www.jameco.com/cgi-bin/ncommerce3/ExecMacro/Jameco/searchResult.d2w/r eport?sort=&search=183425&Go.x=33&Go.y=8 i told my boss about it, and he ordered 10. guess who is going to get one :) ? oops, this is the link http://www.jameco.com/cgi-bin/ncommerce3/ExecMacro/Jameco/searchResult.d2w/r eport?sort=&search=183425 says... One could use a key assignment best done with KEYMAN from Wolfgang Rautenberg and thus have an easy complex (pun not intended) entry shortcut VPN You dont need the , to separate magnitude and angle because the character /_ is used as separator. The /_ character is accessible through the key sequence [alpha] , [right-shift] , [6] . If you find this key sequence cumbersome, you can assign the character /_ to some key or make a custom menu that contains this character. Nick. I work with complex numbers alot in my study of AC circuits. My favorite way to enter them is to use two soft menus. The first is PRG, MODES, ANGLE, which allows me to toggle between degrees and radians, polar and rectangular. The second is MTH, CMPLX, R-C which removes the need to enter brackets or dividers or the angle sign. When the real number is at 0 degrees and the imaginary number 90 or -90 degrees, which is often the case, just put your real number on the command line, press SPC, put the imaginary on the command line (negative for -90 degrees), and instead of pushing ENTER push the R-C soft key. Voila, the coordinates are on the stack in polar form (if the calc is set for polar.) Ciao While I don't understand at all, I'll try anyway :) instead of using integers : [1 0 0 -8] which makes this a symbolic array use float, or 48-style decimal integers: [1. 0. 0. -8.] I hope this helped [VPN] Gecko/20011011 If you can use SysRPL then try: { { :: DirLabel: $ MATRX ; :: TakeOver LastMenuDef@ LastMenuRow@ { } InitMenu LastMenuRow! LastMenuDef! ; } { :: DirLabel: $ STAT ; :: TakeOver LastMenuDef@ LastMenuRow@ { YOUR SUBMENU DEFINITIONS } InitMenu LastMenuRow! LastMenuDef! ; } { :: DirLabel: $ FINAN ; :: TakeOver LastMenuDef@ LastMenuRow@ { YOUR SUBMENU DEFINITIONS } InitMenu LastMenuRow! LastMenuDef! ; } } The method is not invented by me but I cant find out who was the original author. Maybe it's HP :) I saved the program under Illegal name 'CST, and my CST contains this ID: 'CST. This way the program is protected from overwriting by commands like 'MENU'. Best wishes, -- Robert Tiismus Install the hyper-excellent Nosy library and start probing... VPN Hi 48g/gx, was decompiling sysRPL and disassembling ML code, producing a plaintext file easy to read. If I remember well, it was also able to decompile internal libraries & fonts. It was writen quick an dirty in C/X.Window, not designed for speed emulation put purely for decompilation. One strange bug I never understood, is that all was working well, except sin/cos graph ploting that was producing strange results... If someone wanna debug [;-)] Peharps 'cause I wasn't emulating the saturn's bugs [;-)] I'll try to find it back (should be somewhere on a DAT ;-)). If I find it, it'll be put on my web site (www.courbis.com) and I'll post an annouce here. BTW, if you wanna to lear ML for the 48 series (and the 28), my books are freely available as PDF files on the site quoted above. Sorry, most of them are writen in French ;-( : - HP48s/sx french version : http://www.courbis.com/voyage48s.html - HP48s/sx english version : http://www.courbis.com/hp48ml.html - HP48g/gx french version (no english translation sorry) : http://www.courbis.com/voyage48g.html And for oldies : HP28c/cs - french version : http://www.courbis.com/voyage28.html Hope this helps Paul Yes, you can use the library Nosy available at http://www.hpcalc.org/hp49/programming/misc/nosy40.zip Happy decompiling, Nick. double oops :) go here and search for: 183425 www.jameco.com says... All I need is the emacs incredible regexpr manipulating capabilities in UserRPL plus a HP-Way documentation with examples. *** Please, Carsten & Pivo *** Veli-Pekka PS: In Finland we have a saying: Better Bird in Basket than ten on the branch the translation is not very accurate, but in finnish language this is: Parempi Pyy Pivossa kuin kymmenen oksalla Parempi=Better Pyy is a bird unkown in english to me so I use Bird instead Pivo is actually something like posessing or in posses but it's old finnish dialect, which I can't translate kuin=than kymmenen=ten oksalla=on the branch (oksa=branch) So it's better to have a bird in Pivo eg. Peter (which is also my name - translated) than on the branch. Conclusion: Peter F. Geelhoed ! Go and eat a bird ! AND Since I can't translate Pyy a chistmas turkey will do ;-) [VPN] I'm sorry, I've entirely forgotten what HP-Way is. I have mawk and fine these pretty effective. (Though, I remember something about Carsten making a seperate regexp package...) If I understand, you want to connect your HP49 on a MIDI device. You can try this program to send data at 31250 baud : http://www.multimania.com/sgamot/hp48/download/fxmit_v1.2.zip And you can try this one to send midifile (format 0) : http://www.multimania.com/sgamot/hp48/download/midiplay_v0.7.zip (fxmit is required to use it) I know you found some messages telling these programs don't work. I think the best is to try. Midiplay has not been tested by me on a HP49 but I built a version for it. You can have a look and tell me if it works or not. Sylvain Gamot Albert -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag people, I own an HP49G (ROM 1.10) and I'm afreid that someone played with my calculator the other day when I had to leave my room. A program of mine for calculating AC ducts had disappeared and the calculator was terribly slow. I then found out with File Manager that Home Memory and port 1 were full of garbage. I purged normally the ones from Home Memory, but there seems to be no way to clean port 1. Port 1 is full of garbage (0 bytes free) and is full of external references of random (I think) characters. When I move data from this port1 to somewhere, the data is copied to the new port but it's not removed from port1. I can however 'work' normally with this data, putting in on the stack and so on. Please some-one help with this. I use the calculator only professionaly (the only game I use is the built in Tetris) and I am afraid that some stupid guy has linked it with another equipment and destroyed my port1. Tiago Tresoldi charset=iso-8859-1 Try the PINIT command for starters, or else you might have to do a memory want to try holding backspace while warmstarting). It's very difficult to destroy hardware with the use of software alone, so your calc is fine underneath the garbage. I suggest you update the ROM to v1.19.6, as it's much more stable than v1.10. Many bugs, also FILER ones, have been fixed since. Steen LOL ! VPN :: body program ; I'd make my own link between my mac and my HP48, but I don't know the connections between the serial port of the mac and the calculator. Does anyone have done this before? Arthur charset=Windows-1252 http://www.hpcalc.org/details.php?id=3517 -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel Whatever graphical graphical calculator you choose, it will be overkill for a 12 y.o kid. No matter which brand. You should get something like the HP30S, which is a scientific calculator Jean-Yves kid can marketing the their is math I HP new go for spent anyone the having compete really run. 15 use me, is to the be Southern to Computer and calculator archeologists will be very pleased to find such a beauty after some hundreds of years. Imagine the excitement to ;-) Take it with humor and see the advantages of this. Perhaps the companies forget about good code in a copule of years. Then discovering the burried HP48 will perhaps attract the quriosity and demonstrate that you don't have to get a 100GHz PC to do symbolic math. ;-) Nick (as always joking) Same with TI =) I still remember that geography test where we needed the cube root - as you may know, the 92+ does not have an XROOT key. The *whole* class asked how they could get it... It all depends on how good you know to use your calculator. You need to be able of taking advance of its features, and working around its flaws. That's the same, no matter which brand you choose. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman He certainly will be confused when the teacher starts spouting instructions (press THIS button, then THIS button, then THIS button -- without explaining what the purpose is or what operations are being done or...). Your son will certainly be *able* to make more use out of his calculator than most of the other people, and do OK if he's knowledgeable enough of the maths and the calc. Some stuff he can just hunt for a feature; sometimes he'll have to come home and ask you. Who cares what camp she's in? Just make sure that your son can follow what's going on, and show him how to do on his HP49G what the teacher ... wait a second -- he's *twelve*? What the hel-- oh. You're in au. I'm thinking of idiot American schools. OK: make sure that he can do what the teacher tells the other students to do, and give him enough to not have to come home and ask you instead of doing his work at school. OK. What's your point? Not knowing how strong the HP position is in AU I can't say for sure but if the school recommends and bases the class on a Ti then IMO I'd go with the Ti for a few reasons. You might check into what the high schools and such recommend for thier classes as well. First let me say I love my HP's and in my case I wouldn't trade them for the world but if I were starting a child out on a graphing calculator right now it would be a Ti. If the local schools recommend, push, and teach classses around the Ti bottom line is it will be easier for him to learn and use a Ti and it will certainly do everything the class requires, stock out of the box. The HP49 may need programs either coded or installed to do some of the things that are very easy with the Ti. Just a small example, in the HP's when doing multiple graphs there is no indicator as to which graph your working with, that can be quite confusing. Also many of the functions are totally different syntax in algebraic mode or RPN mode entry. To a 12 year old that could get confusing, heck it confuses me sometimes. IMO there should be a HP manual that covers EVERY function wtih a example for each in both algebraic and RPN. The advanced users guide is OK at best but lacks many examples. HP has made it clear calculators are far from their priority now. As I understand it there will be no more new calculator development for HP. The future of HP calculators itself is up in the air, will they still be in stores in 3 years? Will HP still be supporting them in 3 years? I'd like to hope so, but no one really knows. Is the HP a more powerful product than the Ti, right now I think so but in 5 years? no one knows, but unless HP changes it's position as I understand it the HP will get no better in 5 years, the Ti's certainly will. Aside that while I like using my 49G it's far from the quality level my 48GX is or my SX before that, or my 27S before that. In my classes now maybe 10% of the work is calculator work, and even with that small % the books are written to do problems and assist with Ti calcs. That % will no doubt only grow as time goes on. Heck at one point calculators might be all networked in a classroom to share examples and do quiz's etc. It's obvious Ti conciders the educational market a priority and will continue to advance it's prodcuts in the educational areas. That integration will only grow and the inconvience/penalty for using a non Ti calc will too. Using a HP he can bet he will have to look everything up for himself, his book, teacher, and classmates will be of little help. It also means demonstrations done on the calculator are pretty well useless to him. He may learn more about his calculator when he's done but dare I say it he wouldn't need to with a Ti. Some of the stuff is similar but some of it is also very different. He will have to spend more time outside of class asking you, or the www, or the newsgroup. Esp. with the horrible documention on the 49G. Also I think most kids can pick up a Ti and use it pretty quickly, the HP's are a little more complicated you might actuall have to RTFM where most Ti users I know have never opened their manual. It's easy for people to say get the HP when most of them took calculus and their higher maths, myself included, a graphing calc was probably not even required much less the class driven for a Ti. Also remember your asking a group that loves HP's how many will tell you to get a Ti, aside me *gasp*. Now days in these classes they use the graphic calcs every day and as I talk to some people I know in calc now they spend 10-15 minutes a day working with their Ti's and stepping through how to use them and move around graphs etc. etc. It's easy for me to say sure I had to learn to use my HP but I'd have had to learn a Ti myself too because the classes were not including teaching for a specific calculator at the time. It's not that way anymore. IMO you are at disadvantage if you do not have Ti in such a inviornment. That isn't to say that you won't learn more and may be better off learning all this stuff on your own, but it can be frustrating and time consuming. It's kinda like when windows people complain about unix being harder to use those unix people just say quit whinning and learn to use it I think that's kinda how the HP people are. The HP is more complicated to use but you can do more with it. I think right NOW a experienced user will like the HP over the Ti, but a 12 year old with 8-10 years of school left in a Ti dominated educational enviornment is probably better served by a Ti. Sadly even though I own a half dozen HP calculators and not a single Ti I would start him a Ti at first. If he later decides he wants to use a HP he'll have no problems selling the Ti, but finding a HP could be a challege in 5 years :) So the answer is of course to buy both and keep the HP for yourself untill he decides the Ti isn't enough calculator :) Todd On Fri, 16 Nov 2001 16:09:09 +1100, Stephen Well... I'm confussed: My 48GX (ID 93003807) is made in Indonesia, but the screen is blue. ÀAnyone knows why? ...and I hope, it never die. (Apologize for lenguage mistakes) :: CODE GOSBVL =SAVPTR * Save RPL pointers GOSUB ASCII NIBHEX 6464 ASCII C=RSTK D1=C LC(5) 00022 D=C A B=0 A LC(5) 00004 GOSBVL =$5x7 GOVLNG =GETPTRLOOP * return to System RPL ENDCODE ; Here is thedocumentation for the entry $5x7: ***************************************************************** ***************************************************************** ** *A Name: $5x7 - Write 5x7 pxl data for string at Ith char in Row ** ** Category: DSPUTL ** ** Abstract: Writes 5x7 bit patterns for given character string ** starting at specified address (S) offset by a character ** position index (I). Data for subsequent pixel rows of ** each character is offset by (W) nibs from the preceding. ** ** ** D0:S (Start of row) ** B[A]: I (Chr Pos'n (0 - ..)) ** C[A]: N (String Length) ** D[A]: W (Row Width in nibs) ** ** Exit: 5x7 Bit Pattern for string written. ** XM: Font Table Flag (0=5x7, 1=5x9) ** CC iff next chr is at even location. ** ** Alters: CPU - A[A],B[A],C[S,A]; D0,D1; P, CARRY, SB,XM ** ** Calls: PtBits(0), sub8/10(0), Set8/10(0) ** ** Stack Levels: 4 (RSTK=C GOSUB go?Covered CON(5) =cPt2Bits) ** ** Notes: ** (1). The 5x7 font for Chr Pixels is at =PIX5X7, 16 bytes/chr ** The 5x9 font for Chr Pixels is at =PIX5X9, 20 bytes/chr ** -------------------------------------------------------------- ** (2). Generally, the starting address S is the start of ** pixel data for a row of the display. I is an ** option base 0 index specifying where the pixel data for ** the 1st character will start in the given row. The ** address X where the 1st character will start is: ** ** X := S + 1.5*I (each char takes 6 bits = 1.5 nibs) ** ** The width W is the number of nibs that must be added to ** X to obtain the address where the 2nd row of pixel data ** for the 1st char should be written. ** -------------------------------------------------------------- ** (3). Applications - Writing string pixel data: ** ** a) Into a grob. ** ** Implied by (a) is substring replacement operations. ** ie; ability to replace a specified number of characters ** in a grob with others. ** -------------------------------------------------------------- ** D0:S (Start of row) ** B[A]: I (Chr Pos'n (0 - ..)) ** C[A]: N (String Length) ** D[A]: W (Row Width in nibs) As you can see, your WIDTH is incorrectly set, it should be zero if you want to display at the beginning of the screen (cf the documentation Jean-Yves I am new in 48G programing and expect the problem is vry siple. For example: << WHILE A REPEAT ....... % how to change value of A? % ....... END and it didnt work and I have no any idea how to do that. RAF 'A' first, then INCR, DECR, STO, STO+, STO-, STO*, STO/, SINV, SCONJ, .... VPN I had the same problem last year in the Signals and Systems class. Programmable calcs were banned because of the bode plot programs, but I had to deal with complex number manipulation, and the need for some memory variables (I had to do it with a casio fx-82). The only strictly non-prog calc is, I think, the TI-30xIIs, wich has five variables, two-line display an recall expressions. It costs about 10 ¥ (euros) here in Spain. Well, it's not RPN but there are no RPN non-prog calcs for sale in Spain (at least new) If you find another one, tell us, I'm on the same situation! NaN. Hey! Some years ago I studied that same matter and I could use my calculator... :-) Bye. Jordi Hidalgo johil@tv3mail.com people ! I need to know how to open the HP49, i want to paint de back in silver :-) Thx ! just take the blue rivet-heads out. they are in the battery compartment (use a drill) then, separate it in half Does anyone know of a way to expand a fraction like: (2*x+3)/(2+1)^2 to: (2/(x+1)) + (1/(x+1)^2) I have tried the usual alg function to no avail. Micah Try FDISTRIBUTE or PARTFRAC But there is no HW floating point support (yet) :-( PS: I will still bet on the XScale-50MHz as the future 3G-calc-OS running Virtual Platform And maybe even Vassilis Prevelakis vprev@vp.com.gr from http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/forum.cgi?read=12064 would like that name? [VPN] For the Xscale version (third generation) OS: Virtual Platform eg. VP-58GX operand ? http://ca-on.hpcalc.org/hp48/docs/programming/cycles.zip Note the nibbling of the 64-bit accumulators during a full-width operation. Also: No HW multiply, divide, etc.. [VPN] it sounds great! now, if i could put my maple7 (o mathematica) in it... may be, the next HP49G will come with a crusoe chip :) oh yeah, and running linux too BTW: HP49G still rules!!!!!!!!! when you take the thought of others, reality becomes ready made abc How do u exactly convert HP 48 programs to HP 49 format I am talking about the libraries (Oh boy, you saved two whole bytes! I'm sure your mother is proud... or is that two whole keypresses? I'd type the whole word if I were you; practice makes perfect! If you really really meant to just say u for stylistic reasons... ... well, you suck. Hyperreductionist English is evil!) Assuming that you really mean taking HP48 programs and changing them to work on the HP49: Well, one plausable but imperfect way is to write a program that converts all PTRs or GO* to an HP48G? entry point to an HP49G one. This is imperfect because it assumes that the programmer only used certain entry points -- which might not be true; there are valid reasons to only jump to a location after the typical one, say to avoid an unnecessary test or register juggle. When dealing with ML code you have the additional problem of assuming that everything in it is *code*, but that's almost certainly true. The best way I think is to sit down with a program and work your way through it -- which is tedious. If you know how to do this, and have a lot of free time, and really want a certain program: have a blast, and share the result, OK? No, you're talking about CODE objects or SysRPL programs. I tried several times to enter lim or li to get lim in CAT. I remember it worked some times? I am using ALPHA ALPHA LS ALPHA and it works fine to enter small letters, but not in CAT? What I am doing wrong in CAT. ..Heiko Don't know, but here's a workaround: ALPHA ALPHA then hold down LS and type the now lowercase lim. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Try this EQW lim (1/X) for X=0 SIMP UNDO watch the result! Do SIMP UNDO again ENTER and up arrow EDIT ENTER You will get an error message Invalid Syntax. Comments appriciated ..Heiko charset=iso-8859-1 I was probably the first to report this (a long time ago), but it haven't been fixed I see. A shame. Steen Hi together, I made first experiance with pretty print lim (LIMIT) 1. Acess via CAT is difficult, because of missing LIMIT and slow access to lim. How can the cursor step to lim, avoiding to use exessive down-arrow? LIMIT is still existing, but not in the CAT. Hint: it is fast accessible via CALC LIMIT lim 2. sin(X) X / 0. lim gives a lim error numeric input (0. as a real is an exception!)? 3. sin(X) X / 0 lim gives 1 (0 as a zint works) 4. Typing lim x=0 sin(x)/x in EQW and simplify gives 1 in exact mode and hangs up in approx mode 5. Undo exchanges x=0 with sin(x)/x? 6. Typing lim x=0 sin..., ENTER and up-arrow EDIT will not allow to enter again because the = gives a syntax error (= which is shown as arrow in pretty print)? 7. Exchanging = by an arrow will not give the expected result. 8. Wolfgangs 3tog from OT49 is not able to rebuild the pritty print from the stack. I think some reactions are not intended and hope this test contributes to an improvement for further version*s :-) Heiko When i plug the cable into the calculator, it's put dark the screen, and slowly start to vanish until the screen turn blank, the calculator is id938xxxxxx, i heard about a problem with it, what can i do, please help me. cause the calculator have not system and i can't download it into the calculator by this problem. if you plug the cable end to the calc with nothing else connected to the other end, then you probably have something wrong with your cable ( a short in the cable). if the screen fades away, it is because is pulling too much current ( a short) do you have something else connected to the other end of the cable? cristian amoretti says... Did you put the connector the right way up? I.e. with the HP-emblem on top when connected to the calculator. The connector isn't symmetrical, the rounded edges belong on the downside. Arnold. On Tue, 13 Nov 2001 14:36:35 -0300, cristian amoretti charset=iso-8859-1 Do you use a HP49G cable or another connector that fits (100LX for example)? Anyway, if you see the screen change color, it's time to pull the plug! Steen === Subject: Re: [49] calculator problem screen, is help the Are you sure you're using the correct cable? See Joe Horn's cable table at http://pweb.jps.net/~joehorn/cable.htm. James This is absurd. How many people have ever plugged their cable in the wrong way? Gee, it doesn't want to go in, I guess I'll just apply tremendous force instead of examining the cable! Gee, I have a cable here and I think I'll plug it into my expensive calculator without glancing first at the connector -- as if not one of the other cables I use need to be plugged in a certain way! Count me in, I got a nice black screen before unplugging it (in a new world record time :-) The amount of force wasn't that big, just slightly more that the right way up This is exactly what I did, I was downloading a lot of different proggies and I got careless with the plugging... Not that I was worried, because when I pugged in my F1015-80002 several times (applying 12V to my 49) it sustained no damage. (According to Joe it should have been deepfried with onions by now right?) Let me see.... 1. power outlet: 2 holes, doesn't matter how you plug it in, my hairdryer works either way 2. BNC's one hole, earth on rim. Never experienced any problems with them 3. Cable TV, I have some strange receptions on some channels but when I called them they told me that you had to pay for the Playboy channel..... 4. Flat cables, never manage to plug one in the wrong way 5. Telephone, don't work upside down but they do plug in 6. USB, won't plug in the wrong way So don't tell me I should have known better...... .....hmmmmm...... I think that the problem is that on your side of the earth everything is upside down :-) -- This message was written with 100% recycled electrons Pivo No, everything looks right-side-up here -- you must be looking at me wrong. Power plugs here now have the 'one prong is a little larger than the other' feature, and some of them have 'ground' prongs -- both can only be plugged in one way. All the cables and ribbons inside my computer have safeties so they can only be (easily!) plugged in one way -- though older ribbons don't have such. All my power cables have such. Heck... the serial plug on the other end of the connectivity cable needs to be plugged in a certain way! Anyway, I can see how you might have a problem if you're working fast and it doesn't take too much force to push the plug in the wrong way -- but I *am* dubious about people who, having done so, get all the way to usenet before checking their cable. =) Hi to the NG. I've updated my flags document to the revision 1.1 at hpcalc.org. The URL is: http://www.hpcalc.org/hp49/docs/misc/flags11.zip Carsten Dominik). Hope to hear your comments soon. Kickaha PS: I've uploaded also a new document regarding the Reserved Names. I've already some improvements in mind, but I'd like to have feedback on this also. It will be added in next days to the same section of hpcalc.org (49-docs-misc) One comment to fulfill your hope: Please add flag 60, too. (Unit type; SI, if clear, English, if set). Explanation in HP48G Advanced User's Reference Manual, page 3-64, under CONST) ------------------------------------------------------------------------- Ralf Fritzsch Bundesanstalt fuer Wasserbau Federal Waterways Engineering and Research Dienststelle Kueste Institute - Department Hamburg ------------------------------------------------------------------------- Unix _IS_ user friendly - it's just selective about who its friends are. ------------------------------------------------------------------------- Well, there is. Your units 'look' up and the enemies' look down, or the other way around, I can't remember. You were not hoping for a colour distinction, were you? Yeah, that's true. But they show up, anyway. That's the same in the PC game. War ain't easy, is it? I didn't know a word in French (except for the usual words you know 'parrot-like') and I managed to understand the game easily. I had one advantage, though: I had many times played Dune II in my PC. look here for a very good (but in French) tutorial : http://hp.crestey.com/tests.html Having some issues with my 49G, seems at random startups it does a warmboot. I read a thread on that awhile ago and can't remember if anyone came up with any good ideas. I don't loose my memory but it is really annoying loosing all my settings every day or so. It seems to have started to happen since I went to 1.19-6. I tried clearing all memory and removing all libraries and it still does it. It seems to happen only if the calculator has been off for awhile (say more than a couple hours). I checked the batteries and they are still good. However since no one else seems to be having the problem I don't think it is the new OS update. However maybe I'll go back to 1.19-5 and make sure. Todd (SNIP greetings) You are completely right! And this explain why I can't find lot of informations about this variable. Anyway, going here and there in the hidden directory (thanx to all that replied me giving hints in this direction or in any other one ^_^) I discovered the variable Alrams, which I think must refer to the scheduled alarms. I haven't the time to play around to understand its structure, but I don't think it will be a uncredible task... I hope to do this in next week: I'm a very busy person, don't you know?! :o) (SNIP about HP48 Insights from B Wickes) I agree! I avoided an headache thinking on it! Kickaha Oops, the program is in part two. I won't apologise since I'm sure you've had a good time reviewing part one. :-) Bye. Jordi Hidalgo HPCC member #1046 Argh! I think i hate RPL ! a friend of mine is using a ti 89 and i m running a 48g+ I wish i could access a variable as easily as he does... why can t i do << 'Phrase' sto Phrase Good Bye 'Phrase' sto Phrase ?? anytime i reinitialized myt var looked like <