A25. I guess because they put today's software to run on ancient hardware. Like using cerosine for a car made in the 20's. ;-) Greetings, Nick. P.S. Many will say yes but the software is highly optimized, blah, blah, blah. That's just another way to accept how old the hardware of the 49 is. ==== Raymond Kristiansen wrote: I heard something like the TI EOS is table based, the HP CAS uses the actual algorithms. That is, the TI can do a table lookup very quickly, while the HP needs to ponder the problem first ;-) ==== Thats just another example of the advantage of the MC68000 running at 12MHz! vs the poor Saturn at 4MHz. Yes, the software of the 49 has a lot of features that the TI lacks, but... hey there's Flash ROM. The 1.5 MB of the 49 won't help to make it faster than what the processor can do. And many of you talk about the advantages of having a SysRPL compiler on calc and all that stuff, but, again, the TIBasic, even though is a really slow implementation, beats in many cases the UserRPL equivalents. And please, let's not talk about simplicity, BASIC is just that, a really easy to learn programming language. As an example( because I know you love them), I own a HP48, and a TI89 (yes I didn't upgrade to a 49) and I had an exam of Industrial Automation last week. I needed some help for Petri Nets and I wondered if there was some programs written for any of the calcs. Well Lauren Damay wrote a really nice program for Petri nets analysis for the TI89, the HP48, and the 49. So I decided to use both and see which one was better. It was almost the same for both...besides a small thing. The ti89 version was entirely written in TIBasic, while the HP48 was in SysRPL!. The Ti was still faster than the HP nowadays. So, like I always said, HP makes your life miserable, making the easy things, look harder. Someone needed to learn SysRPL (a very useless language if you take into account that besides HPcalcs there's nothing else you can program with it) just to make a program as fast as one written in TIBasic. Anyway, I still like my 48 a lot. cybernesto ==== Raymond Kristiansen wrote: They surely are not. TI89's CPU is Motorola 68000 (a 16-bit microprocessor), running at 10 MHz (or 12?) while HP49's is a Saturn (a 4-bit microprocessor) running at 4MHz. Saturn's word size is 64-bits, 68k's is 32 (or 16?). Sorry for this incomplete information, but I don't know any of these processors, this is just to tell the CPU's are not the same. Someone with more expertise on this matter could explain better the differences between both calculators. As to his second question, some argue that although the 89 has a better and faster microprocessor, its OS does not explore it well. The 49G, on the other hand, harnesses the most out of its old Saturn microprocessor. But as HP is getting out of the high-end calculator business, I don't know if these discussions have any purpose at all. G. W. Barbosa Content-Transfer-Encoding: 8bit ==== der d212mon wrote: Could you use state space instead? non-linear systesms probably don't work too well on any calculator, but I'm sure you can find ways around that. -- Aaron ==== cybernesto wrote in message news:<9u6edc$nf8$1@news.rz.uni-karlsruhe.de>... Evidently you've conveniently ignored all the advantages of a poor Saturn running at 4MHz! vs the MC68000 running at 12Mhz. (This is sarcasm. If you don't understand how important software is when considering such things, try for a moment to relate these concepts to yourself. Hint: your mind is software; your body is hardware; a great deal of the memes you have are products of man-years spent on improving the software. (A 486 with Linux runs faster and is more useful to me than a Pentium III with Windows ME.)) It will when I want to use 1.5MB =) BASIC is not simple: it is evil. The advantages of UserRPL, SysRPL, and ML are sufficient enough that I would never leave them for such a stark and weak language as TIBasic -- I've *used* it, after I'd become a programmer and knew something about languages, and I don't like it at all. I doubt your claim that it beats UserRPL in many cases. Examples? (OK, on BASIC: it's really *not* simple; just weak. It's not evil unless it's given to otherwise innoccent (virginal?) pre-programmers. Perhaps I'm biased on this. Perhaps you are, too.) Two arguments tempt me, and since I have time I'll give them both. Argument 1 (sarcastic): That's funny! What else can you program with TIBasic? Why do you like your 48 a lot when HP (I assume you mean the HP48 you have, given the rest of that sentence) makes you miserable? The SysRPL was as fast? Maybe the author learned it first, liked it better? I mostly program in SysRPL instead of any case of UserRPL, but I have Emacs. Argument 2 (philosophical): Obviously HP (see previous note) doesn't make *my* live miserable, so I'll ignore this part. It makes easy things look harder? Well, I certainly don't know where you get that. Examples? A language is useless if you can't program anything else with it? *Wow*. I'm tempted to ignore the opinion of a limited-or-non programmer, but here's my view of such things, conveniently spoken by a wise person in quotable form (translated here, because I don't remember the exact phrasing and anyway it wouldn't totally apply): The effort it takes to optimize a routine is paid once, when you write it. The effort it takes to run a routine is paid every time it runs. There, isn't that wonderful? I could use TIBasic, and be eternally frustrated by its lacks, the wall it leads me too -- or I could use the UserRPL/SysRPL/ML tree, in which everything is mine to command, I have a broad potential vocabulary, and nested levels of increasing flexibility to enter when necessary. I've never ever disliked or been disappointed by this arrangement -- not even when, years ago, UserRPL was my first language and I barely realized that what I was doing was programming. Hmm. My arguments seem to center around This doesn't bother me, and I don't understand why it bothers you. I'd like to add to this, and I suspect that problem is mainly in the user. Sorry if this post is *too* caustic -- but what you send is sometimes what you get, ne? If you think of arguments that aren't as hollow as some of mine (I don't like it. I do.), I'd be delighted to read them. Content-Transfer-Encoding: 7bit ==== cybernesto wrote: Poorly written sysRPL? I wrote PowerPlot in almost all sysRPL, with the exception of some assembler lookup tables and some of the graphics. I'm quite certain (although never raced) that it is comparable (and can certainly be made to be faster by decreasing resolution) in speed to the TI-89. Compare my graphing program to the one in the HP48's ROM. It's at least twice as fast and both were written in sysRPL. Algorithm design specific to the programming language in many cases can be a more significant factor in how fast you get your results than either processor speed or usage of a certain programming language. As an example, since sysRPL makes use of the data stack interface to programming, the proficiency of the programmer at utilizing the data stack efficiently could make a noticable difference in the speed of the program. -- Aaron Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit ==== Ok, my intention is not to discuss about the programmer. The point was, In User RPL it was way long slower than the same program on TIBasic, ok. That was it. If you want to make things fast in a TI89, you program it in C, or in 68K asm, both, industry standard languages. C doesn't need an explanation. 68Kasm, well it's not a popular processor in this days, but many Motorola microcontrollers follow its architecture. In Hp instead, you need to use the User,Sys,ML tree (nice expression Julian) which you learned ONLY to use it on your calc. USELESS anywhere else. TIBasic is inferior to RPL, yes, I've noticed that too. But for complex things, C kicks butt!. By the way, I like my 48, just because if was my first calc. I used it during my entire Engineering studies and I've got accustomed to its 'complications'. I don't regret, because in my country, HP calcs are the standard, and most people don't know how use it, besides the features a normal scientific calculator. With what I learned, I can help people get out of the mess and leave them thinking I'm a genius, because I was able to make the 'obscure and complex machine' work. Aaron Wallace schrieb in im Newsbeitrag: 3C072AFC.C70AD015@utdallas.edu... HP quite be Compare my fast and get your language. ==== cybernesto wrote in message news:<9u6edc$nf8$1@news.rz.uni-karlsruhe.de>... Yep! Hmm, I don't know if SysRPL is directly comparable to TIBasic. SysRPL is somehow between the user/high-level language, and the machine/low-level language. At what level would you put TIBasic? I would prefere User RPL on a speedy processor. No, no, no! It is a performance problem when someone doesn't use User RPL. The language itself is very flexible and capable of doing (almost) anything. But the demands of today are simply to high for the small saturn executing User RPL at 4MHz. Is there anything else that uses TIBasic? Greetings, Nick. ==== cleverjulian@hotmail.com (Julian Fondren) wrote in message news:<76ea4fd3.0111292029.20ff0e9a@posting.google.com>... Which are? It is not that simple with our brain/mind. The evolution of our hardware and software went (and still goes) hand in hand. As the software is getting more complex it really induces the further evolution of the hardware, which in turn gets more complex to be able to run the new software more efficiently. This evolution of hardware influences evolution of more complex software, and so on. It is a highly self referencing system. I don't think that the strict separation of soft- and hardware is possible for our minds. Yes, but a Pentium III with Linux would also run faster and would be more useful than a 486 with Linux. Huh?? :-/ Allow me please to add a few words about User RPL. A language specificly for formulating and solving math related problems, should reflect the complexity of the underlying method itself, the mathematical method. The rich set of commands of User RPL on the 49G is flexible/powerful enough for the formulation of such problems. It lets me choose the most appropriate way for formulation, lets me influence to a great extend the way to the solution and lets me combine the already existing commands in a huge number of ways, to create new progs that can further enrich the already existing system. The use of SysRPL makes the available power even bigger, but the question is: Did I buy the calc to solve problems, or to become a SysRPL developer? The needed effort for learning to use SysRPL is (for me) much bigger, than for User RPL, somehow replacing the goal which (again for me) should be, to be able to get answers faster than doing all by hand. This goal gets replaced by the programing itself as a goal. (Watch this group and count the math related questions against questions about menus and key assignments to understand what I mean.) Of course it is necessary to learn some programming language in order to get answers for problems, for which a programable machine doesn't provide a built-in command. But should programing itself replace the search for answers of math problems? To put it in other words, should it be expected that the user of Mathematica goes beyond the usage of the (very very) rich set of available commands, and starts decompiling and compiling the whole system again, just to find how the menu item XY, can be changed to YZ? (Bhuvanesh, your turn ;-) ) User RPL allows (theoretically) to do many many things. You can write new routines for expanding, for exchanging vars, for solving etc. There is a way for just about anything. The reason for not using it, is not that it doesn't allow to program a certain solution for a math related problem. To me the reason seems to be the weak performance of the User RPL program on the ancient hardware. For example, the provided set of User RPL commands allows to completely explode some expression in its components, check what each component is, follow rules for replacement etc. But the problem is, how long would it take for a program written with User RPL for example, to take an arbitrary expression, analyse it, and distribute * over + ? The heavy usage of SysRPL or even ML with the expectation of better performance, is not a proof for the weakness of User RPL, but a proof for the weakness of the hardware of the 49G. I think that nobody complains about the complexity of software like Mathematica. Take alone the huge number of ways to define something with Mathematica. (Bhuvanesh, your turn again ;-) ) I can of course complain about the poor performance when Mathematica runs on my old Performa 6200. But if I do complain, I don't complain about Mathematica's methods for doing things. I complain about the hardware. So the greek guy is listening with looong ears ;-) I would say, the effort it takes to concieve/optimize a routine (for formulating/solving a math related problem) is the deep thoughts, about the usage of the available and the searching for the principal existence of a such a usage. In other words, check if the available languge is built in such a way that it allows to build something new, that consists of the old. Did you at that time work for some secret agency? So secret that even you didn't know what you were doing? ;-) Hmmm, it is really no fun, programing something simple, and then waiting half an hour to see a result. I think that this can frustrate many users, who then blame all on User/Sys RPL. I mean, you must understand that many users (including me) are just bored/disappointed sitting at the desk and waiting for a simple answer that takes aeons to show. Most of the users don't really care how it is derived. But they care if it is correct and how long it took. About being caustic, well, it sometimes helps clearing the atmosphere. ;-) Greetings, Nick. ==== Aaron Wallace wrote in message news:<3C072AFC.C70AD015@utdallas.edu>... I agree with you Aaron, but you must accept that even your programs would run even faster on a, say, 10MHz Saturn processor. In addition: It is a matter of time until the dark side will also highly optimize the existing code on TI calculators. When optimization has reached such levels as on the HP49G, then the TI92 will blow up *any* HP calc it finds on its way. Greetings, Nick. Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit ==== really easy Definitely user/high-level. I was trying to point out the advantage of the hardware speed. You don't need to learn a low level language in order to get fast results. Anyway, I've read your other posts, and it seems that unlike Julian, you could understand better my post. (yes last nice for HP Yes. Me too. And the 48 look. Yes, your right. It's a performance problem. But I'm a performance lover, and if that involves learning SysRPL, I don't like it that much. As you said, I've got the machine to get results, not to make a career in calc programming. else in No, that wasn't my point either. But you must admit, that the basis of TIBasic, are the same as for VisualBasic, QBasic, etc. Programming languages easy to learn, and easy to use. Anyway that part of my message just intended to point out that the programming languages of the TI's (Basic, C, 68K asm) are something useful to learn, because now, almost everything can be programmed with such languages...BasicStamps, PIC, DSPs, Microcontrollers from Motorola, Atmel, Intel, etc. So it won't be a waste of time to learn them. You can take a lot of advantage of it. cybernesto ==== cybernesto wrote in message news:<9u7kom$qcb$1@news.rz.uni-karlsruhe.de>... Well, how easy is it to implement an algebraic manipulation with C? Or symmetry operations? Or any other symbolic manipulation? Greetings, Nick. NNTP-Posting-Host: hhi504.stud.uni-karlsruhe.de Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 ==== Ok. Last question I'll answer. Geez, how hard is to post something without getting questioned about everything!. Symbolic manipulations are really easy using TIOS entry points, and the expression stack (eStack). Ok. There you are. Thanks. Bye cybernesto References: <1006986372.30728.1@eos> <9u6edc$nf8$1@news.rz.uni-karlsruhe.de> <3C072AFC.C70AD015@utdallas.edu> <9u7kom$qcb$1@news.rz.uni-karlsruhe.de> Content-Type: text/plain; charset=us-ascii ==== cybernesto wrote: I'm still scared of the ASM syntax which is used for 68k and x86 processors. C already adds some overhead. It's sometimes even very easy to write better code than a compiler would. An example: As long as you *know* your calls do not need more than 6 RSTK levels, you can use GOSUB calls instead of a self-made return stack or RPL. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Content-Transfer-Encoding: 7bit ==== Nick Karagiaouroglou wrote: of course. I don't think anyone here would dispute that the hardware in the HP48/49 can be classified as ancient. I only wanted to illustrate that over-simplifying something can lead to false (or insufficiently qualified) conclusions. That's not to say that I disagree with what cybernesto was saying.. to be honest I really don't know if he is right or wrong. I was just pointing out that the example he used was insufficient to draw the conclusion that he had made. I'm sure it will. No one (in their right mind) would dispute that the 89/92 is technologically superior to the 48/49. -- Aaron X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 ==== So are RPL languages. I spent maybe two weeks learning SysRPL. It's not hard, and when you have to learn a specific programmong language, the syntax is all that you need to learn. I don't find it more or less difficult to learn, just different. Basic is not simple BTW. nice for HP Then that SysRPL programmer must have lacked talent bigtime. I would say anytime that UserRPL would beat TIBasic at any normal programming task. No, it's called flexibility. else SysRPL is no more useless than TIBasic. A programming language that I have used for 10 years doesn't seem useless, does it? Anyhow - a programming language is not hard to learn, it's just syntax. Regards Steen Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== Hmm, you could buy a 49G with double speed module. :) Seriously, I guess we are all spoilt by Moore's law. What do you expect? It's a $200 calculator, not a $800 palmtop or a $10000 workstation. Granted, the TI 89 is about the same prize, and it's faster. But it won't solve any large-scale problems either, and HP calcs have other advantages. Slow is beautiful. Yes I know my poor 49G is working hard when the hour glass shows up. ;-) Cheers, Albert -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit ==== Julian Fondren wrote: Are you running mainly operating system on your computer ? Jack Message-ID: References: <76ea4fd3.0111121622.23a3fddf@posting.google.com> ==== JF> OK. I do everything on-calc, except occasionally print out or read on JF> a computer some documentation. I do this for comfort and economy, since JF> it is, for me, both difficult and time-taking to access a computer that JF> is both connected to the internet and available for me to connect to my JF> calculator. I agree 100%. What I don't have on the calc, I don't have with me. It is no fun to carry lots of printed documents. JF> This is easy. First you set flag -91, then you use the matrix editor JF> to quickly build a list of topic/command/whatever->string lists, give JF> it a title string, and then make a CHOOSE out of it. Pass the result JF> to SCROLL. Isn't that nice? It's a little tedious for libraries like JF> Mawk, but most should be no trouble. For libraries which document their commands internally using the message handler (currently AFAIK LTools, Emacs, ConstTools, UTool, OT49), the CdBox command in LIBman is a very similar and general way to explore the commands of a library and to look at the builtin help. We just have to persuade more authors to put help into their libraries. - Carsten ==== Carsten Dominik wrote: Carsten, isn't it a better idea to reply in this NG? There may be more people interested in undocumented pointers. Actually, the emptiest barrels in this NG are the ones who waste bandwidth all the time by not using the mouse to cut superflue text as has been done in this message :-) Here an important undocumented command from your list: CK0ATTNABORT This is just the program executed by the two UsrRPL program delimiters which actually denote programs. The command does essentially ?ATTNQUIT, i.e., nothing if CANCEL has not been pressed. Otherwise clears the ATTN buffer and the key buffer and then ABORTs. However, the UsrRPL program delimiters for local environements do completely different things although denoted by <<, >> as well. The normal program delimiters can mostly be stripped, (e.g., by ->XU from OT49), the latter not! Also the UsrRPL-word UNTIL denotes just CKOATTNABORT. Hence, the UsrRPL-program << DO blabla UNTIL foo END >> may in general and without harm be shortened to xDO blabla foo xENDDO This is a slightly more cautious implementation of :: BEGIN blabla foo COERCEFLAG UNTIL ; since both, xDO and xENDDO, first run CKOATTNABORT which sometimes is a useful interrupter in SysRPL programs :-) - Wolfgang Message-ID: References: <3C064C18.BCE983FD@math.fu-berlin.de> ==== WR> CK0ATTNABORT Added, thanks. - Carsten ==== Wolfgang Rautenberg wrote in message news:<3C064C18.BCE983FD@math.fu-berlin.de>... Particularly as the entry database does not have any kind of 'newly added' listings -- it's a big document. I handle this with the unix 'diff' on the HTML, but non-unixy people are probably going to have trouble. X-Proxy-Client: cung@uiuc.edu from lincoln-pl-5.flexabit.net ==== hi, is there an autozoom that will best fit the graph? what i mean by this is like on the TI85, there is a zoomfit. is there something like this on the hp48g+? thanks NNTP-Posting-Host: aputeaux-104-1-2-63.abo.wanadoo.fr ==== Yeah sure! once your graph is done, and you re under PICTURE mode (you can move the cross hair) there s a zoom box on lower left corner..prees it (press A) then press next (under F) and then press F (or is it A?) Should be ZAUTO like Zoom Auto... You re done ! Pied gc a 216crit dans le message de news: iNSN7.1503$Nz1.26075@vixen.cso.uiuc.edu... is like hp48g+? References: <9u3iak$flq$1@wanadoo.fr> <9u3m08$vj5$06$1@news.t-online.com> <3C055CB6.FC9B51B8@iname.com> <9u3t0l$hud$04$1@news.t-online.com> Content-Type: text/plain; charset=us-ascii ==== Raymond Hellstern wrote: [...] That may be one of the reasons. I only program on the calc (except some hacking on Jazz, which is SASM syntax). But there I prefer to write one idea on one line. E.g. LC(5) NUM ?A=C.A RTY seems much more readable than LC(5) NUM ?A=C A RTNYES And the most omitted. I'm not an exception here either. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== Hi, Maybe, but the MASD sources I saw so far seem to be nearly all made on the 49 (or alike) itself, and so they look somewhat squished. It's not that different to using macros for SASM. I did 'cross development' in 1999 and 2000. For this I used source files made with my favourite editor, then the 49SDK for compiling and testing. The SDK's built-in editor doesn't support tabs, and screwed up my source file formats, so I didn't use it for editing;-) But maybe we mean something different when saying 'cross development'. I did it on the PC, so I didn't have to worry about the keyboard of the 49;-) And there was no need for me to use MASD. I used exactly the same source files, instead of re-entering the code on both platforms. Of course the above makes no sense. You could take a look at the RPL48 sources, or some other sources by Detlef or me. Then you know what I mean. Sometimes it's practical to have kinda data flow matrix, and a header which describes the IN and OUT conditions. It simply saves time;-) The coding and commenting suggestions in SASM.DOC are a good starting point (Standard headers, etc.)... Of course you can, but who does ? ;-) The MASD sources I saw so far were more or less poorly commented. Maybe becuse they were made on the calc itself, and there every key stroke counts;-) Regards, Raymond Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== Hi Thomas, Thomas Rast schrieb im Newsbeitrag news:3C06242C.6F7B3BD9@iname.com... Up to 1990, when I didn't have a PC, I used to program on the 48 itself very intensively. (Before 1990 it was the 41 and the 71B, of course) My first disassembler was written on the 48 completely, equipped with only a 32K RAM card and a serial printer. When joining W&W, I had the chance to use a HP-UX workstation and a PC for development. I don't agree here. For me the second version is much more readable. I didn't document all my code as I wished I had done. But it's getting better. At least the more complex code slices have suitable headers and inline docs, because I don't want to waste too much time to re-document the code every time I look at it;-) Regards, Raymond <9u3m08$vj5$06$1@news.t-online.com> <3C055CB6.FC9B51B8@iname.com> <9u3t0l$hud$04$1@news.t-online.com> <3C06242C.6F7B3BD9@iname.com> NNTP-Posting-Host: 217.11.167.223 ==== Le Thu, 29 Nov 2001 13:03:56 +0100, Thomas Rast a 216crit : I think that the more radable on a computer screen is : % Exit case : blabla LC(5) NUM ?A=C A RTNYES Just add a comment and it becomes more readable than the first version with all on a unique line (but this version is more readable on the calc ! That's why we need intelligent transfer tools, able to reformat an asm file...) Bye. References: <9u3iak$flq$1@wanadoo.fr> <9u3m08$vj5$06$1@news.t-online.com> <3C055CB6.FC9B51B8@iname.com> <9u3t0l$hud$04$1@news.t-online.com> <3C06242C.6F7B3BD9@iname.com> <9u5so2$gqu$02$2@news.t-online.com> Content-Type: text/plain; charset=us-ascii ==== Raymond Hellstern wrote: Well it seems to be a matter of taste then... lets not get into religious questions ;-) The problem I have with most of my code is that it constantly changes its internal interfaces throughout the development. And when it's all done and working, I'm too lazy =) Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== Hi Thomas, That's right. I kinda 'grew up' with the HP/SASM syntax, and this may be a reason why I can better read/write this syntax. I also didn't want to say 'use this' or 'use that', I simply wanted to tell what's my favourite choice. of course not;-) After all, it's just a hobby. I know that from own experience, too;-) That's why I try to put in some headers before actually starting to write code. Sometimes when I have an idea for a program, I start one evening, often don't have time to go on the whole next week, then restart maybe two weeks later. So I usually make stack diagrams and in/out descriptions before suspending the development. Regards, Raymond Content-Transfer-Encoding: 7bit ==== On Thu, 29 Nov 2001 17:50:47 +0100, Raymond Hellstern wrote: I think he's talking about MASD's high level control structure capabilities. Still, there *should* be no trouble implementing similar capabilities using conditional compilation and macros right? Well, unfortunately until the HP-Tools stop choking on such simple things as push MACRO sp, value sp = sp + 1 stack$(1) = $2 ENDM which it should be able to process ( and I can't find anything that implies otherwise in SASM.DOC ) there will be no such facility without performing great contortions. Of course, if HP didn't have a ridiculously restrictive license on the tool source code then we could have corrected this feature long ago. ---------------------------------------------------------------------------- --- Jonathan Busby - Remove the random permutation of NOSPAM from my e-mail address before replying. ==== Hello Jonathan Busby wrote in message news:bc8e0ucsivghuv8f8qm78eqm6rpfp51vcc@4ax.com... Which ridiculous restrictive license ? If you want to modify the source code to add new features, nothing prevent you to do so. Then you can just send me your modifications and I will distribute these changes. You have full access to the source code Jean-Yves ==== Carlos Cue. wrote in message news:<9u4ude$eeu$1@diana.bcn.ttd.net>... Samson Cables appears to be able to ship to Spain. It sells HP49G calculators for 155.95 USD (plus $20 for the package that comes with an interface kit; I'd suggest you just buy a cable too from them and save $5). ==== ugh, excuse me. see www.samsoncables.com ==== Julian Fondren escribi227 en el mensaje news:76ea4fd3.0111291254.acdbfd6@posting.google.com... news:<9u4ude$eeu$1@diana.bcn.ttd.net>... HP49G Thank you very much JULIAN, I try to purchase there.. ==== Carlos Cue. wrote: Hola, Si eres de Barcelona puedes mirar en ICT, calle Entenza 224 (creo). Siempre t qda el socorrido Corte Ingl216s... Saludos Ricardo -- Linux Registered User: 202 170 Kernel 2.2.19 http://193.145.89.239/~rblasco/ Un222os, hermanos linuxeros 640 Kb de memoria son m207s que suficientes (IBM AT Designers - 1982) Bueno, espero que esto no aparezca en la versi227n final (Bill Gates en la presentaci227n de Windows Crash Debug '98) ==== In article <20011125141058.15535.00001715@mb-mn.aol.com>, PawlRevere2 wrote: Organization: http://groups.google.com/ Lines: 6 ==== On the HP49 it is possible to expand for example (n+1)! to n!*(n+1) . But is there any built-in command that does the opposite? Greetings, Nick. ==== [beep] [beep] if all fails, use the built-in 'eyeball' (yours comes with cutting laser too) [beep] [beep] In article , Nick Karagiaouroglou says... ==== [long sad beeeeeeep] It did failed all. Starting laser activating sequence... (Fire and smoke... ;-) Eyeball ready. COLLECT. I said COLLECT!!! Do it! Would you %&$'*! COLLECT them ?!!!!!! ----Nop, nothing------- Deactivating laser. (Save batteries ;-) ) [beep??? :-/] Greetings, Nick. rcobo@eng.morgan.edu wrote in message news:... NNTP-Posting-Host: dutw623.wbmt.tudelft.nl Mime-Version: 1.0 ==== And I didn't even count that one. (in light of VPN's autoreply mode :-) -- This message was written with 100% recycled electrons Pivo ==== Peter Geelhoed wrote in message news:<3C0608CA.78B9FA79@student.tnw.tudelft.nl>... Well, I didn't reply because VPN had said everything for me, and because I don't like such repetitions. I'll be sure to post something next time. That's too bad. He was probably very sincere. ==== Thomas Rast wrote in message news:<3C04F4DD.4784419C@iname.com>... I certainly wouldn't want to implement Perl's syntax =) Or awk's, for that matter... what important functionality of it requires a full interpreter? Thank you, Thomas ==== hello, how can i open the EQW in a program so that it automatically comes up with a custom menu (instead of the default menu)? thanks ==== rcobo@eng.morgan.edu wrote in message news:... [Beep!! :-)] 43.1 KEYEVAL @ for activation of EQW About custom menu in EQW: (VPN's idea): Create program << @Title of choose box, none if empty { DISTRIB EXP2POW ...} @list with your commands, progs etc 1 CHOOSE IF THEN EVAL END STOre this in STARTEQW in HOME directory. Now whenever you are in EQW and you select some expression/subexpression, you can press [blue-shift] then [MODE] for getting a choose box with the commands in the commands/progs of the program STARTEQW. They act upon the selected subexpression only. Greeting beeeeps!! Nick. X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 ==== Try this: 1.) 6x^(2a+1)-4x^(a+1)y^3-3xy^6 2.) (x^2+2)^(2/3)*(x^2-3)^(-1/3)-(x^2+2)^(-1/3)*(x^2-3)^(2/3) I didn't give a real example because I just thought there was maybe a bug or something. A friend of mine tried this on his TI89 and he says it worked, but for me it gives the same thing back. Steen Schmidt wrote in message news:pFgN7.1565$K4.275970@news010.worldonline.dk... the for the NNTP-Posting-Host: hhi504.stud.uni-karlsruhe.de Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 ==== If you want to horrify yourself about the factor capabilities of the 49 please try this (1 - COS^2(x) ) ^4·(1 - SIN^2(x))^3 ·(SIN(x)^2 + COS(x)^2 )^5 A really stupid expression if you look at it carefully, and easy to simplify mentally. The TI89 returns the right answer in less than a second (which as you may guess is sin^8(X)cos^6(x). The 49 with all the flags in default state, returns an expression that could drive a hurried student in an test crazy. Ok, Steen was able to simplify this expression through some manipulations, but, he knew the answer before, and again, HP is making your life more miserable by making easy things complex. cybernesto NNTP-Posting-Host: mix-grenoble-110-4-76.abo.wanadoo.fr Mime-Version: 1.0 ==== cybernesto a 216crit : HP (more precisely I) decided to favor people who have a brain. The operation the TI did is *not* factorization, it is trigonometric simplifications. Factorization is an operation over polynomial. Since this expression is not a polynomial with respect to x, the first task is to make a change of variable so that it becomes a polynomial, here it is a polynomial with respect to s and c where s=SIN(X) and c=COS(X). If you call FACTOR you will get the factorization of the polynomial with respect to s as main variable (exactly as if you had done the change of variables yourself). It happens that due to trigonometric identities you can rewrite this expression as a polynomial with respect to s only or to c only something you can ask the 49 to do with TRIGSIN or TRIGCOS. Then factoring will return s^8*(1-s)^6*(1+s)^6 (FACTOR after TRIGSIN) which is the correct factored expression. Next time, please learn a little bit maths before stating stupid assertions. ==== many people are missing the main point. for me a calculator is not a performance machine that spits out rigth answers fast, but a learning tool. i consider the HP49G a better tool for learning (math and programming) than the TI89. the HP49G allows you to be part of the problem and to do that, you need to know how things work. strengh by itself is nothing. meaning is everything. the potential for learning math and programing should not be compared with speed and graphics. that is what many peole fail to see. the HP49G is a teaser, and teasing is the best way to learn :) the animal cannot only talk, but now it can also program. In article <3C0728F1.59115211@fourier.ujf-grenoble.fr>, Parisse Bernard says... NNTP-Posting-Host: hhi504.stud.uni-karlsruhe.de Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 ==== Hey, give me a break you genius. I'm aware of who you are, and what you did, and before reading your post, I had respect for you. But you are noone to tell me Next time, please learn a little bit maths before stating stupid assertions., If I posted this example was just to show how the automatic operations of the freaking calc whose CAS you wrote can make a stupid expression a complete mess. Call it factor, call it simplification, I don't care, english is not my first language, and is your 'factor' algorithm the one who makes the calc go crazy. test Yes, if you didn't check the answer that your CAS returns, Newton, I'll copy just a little bit of it -((COS(x)^8 - 4·COS(x)^6 + 6·COS(x)^4 - 4·COS(x)^2 + 1)·SIN(x)^16 + (5·COS(x)^10 - 23·COS(x)^8 + 42·COS(x)^6 - 38·COS(x)^4 + 17·COS(x)^2 - 3)·SIN(x)^14+(10·COS(x)^12 - 55·COS(x)^10 + 123·COS(x)^8 - 142·COS(x)^6 + 88·COS(x)^4 - 27·COS(x)^2 + 3)·SIN^12 + (10·COS(x)^14 - 70·COS(x)^12 + 195·COS(x)^10 - 281·COS(x)^8 + 224·COS(x)^6 - 96·COS(x)^4 + 19·COS(x)^2 - 1)·SIN(x)^10 + (5·COS(x)^16 - 50·COS(x)^14 + 180·COS(x)^12 - 325·COS(x)^10 + 325·COS(x)^8 - 180·COS(x)^6 + 50·COS(x)^4 - 5·COS(x)^2)·SIN(X)^8 + .... And now I want you to tell me if THAT is an elegant result for such a stupid expression. Yes of course, through some manipulations you can get the right answers, but hey, Mathematica, or Maple, don't give such results, do they Euler? Yes, factoring is about polynoms, so how about trying this poly Fourier, 2·(x^2 - y^2 )^6 - (x^2 - y^2 )^5 ·(2·x^2 - 3) Why the hell your super CAS returns this automatically? -((2·y^2 - 3)·x^10-(10·y^4 - 15·y^2)·x^8 + (20·y^6 - 30·y^4)x^6 - (20·y^8 - 30·y^6)·x^4 + (10·y^10 - 15·y^8)·x^2 - (2·y^12 - 3·y^10)) I mean, if he is unable to simplify it without your incredible FACTOR command, why the doesn't leave it as it was. What's the purpose of messing the entire thing? uh Leibniz? SO IF THIS WAS STUPID ASSERTION; EXCUSE ME BUT AT LEAST I TRY TO PROGRAM THINGS; THAT DO WHAT USERS WANT; IN THE WAY THEY WANT; NOT THE WAY I WANT IT. Thanks, a lot, and please forgiveme if it sounded like trolling, is, just that I didn't offend anyone in my previous post, to receive such a STUPID classification. cybernesto ==== Bernard Parisse wrote: <> Correct, it is *trigonometric* factorization (TrigFactor[ ] in Mathematica). rcobo@eng.morgan.edu wrote: A matter of opinion, as always :) I have found the TI-89 (actually the TI-92+) to be a great learning tool. That is where I first developed my love of computer algebra. So you also are saying that the HP49G CAS is slow? (It's unexpected, that's all.) Bhuvanesh. ==== Parisse Bernard wrote in message news:<3C0728F1.59115211@fourier.ujf-grenoble.fr>... Mr. Parisse! When I use TRIGSIN and then FACTOR on the expression (1 - COS^2(x) ) ^4·(1 - SIN^2(x))^3 ·(SIN(x)^2 + COS(x)^2 )^5 I get -s^8*(s-1)^3*(s+1)^3 (I think you misspelled ^3) But could you tell us which command or command/sequence can be used on the 49G to tranform (1 - COS^2(x) ) ^4·(1 - SIN^2(x))^3 ·(SIN(x)^2 + COS(x)^2 )^5 to sin^8(X)cos^6(x) without taking it in the EQW and applying separately the commands TRIGSIN, TRIGCOS and TRIG? I also accept that simplification depends on how one wants it to be, but I can hardly say that the 49G simplifies the above expression. Try SIMPLIFY on (1 - COS^2(x) ) ^4·(1 - SIN^2(x))^3 ·(SIN(x)^2 + COS(x)^2 )^5 and if you think that the result COS(X)^14-4*COS(X)^12+....COS(X)^6 is more simplified than sin^8(X)cos^6(x) then OK, you win. (As always) Also try TSIMP in real mode on the above expression and see if the result can be considered more simplified, than the input itself. Sometimes the other side can also be right, expecting not the commands there to be the same like commands here, but expecting some easy things to remain easy. In this case the TI89 does simplify while the HP49 doesn't. Greetings, Nick. ==== Hi Rcobo! Despite the fact of being a part of the problem and knowing about trigonometric relations I just can't rewrite (1 - COS^2(x) ) ^4·(1 - SIN^2(x))^3 ·(SIN(x)^2 + COS(x)^2 )^5 to SIN(X)^8*COS(x)^6 with the built-in commands of the HP49G. How would you do that? Write a new program for this? Buy a TI? Assimilation and robotification of my brain? Greetings, Nick. rcobo@eng.morgan.edu wrote in message news:... NNTP-Posting-Host: hhi504.stud.uni-karlsruhe.de Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 ==== Even if you called us the dark side your statement is a compliment. Mr Parisse sent me to learn math, because he says I'm stupid. C'mon. If he didn't like my post, he just had to say it. Not to offend me. And if my message let you find such a flaw, then it wasn't so stupid. Thanks. Now I feel better. cybernesto ==== hi chief, comparison: my maple (with a power wire hooked up to the wall and everything) is not able to manipulate expressions as my willy-nilly desire either. sometimes it does. one day, i'll worked to feed the computer (as oposed to the computer working to feed me). that day, i'll apply the nowledge gained through my HP49G so that i will claim my rigthfull freedom over the machine. In article , Nick Karagiaouroglou says... ==== hi chief, comparison: my maple (with a power wire hooked up to the wall and everything) is not able to manipulate expressions as my willy-nilly desire either. sometimes it does. one day, i'll worked to feed the computer (as oposed to the computer working to feed me). that day, i'll apply the nowledge gained through my HP49G so that i will claim my rigthfull freedom over the machine. In article , Nick Karagiaouroglou says... ==== In article <662e00ed.0111300553.101b91a0@posting.google.com>, Bhuvanesh says... yes, the final answer to everything is based on opinions (isn't that a pseudo relativity theory?). i am glad you had a good time. right now i am were you used to be (learning). well, the CAS in itself is a work of art (I compare it to dancing flamenco on top of a telephone pole). speed is not important to me. neither i have unrealistic expectations from the calc. I am sure the TI-92 is also a great calculator :) cheers NNTP-Posting-Host: mix-grenoble-110-4-239.abo.wanadoo.fr Mime-Version: 1.0 ==== cybernesto a 216crit : I never said you are stupid. I said you made a stupid post, that's completely different. And the main reason was that you were speaking of factorization instead of trigonometric simplifications to support your claim that the HP49 answer was not what one should expect. As a reminder, when you press enter in algebraic mode or EVAL in RPN mode, the expression will be fully expanded with respect to all rational variables (and written in recursive ascending or descending degrees depending on the corresponding flag), hence the HP49 answer is exactly what one should expect. NNTP-Posting-Host: mix-grenoble-110-4-239.abo.wanadoo.fr Mime-Version: 1.0 ==== You can not easily, it would require OBJ-> But the question should be: why would you consider s^8*c^6 a better form than s^8*(s+1)^3*(s-1)^3? s^8*c^6 looks like a 2 variables polynomial which a polynomial system will consider as more complicated as a 1-var polynomial. It's just that you get a 1 var polynomial (expanded). You can of course FACTOR it. SIMPLIFY returns expanded expression. TSIMP is not well adapted because introducing complex numbers here is not appropriate. Yes the TI is more adapted to these texts of exercices or exams because they were thought to be solved by hand without calculation (and frankly if I had to work with the above expression, I would simplify it myself to sin(x)^8*cos(x)^6 because it's much faster to enter it that way on the calc. The same is true at a lower level: everyone would enter 10^6 instead of 10*10*10*10*10*10...) But if you try expressions that have not this reasons to simplify you would find the HP more adapted. For example, if someone has a TI at hand (I don't have one at hand), I would be interested if he could give us the result of an experiment: take the original expression but replace sin(x) and cos(x) by s and c, expand and factor the result, does it work? X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 ==== (20·y^8 - My '49 returns '-((2*y^2-3)*(x-y)^5*(x+y)^5)', so where's your error? The factor command shouldn't decide if something is simpler than anyhting else, it should just *factor*. What about this: 68497228800 FACTOR -> '2^10*3^5*5^2*7*11^2*13' Should FACTOR just return 68497228800, since it's much simpler than the output? I think not. FACTOR does what it's supposed to, and so does EVAL, SIMPLIFY, LIN, COLLECT and all the other simplification commands. When you know how to manipulate these things, hence know the math behind it, you can use the EQW and get the exact result you want. The TIs are for schools, where you can only expect one out of 5 different problems in each category. The lacking flexibility is not overcome by speed in a real world situation. Regards Steen Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== Hi, I think the main problem here is a misunderstanding of what these rewriting operations actually do. When you say simplify you expect the calc to do some magic and return some kind of shortest equivalent expression. This is not the way the HP49 CAS works. Instead, you have a set of well-defined rules which perform certain rewriting steps, and you must decide for yourself how and where to apply them, and in which order. This is the way the HP49 CAS has been designed, and reportedly this approach offers much more flexibility and works better on complicated (non-schoolbook) problems. (BTW, I think that there is also a theoretical reason for preferring the manual approach, namely the undecidability of the word problem in equational theories, which makes it hard to come up with an algorithm for always doing the right simplifications in non-trivial theories.) But I also see a problem here, namely the docs. :) While many rewriting rules are described in the AUG very explicitly, I cannot find more than a sketchy description of SIMPLIFY/EXPAND. (In fact, I couldn't find an entry for SIMPLIFY at all.) I'd guess that SIMPLIFY and EXPAND are the same, and perform a combination of factoring+canceling out common factors in num/denom of a rational expression, then expand using distributivity? Could someone who knows more please explain? Cheers, Albert -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag ==== rcobo@eng.morgan.edu wrote: I'm still learning too. Now I have the added advantage of being able to ask my colleagues about various algorithms and such :) Hehe :) Well, it depends on what you need from it. If you wanted to solve a 5000*5000 linear system of equations, you naturally wouldn't want to wait too long for the answer :) Happy learning, Bhuvanesh. Mime-Version: 1.0 ==== Hi! I hope that somebody can help me. I have installed HP Comm38 1.0r4 and PC Comm Kit V3.0r4 on my notebook computer. My computer is LG-IBM Thinkpad X20 and its OS is windows Korean Version. I have no problem to transfer and get files between 38G and the computer. With 39G, I can transfer a file from 39G to the computer but when I tried to transfer the file back to 39G, I encounter file name error. With 48GX and 49G, also I can transfer files from the calculators to the computer. But I can not do it from the computer to the calculators. The same file name error message. When a file was transferred to the computer, the icon was changed to an unidentified application file. In 39G case, when a file was transferred to the computer, the icon was the HP calculator type. I suspect that this might be some problem caused by the multibyte Korean character representation in Korean Window version. Is there anyone who encountered this problem and know how to resolve it? Desperately seeking for help! TS PS: I am sorry to you who are confused by my poor English. ==== Ricardo Muchas Gracias por ayudarme q sin duda lo has hecho mucho!!POr cierto ahora va la pregunta del mill227n, donde puedo comprar la 49g por ese precio,porque llevo un tpo buscando precios y todos disparataos(al menos 40000),as222 que si eres tan amable de indicarme donde puedo pillarla a ese precio!! Muchas gracias de nuevo... ==== Carlos Cue. writes: Hola Carlos, En Barna, prueba con el CPDA, la tienda dentro de Industriales, en la Diagonal. No creo que encuentres precios m207s baratos. Pero puedes probar tambi216n en el barrio de la Barceloneta. Y, como dijo Ricardo, no dejes de pasar por el ICT, y pregunta por la funda y el CD que sol222an dar... Luego te quedan el FNAC y el Corte Ingl216s. El primero suele ser m207s barato. Pero lo bueno del segundo es que si encuentras un defecto (cosa MUY frecuente en las HPs de hoy en dia) pues te la cambian sin hacerte ni una preguntita. Bueno, a m222 me la hicieron al devolver una 19BII, pero no me apeteci227 responder (forma educada de decir que no me sali227 del forro de los co*ones :-) Bueno, si encuentras algo mejor, h207zmelo saber, ¿vale? Hace ya alg234n tiempo que no voy de pesca. Por cierto, por qu216 no hablamos en Catal207n? Ya puestos... Apa, que et vagi b216! END OF LOCAL SERVICE. Some words in English now: Carlos asked for a cheap place to buy an HP49G in Barcelona. I said that he should wait some time, until it is discontinued, and becomes a piece for collectors. Since the incredibly low value this machine will have among collectors and true specialists on calculating machines, he will be able to find a '49G very very cheap. Almost free, as free as this translation has been. :-) Regards, Bye. Jordi Hidalgo HPCC member #1046 johil@tv3mail.com ==== En la Universidad Polit216cnica de Valencia, hay un par de particulares que venden calculadoras HP. Tienen los mejores precios con mucha diferencia sobre tiendas (papeler222as, FNAC, Corte Ingl216s). Seguro que en Barcelona pasar207 lo mismo. Yo de ti preguntar222a a alguien que est216 estudiando alguna ingenier222a en Barcelona. Mithrandir. ==== Hi Folks, So I have an HP48SX that I bought a long, long time ago, brand new! It does pretty much everything I need it to, but it would be really nice to add some ram to it. The HP 128k ram cards are big coin, when you can find them. They typically are as much, or more expensive than, a brand new 48 G+/GX! I found some other third-party cards at http://www.uuhome.de/oklotz/index_e.html and they seem much more affordable. My question is, has anyone bought from this vendor and, if so, has the product performed as it should, ie exactly like HP's RAM cards? I'd be interested in hearing from owners of, in particular, SXs, but I'd be happy to hear from anyone else, too. The price for these cards certainly seems right!! I've got ROM-E, I think, for what it matters... Thanks in advance!! -- ____________________________________________________________________________ ___ Christian B. K. Schroeder, MSc (Physics) PhD Candidate, Dynamic MRI Microscopy University of Guelph Department of Physics Guelph ON N1G 2W1 CANADA Phone: (519)824-4120 ext. 8542 Fax: (519)836-9967 ==== [edited for clarity] Scott Klettke wrote in message news:CnfN7.19$tI2.1155@client... It might; I started on the 1.16 and can't talk about 1.10. You should upgrade anyhow. I don't observe anything that really fits a 'fade out'. If you had been more precise in your report I might be able to help you more -- but then, my failure to notice what to you is greatly obvious might itself be evidence for damage or bug in your 49G: Upgrade to 1.19-6. ==== mithrandir@terra.com (Mithrandir) wrote in message news:<92e02401.0111281138.7949638@posting.google.com>... Is it in your PATH ? Do you have the tables (library 993) installed and jazz all in port0? jazz should give you more than just invalid token as an error anyway. *which* is the invalid token? As for sources, drag'n'drop the text/file. If you lean ED, you will write a lot faster on your real calculator than on any other editor. Emacs(hp) comes close from what i imagine. ==== First of all, thanks to all of you for your help and interest. Second, I've found the problem with Jazz. I had written {ZERO} instead of { ZERO }. There was no problem at all with the transfer to the calculator. I have now compiled the program and it works ok. As for HPTools, I'm still not sure of what was happening, but for what I've been reading, the problem must be that SASM hasn't the correct location of the opc file. For the moment, I'll stick to Jazz, and I may try HPTools when I finish this program. One last thing, most of you are referring to HP49, in case you're interested, I'm using HP48. One question, if I want the program to run on HP49, I'll just have to recompile the program using 49's entries? Oh, and one more thing, when I program in User I do it with the calculator, but for System I use the computer because of two reasons: code is larger and it is a pain in the ass to write that much with the HP keyboard, and I'm not yet willing to buy a memory card to install Jazz, UFL and so on. Thanks again, I'll post the program when it's completed. Mithrandir. Content-Transfer-Encoding: 7bit ==== Mithrandir wrote: You'd get used to the keyboard. I can program faster on the '48 keyboard than on my qwerty using emu48. The best reason(s) to program in sys on Emu48 are: (1) if you mess up, you can easily recover from TTRM, (2) much faster to use Jazz on emu48 to debug / step through code. I typically write code on the calc itself, then debug on the computer. -- Aaron ==== Hello Mithrandir wrote in message news:92e02401.0111281138.7949638@posting.google.com... First of all, I suggest that you upgrade to a newer version of the HPTools v3, they don't require a SASM.OPC file anymore. If you download my package, you will find some examples that you can compile directly quite easily. Then you have some good tutorials available by Eduardo Kalinowski All you need can be found at www.hpcalc.org Regards Jean-Yves ==== mithrandir@terra.com (Mithrandir) wrote in message news:<92e02401.0111291502.5d9cbcb0@posting.google.com>... (I haven't looked much into this, so be warned; the following is at least based on what I've seen of documentation and what I intuit:) If you have the source, much of it should recompile fine for the HP49. Some entries I know on the 48 do not exist on the 49 -- see the listings on hpcalc.org for information about these. Oh, I thought you were using the HP49. I don't actually keep source code for my SysRPL programs; I just decompile them. As for large code... well, I'm primarily a Forth programmer. Thanks for explaining your position =) ==== As a beginner on the HP 48GX, I am trying to find out how do I install the famous BZ and TED programs on my HP 48GX, or where can I find detailed instructions about how to install ( I`m aware of the HP calc.org), and how to use it ? Greatfull for help Jan ==== You can download BZ 1.2 from http://www.hpcalc.org/hp48/utils/compress/ and also download BZM 1.2: it will help you for using the compressor. Really it´s very comfortable... Raul downhole@online.no (jan) wrote in message news:<6d23acde.0111300559.7ee74929@posting.google.com>... Xref: newsmst01.news.prodigy.com comp.sys.hp48:144180 Hello all, Surfing in the net, I re-discovered a MathML website. It has some nice articles about using xml parsers for editing, showing,.. math expresions.. You should take a look at: http://www.mathmlconference.org One I have found interesnting about hadwriting interface for math: http://www.mathmlconference.org/2000/Talks/suzuki/ Have a nice day, J.Manrique CdU de la ETSIG ==== maple7 already comes with mathML :) i use maple7 quite often, but i haven't used mathML yet maybe when it becomes more popular www.mathsoft.com In article <2AN3GqfI5UMu@opalo.etsiig.uniovi.es>, Lopez de la Fuente, Jose Manri says... ==== There's a MathML export program for the 68k calculators at my site: http://triton.towson.edu/users/bbhatt1/ti/ I'm very interested in MathML. Being based on XML is particularly useful. You can search for subexpressions etc., and it is an unambiguous syntax. I use techexplorer (the free edition) for displaying MathML. MSIE will probably support the MathML DTD pretty soon, and there's a MathML project for Mozilla too. Bhuvanesh. zz921174@opalo.etsiig.uniovi.es (Lopez de la Fuente, Jose Manri) wrote in message news:<2AN3GqfI5UMu@opalo.etsiig.uniovi.es>... ==== rcobo@eng.morgan.edu wrote: Mathematica does too, of course :) Bhuvanesh. Content-Transfer-Encoding: 7bit ==== My brother is a Civil Design Draftsman /Surveyor for a small company in Tennessee and when I told him that there was some speculation that HP was going to stop production of the HP48GX, he did some research. He and his boss contacted several companies that sell or manufacture surveying equipment( Including RAM cards etc for the GX) and was told that HP had told these companies that the GX was not going out of production. Whether this is true or not we do not know,but apparently HP has told these companies that the GX will still be around. Take this information with a grain of salt. Just another bit of info to add to the mix. Harold A. Climer Dept. of Physics,Geology and Astronomy U. Tennessee at Chattanooga X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 ==== well, if they are going to still manufacture them, then why are they pulling them from OfficeDepot?? talk about major sales lost via that franchise. Harold A. Climer wrote in message news:9g8f0ugh0hbagsm9at8san85a7872ebuvc@4ax.com... ==== Harold A. Climer wrote in message news:<9g8f0ugh0hbagsm9at8san85a7872ebuvc@4ax.com>... The GX will always be around. If you have troubles getting one, try ordering one from greece... here these things are not used, but stores have them... i haven't seen the 49g here yet, i can safely bet that they don't know what is going on in the calc world as a whole. George P. Tsiros Physics dpt. University of Patras, Greece www.upatras.gr Struggling to get a degree. ==== I've made a DoInputForm. It's fields only accept reals. Initial values are 0 for all of them, but the user can erase that value leaving the field empty (this will return bint MINUSONE, which is not a real). Shouldn't it give an Invalid Object Type error when the user tries to do this? This is important, because I was planning on making no argument checking because DoInputForm should be doing it, but I now see that I may have bints instead of real numbers. What do you think about this? Will I have to check for reals two times then? If so, ther might be no point in having DoInputForm only accepting reals. Mithrandir. ==== mithrandir@terra.com (Mithrandir) wrote in message news:<92e02401.0111300740.5a7658d8@posting.google.com>... On all field of input forms, the user can either enter a value, in this case the system will only accepts inputs of the given type OR leave it blank, in this case the system returns MINUSONE to indicate that. I'd say it's not checking for two argument types, it is checking for valid inputs... By the way, it is possible to force a user to enter something in the field always by using message handlers. -- You want it in one line? Does it have to fit in 80 columns? :-) -- Larry Wall in <7349@jpl-devvax.JPL.NASA.GOV> Eduardo M Kalinowski ekalin@bol.com.br -- **Novo e-mail** **New e-mail** http://move.to/hpkb ==== Hi! Could anyone tell me which are the steps to improve the language programming of 49? And relatives url where I could download all the material to study the language? Thank to everybody has good advices! Bye ==== [Let me remind you all that there's no need to answer all questions: the winner will be the first who answers the MOST number of questions correctly] Would you like to know more about the HP-12C? Things such as programming scientific functions, reviewing emulators, watching its new Agilent chip, learning all about codenames and varieties, etc. Or maybe you would enjoy reading a thorough article on Compaq's merger, or what happened at the HHC-2001? Well, all you've got to do is get the September/October issue of Datafile. And guess who got a spare copy? I just want to make sure that it will go to the right home, hence this: ***CHALLENGE*** Five questions. The prize will go to the first person that correctly answers the most number of questions. Please, email your answers to: contest@sputnikmail.com Ten days to submit entries, until Sunday 9 December (according to Spanish time). Then I'll announce the winner and send the magazine to him or her. [Note to current HPCC members: Of course you can participate, although you've already got the prize! ;-] Ok, here are the questions: 1. Who suggested to Henry Horn (HP Key Notes Editor) a competition to name the HP-41C programming language? 2. Who wrote the HP-67 owner's handbook? 3. Who founded the HPCC (formerly PPC-UK)? 4. Which famous filmmaker wrote a letter of praise to HP after purchasing an HP-35? 5. Whose was the idea of hiding an Easter egg in the HP48G command RULES? Pretty easy, eh? I hope my poor English won't have caused too much trouble. Good luck! Regards, Bye. Jordi Hidalgo HPCC member #1046 johil@tv3mail.com PS: Remember to send your entry to contest@sputnikmail.com ==== Hi! I m starting to do seriously headaching math and wish to know if my 48G+ can solve this system and how to do so: / x+y=7 | ln(x)-3ln(y)=ln(4)-ln(3) Or perhaps the same kind of system without the ln's would be already great... Pied Ps: i have erable installed if this can help NNTP-Posting-Host: uds207-38.dial.hccnet.nl ==== use solvesys (available at www.hpcalc.org ) Caspar -- Pied Luxi Fere <[ptitpieds]@[hotmail.com] Remove Brackets> schreef in bericht news:9u8gk1$c36$1@wanadoo.fr... can ==== glens@fvsd.ab.ca (Glen T) wrote in message news:<45f00d79.0111280847.63e3e2f6@posting.google.com>... DIMLIMITS is there, but as a flash pointer ^DIMLIMITS, which one calls with FPTR2 ^DIMLIMITS. I've seen and updated list of dispatch types for the '49g here some time ago. Carsten Dominik has compiled a reference of entry points for the '49g, it should be availabe from www.hpcalc.org, or search the newsgroup for the address. Sure that would work. -- Hire the morally handicapped. Eduardo M Kalinowski ekalin@bol.com.br -- **Novo e-mail** **New e-mail** http://move.to/hpkb ==== Hello Glen T wrote in message news:45f00d79.0111280847.63e3e2f6@posting.google.com... DIMLIMITS is still here, except that you have to access it using flash pointer. With the 1.19-6 ROM, you just have to put at the beginning of your source code: FEXTERNAL DIMLIMITS Then use DIMLIMITS in your code as usual, the compiler will replace it with a flash pointer call automatically (just as HPTools would do) A loop is probably the solution Jean-Yves NNTP-Posting-Host: ai05.aula.eis.uva.es Mime-Version: 1.0 ==== Nick Karagiaouroglou escribi227: [...] [...] Maybe you can even do (with the new ROMs): 1: (24*PI*n1+5*PI)/24 @ That is, after your EQ-> NIP ABS PI < 'n1' SOLVE And just the same with the other set of solutions. ==== ?iso-8859-1?Q?M?imo Casta?da? Riloba wrote in message news:<3C06178F.B9036E6A@alumnos.uva.es>... Hi M207ximo! Yes, you are right and your idea is the better one. It is only that the created inequality: |(24*PI*n1+5*PI)| -----------------<=PI 24 cannot be solved for n1 because you get the error message: Parameters not allowed. Obviously the CAS considers in this case PI to be a parameter. But it is possible to get the inequality in the EQW and ->NUM everything except n1. Then the inequation can be solved with SOLVE. Or after do first PI / , and then ABS 1 < 'n1' SOLVE The resulting inequality can now be solved with SOLVE because no parameters are present. Thanks for the very nice idea! :-) Greetings, Nick. ==== It's a kludge but works. Runs an AppleScript which is a stay open application stored in the emu48 folder with the alias in the startup items folder The script uses theAkua Sweets OSAX Scripting Addition to simulate dialog button pushing. If you attempt to run this script as is you may need to tweak the coordinates depending on your monitor resolution. ******** --Script Update emu48 Clock after waking on «event pmgtwake» tell application Finder activate select file emu48 open selection -- bring emu48 to front, discard current instance input state {pointer location:{{445, 258}}} --select Don't save button input state {button down:true} -- press Don't save button --start new instance end tell end «event pmgtwake» *********** If copy&pasting this script ensure the line wraps correctly Questions to gnarlodious AT earthlink.net --  Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit ==== John H Meyers wrote: That is true for binary, but fortunately it is increasing only for the numbers that are generally used rarely. Again, how many times or everyday user of calculator engineer really uses numbers like 2^2000, etc ? With the exception, that it is programmed to keep rational numbers in the fractional form as long as it fits the preprogrammed integer length. Hmmm. I must be missing something in that logic. Did I recently heard, that HP gave up on calculator business ? Just sarcastic. I know Microsoft is not an example of the greatest stuff around, but neither HP in a calculator business recently. Lets face it. HP49 is a piece of junk from the hardware standpoint. It is like HP was trying to get away with keeping around the equivalent of dressed up MS-DOS. HP28 hardware is older than the whole Windows family, and it is only improved several times in a clock speed. In the meantime Microsoft released at least several major versions of Windows that differ from original MS-DOS significantly. This is why sticking with BCD has no meaning to the engineer. Engineers that have to work with real world data rarely need anything beyond 3-4 first digits of the final data to be precise. They do not care if internally it is BCD or binary as long as it runs fast. This is especially not important on the calculator that is not used beyond very short programs. The only real life example of the need of BCD I've seen during this discussion was in the financial industry running millions of accounts, myriad of financial operations and billions of dollars suppose to add up to the last cent at the end of the day. I would like to see some day such real life example running on a handheld calculator. John. Invest $30-40 into 256 MB of RAM !!!!!!! :-) You are incorrect !!! It is neither binary nor decimal. Slide rule is not even digital. It belongs to the old class of analog machines and like them are only binary in nature by the fact that they are built from atoms and voids alternately and have limited number of quantum energy levels altough smeared along the energy spectrum because of the Heisenberg Principle ;-). It is a conversion procedure built into the slide rule (namely drawn markings of short lines with one long every 10-th) that translate decimal input into the analog slide of (binary) atoms voids and then back to the decimal readout for the human interface !!! (I guess, I'm really stretching here, but it must sound scientific for an artistic soul ;-) ) Jacek Content-Transfer-Encoding: 7bit ==== On Sun, 25 Nov 2001 21:29:31 GMT, Jacek Marchel wrote: That is what I was talking about initially, you seem to be talking about something different and the rest of us seem to be talking about yet something else. This debate has gone on so long that I've lost sight of what exactly it is we're debating. Maybe that's the problem : there was no debate in the first place. See my other post to Steen on this subject. Carefully chosen means contrived in this case. Your example proves nothing since you can achieve the same effects from division in binary. That is why I am not using division. That is not the point. I don't know where you've been but I have not been debating that. If I were then I would be full of BS. It takes highschool no, junior high math skills to figure out that binary floating point will be more accurate for the same number of bits. I am at a total loss as to why you think we've been debating something so trivial all this time. I don't know what this is supposed to mean but it is blatantly wrong ( at least for binary values displayed as is. ) Look in any elementary number theory or abstract algebra book as to why binary can't represent most decimal fractions exactly. Or see one of my recent replies to Steen. You STILL don't understand what we are debating. I know binary will be more accurate with respect to division. The debate is about whether simple additions, subtractions and multiplications in binary will lead to problems as opposed to BCD because the user input decimal data can not be represented exactly. Sigh... I am not talking about these operations. I am talking about MULTIPLICATION, ADDITION AND SUBTRACTION. If you can't understand why this matters then you are too far gone for help. Perhaps, but in a user-centric environment when people expect the machine TO DO JUST THIS at least for simple operations it is necessary. I'm sure a student would like it if they got 1999999999999.99998 or whatever on a test instead of 2000000000000. Or if your casual user typed in ( .6 + .7 ) and got 1.2999999999999998 instead of 1.3 . What you are saying is like we should all program in HEX code because it is easier and faster for the machine to deal with. I agree that lesser precision would be sufficient but the fact is that the calculations are still done in 80-bit floating point. Do some research on Borland C++ and the x87 family of floating point instructions. You will then learn that the CPU does all calculations in an 80-bit floating point 8 level stack. It would be pointless ( because the number of clock cycles an operation takes is independent of its precision ) and less accurate to round the result of every operation to 64-bit. See ftp://download.intel.com/design/Pentium4/manuals/24547004.pdf page 187 for more information. To quote the above manual : As a general rule, values should be stored in memory in double-precision format. This format provides sufficient range and precision to return correct results with a minimum of programmer attention. The single-precision format is useful for debugging algorithms, because rounding problems will manifest themselves more quickly in this format. *The double extended-precision format is normally reserved for holding intermediate results in the x87 FPU registers and constants. Its extra length is designed to shield final results from the effects of rounding and overflow/underflow in intermediate calculations.* However, when an application requires the maximum range and precision of the x87 FPU (for data storage, computations, and results), values can be stored in memory in double extended-precision format. - Note the statements in between the stars. Use a long double and I will show you the same behavior with powers of 10 in binary. :) But this point is moot because I am not talking about division and even if I was there would be no debate since binary obviously is more accurate. Yes, I already said that in many posts. This fact IS COMPLETELY OBVIOUS. Why would we be debating it? I withdrew those examples. Obviously you didn't read that either. And I'm not surprised. The fact that you've been trolling on this subject since 1998 ( ! ) is a testament to the fact that no one will change your mind even if they put the concrete proof right in your face. ---------------------------------------------------------------------------- --- Jonathan Busby - Remove the random permutation of NOSPAM from my e-mail address before replying. Content-Transfer-Encoding: 7bit ==== On Sun, 25 Nov 2001 14:20:09 +0100, Steen Schmidt wrote: Yes, and you have the right to be. If it wasn't for your very specific explanations then we wouldn't have pinpointed the source of the confusion. My comments were egotistical and stupid. Anyway, I think this debate is an interesting experiment in human relations. It's an example of what happens when you get many people together who are individually debating something different yet they are unaware of any differences. :D The debate itself is vacuous. It's seems to have taken on a life of its own independent of any concrete argument. When we finally know what we're talking about then there is no debate. Yeah that's right, we are talking about 3 different things here. I'm talking about binary floating point numbers as defined by the IEEE standard displayed as is and the operations of addition, subtraction and multiplication performed on these numbers as is. Others are talking about the same types of numbers but stored at a higher precision and displayed at a lower precision and with special care taken with regard to sensitive functions such as IP/floor. And finally, you are talking about floating point numbers of the form M*10^E where M and E are binary which is *totally* different than the other two subjects. Heh, if someone would have actually taken the time to *look* at the format of your numbers when you were spelling it out instead of just glossing over it then we would have caught this much earlier. ;) Regards, Jonathan ---------------------------------------------------------------------------- --- Jonathan Busby - Remove the random permutation of NOSPAM from my e-mail address before replying. Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit ==== Jonathan Busby wrote: Don't fool yourself. Of course, you do use divison, for the initial value. Based on that intital value you are trying to show me, that using iterations you can enlarge intial miniscule imperfection in the representation of the real value to any desired difference. What I PROOVED to you, is that finite representation has preferred intial values that are represented exactly in one coding and of course imprefectly in the other worse coding. Then using exactly the same trick of multiplying I enlarged this initial small BCD error to the value that prooves BCD is as bad as binary or even actually worse, because it is 1000 times less exact to start with. For example starting with: 1/67108864 value and using only MULTIPLICATION by 671088640 and then SUBTRACTION of 9 will render HP48 nonsense in just 12 repeated steps. Binary will maintain perfect outcome all the time using just 26 bit mantissa (8 BCD digits of internal accuracy). If you want to proove that binary somehow will fail addition and/or multiplication, please use real number that do not have any fractional part because this is obtained by dividing integer by 10. I didn't say displayed as is. That is where you have a problem of understanding. What I have said, is that the binary number added or multiplied by the other binary number is mathematically EXACT, PERFECT sum or multiplication of the initial values unless overflow of the allowed mantissa size occurs. You somehow assumed that binary when does not represent exactly displayed value is then not exact in multiplication or addition. DAAAA. That doesn't require rocket science. EXACTLY the same problem is with BCD. When BCD does not represent particular rational number exactly, then it's multiplication and addition IS NOT PERFECTLY ACCURATE !!! And of course, because BCD is by nature thousand times less accrate to start with, so it's initial error in general real representation will grow much faster in such repetitive processes. This can be easily prooven using simple recursion of using 1/3 as initial for MULTIPLICATION by 3 and SUBTRACTION of 2. Both codings do not represent 1/3 exactly but binary will outclass BCD of the same size easily. 1/3 is perfectly acceptable user input on any known to me calculator. You are somehow misled that the user is capable to input only single real number with decimal point and will never use division or any scientific function. I don't consider using just this as an intelligent benchmark of any calculator including financial. Exactly. That is why they suggest to use 80 bit binary when the extreme precision is required. You are strongly encouraged to find similar statement about using BCD in FPU. If you don't understand that BCD is even more vulnerable to overflow and underflow and to accumulation of error due to the much less accuracy of 80 bit BCD, then I can only be sorry for you. Jack ==== this will probably happen by fall of next year, is what i predict.. any thoughts? will it live long or just become another memory, gone faster than these dot-coms dot-bombs?? ==== Thanks for the help. That works great. And I didn't seem to have any trouble with carriage returns.......... Ed ==== Hi! n4v4jo: Please, see the owner manual instruction in Section 28 (28-1;28-2 and 28-3). Other options is with de emulator Emu48 of Sebastien Carlier (www.hpcalc.org) and create.exe archive, for expand the ramcard.bin memory. Ins this emulator with the ramcard emulator you have a configurate in real mode and can see the ports (1...5) for 512 kbytes. Best Regards from Neuqu216n - ARGENTINA.[:¬) Miguel Angel CAPORALINI HERK **************************************************************************** *** n4v4jo@yahoo.es (n4v4jo) wrote in message news:<7de70cb3.0202110558.1349d0e5@posting.google.com>... ==== Hi Yes Steve, you right but I mean If I have to wire the 512 single chip as a 128 one and will the HP distribute all memory in diferents ports automatically. You know, I wouldn´t like solder the 512 chip and after this maybe the HP will detect just 128. Thank you Hasta la vista etsteve@yahoo.com (Steve Sousa) wrote in message news:... ==== n4v4jo@yahoo.es (n4v4jo) wrote in message news:<7de70cb3.0202120759.aa4fb74@posting.google.com>... This is what i wanted you to read: How to you upgrade your calc using a single 512k chip, V1.00 Note: signal names are as they appear on the HP48 schematic by Marcel Flipse at: www.hpcalc.orgdocsmischp48sch.zip CE1-----------------| | OR ------- 512k ram A17 P2-A17----| | | AND ----| +-| CE2.2---+----------------------- 512k ram A18 nCE0 --- NEG ---| | OR --------| CE1----| | | NAND ---- 512k ram nCE | OR ----| +-| CE2.2--| | | Vcc(on)--------------------+ This gives you the following addressing scheme: A17 A18 IRAM 0 0 Port1 1 0 Port2 0 1 Port3 1 1 Neat, huh? The circuit can be build with only two IC's: a quad OR gate, and a quad NAND gate. You should remove the 32K ram chip from your calc, in its place you solder the 512K chip, you do that by conecting all data lines and address lines A0 to A16 to the 512K chip, you'll be left with two address lines and the chip enable unconected on the 512K chip, you must connect those to the circuit above. It will recognize 128K on ports 1,2 and 3 automatically. You can't solder all pins of the 512K chip on the same pads as the original 32K as the pinouts are not the same. You can however, connect those which are common. Steve ==== etsteve@yahoo.com (Steve Sousa) wrote in message news:... And, of course, like in all upgrades, you must set the card detect pins properly, always high in this case. Steve Sousa ==== Veli-Pekka Nousiainen escreveu na mensagem <9rk705$1ec$1@news.kolumbus.fi>... add: a complete eq lib(+/- 2000 eqns for math, eng, sci, physics, etc) 640x240x256 LCD digital sound 33600 serial 57600 modem network card Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== Hi! Judging from the downloads on my site it looks like GraphWriter is at least useful to some. I know that the manual is a bit terse, so taking up a nice idea of Nick the Greek, I would like to post a few articles about using GraphWriter for doing graph algorithms on the 49G. Anyone interested? If so, I could start this Graph Olympics as soon as the supposedly stable version GraphWriter 1.3 is out. Disciplines, err I mean topics ;), would mostly revolve around graph algorithms written in user RPL, including: 1. The basics (the GraphWriter adjacency list data structure, and a brief overview of the GraphWriter library). 2. Some common graph types and how to create them. 3. Popular embedding algorithms (random, circular, bipartite, grid, spring, ...). 4. Basic graph algorithms (testing for loops, simplicity, bidirectedness, etc.). 5. Graph traversals (depth-first search, connected components, etc.). 6. Basic network algorithms (shortest paths, minimum cost spanning trees). 7. Advanced network algorithms (bipartite matchings, network flows). 8. NP-complete graph problems (graph coloring, cliques). If you have anything to add to this list, just let me know... Also, I still have to program most of these algorithms in RPL, so contributions are welcome. (How about a graph algorithm challenge? ;-) Awaiting your responses. Cheers, Albert -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700 ==== And why not BLOCH & Ford-Fulkerson (flow max)? Your library GraphWriter is great Regards Laurent Damay Albert Graef a 216crit dans le message news: a4bn94$332$02$1@news.t-online.com... least the RPL, spring, still References: Content-Type: text/plain; charset=us-ascii ==== Albert Graef schrieb: [...] [...] Yes, of course I'm interested! BTW, what are the labels (tagged values) good for? Can I use them go give names to nodes which then appear in the graphs? (If it's in the manual, excuse my ignorance - but I haven't been able to find it until now.) Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Mime-Version: 1.0 Content-Type: text/plain; ==== Moin Albert! Albert Graef schrieb im Newsbeitrag news:a4bn94$332$02$1@news.t-online.com... the I'm very interested. Thanks, Micha. Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== Well, network flow was in the list. Regarding BLOCH, could that be a heuristic graph isomorphism procedure? I'll have to look that up, do you have a reference? Thanks a lot. :) Cheers, Albert -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== Thomas Rast wrote: Hi, Thomas! The labels are used to store any kind of additional data associated with nodes and edges, depending on the problem at hand. Such as the length of edges in the shortest path problem and edge capacities and costs in network flow problems. As another example, colors could be stored in a graph as node labels by a node coloring procedure. And yes, you can also use them for the purpose of identifying nodes and edges. Such as in {{ ::Hi, I'm the only node here! }}. I decided to use tagged values for this purpose to make it possible to access different label values associated with a node or edge, using the tags as field names. The GraphWriter library provides the GETLABEL/SETLABEL operations to help with this. Note, however, that at this time the labels are *not* shown in the graphical display. I don't think this would be practical, so for now the labels have to be viewed and edited with the editor's choose/inform box interface, the browser. (But I'm always open to suggestions. If anyone has an idea how to display the labels on a 131x64 LCD without producing too much clutter, I'd be glad to pick it up.) Cheers, Albert -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag ==== Do: / 2 / 1/(x^4) dx / -1 / Calc gives.... -3/8 (must be using blind FTC) Correct answer... divergent... +infinity ????!!!! Wow... couldn't believe today that my calculator was just flat out wrong! I do EVERYTHING by hand before asking Hewy, but this time he was wrong! ;-) Oh well... Man triumphs over machine again, ey? ;-) -Palm Man Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 ==== out was I guess you're using the HP49G? You have set silent mode ON. If you set silent mode OFF, the calc will tell you that it has encountered a singularity, and it'll ask you if you want to proceed with the integral. You can't both have silent and not silent, right? Regards Steen ==== Hi, there, i´m using my calc (hp49) to solve a simple function, 6X^2-8+40cospiX and by hand i got =248, but the result in the calc is a lot diferent (-32) oops, that is when X=6 why? why? why? some times i turn crazy... Content-Transfer-Encoding: 8bit X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 ==== Is this a simple function? What is cospiX? Is it 'COS(pi*Xo)? I assume it is. How did you solve this by the way? SOLVE can't do it. Have you used ROOT or the numerical solver GUI? X=248 is not a solution. It IS 'X' that you're solving for, right (X can't be both -32 and 6 at the same time...)? The four solutions are approximately {'X=1.4616' 'X=0.4456' 'X=-0.4456' 'X=-1.4616'} Regards Steen Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== P.S.: a recent thread discussing these issues can be found here (remove silly line breaks from the link as necessary): http://groups.google.com/groups?hl=en&th=4f3b903514749bd6&rnum=8 -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag References: Content-Type: text/plain; charset=us-ascii ==== Albert Graef schrieb: In rament.S, must be somewhere on hpcalc.org. EQU TOLVar216 80E64 However there's no GARBAGE1 entry listed, maybe you mean this one: EQU GARBSCRATCH1 80FAD Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== Thomas Rast wrote: Thanks for the hint, I found it. :) Hmm, I have no idea what the call to GARBAGE1 in CdB's code is supposed to do, any ideas? Well, of course I can just try it out on the emulator. -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag ==== Hi ! All: Example, for the HP-49G, ROM Revision 1.18 (C): I have a Global variable of {data arguments}, named DBASE and other with the components SORT in ORDER, named COMP, with CHOOSE command. Please, help me. Thanks in advance. Best Regards.[:¬) Miguel Angel CAPORALINI HERK - Neuqu216n - ARGENTINA ==== Hi! Is it possible to get any information, lets say for example the size, of the object in port 2 when it's larger than free memory? Say it's more than 20kB large and I have 20kB free memory. I got ACPTR pointer to this object in port 2. I want to know the exact size of this object. If I try to recall it, the result will be the error message about not enough memory... Is it possible, and if it is then how, to map port2 active and poke inside it by means of assembly to get the interesting information? Robert Tiismus ==== There is something you can do. however: -=THINK VERY SERIOUSLY BEFORE DOING IT=- -=I, FOR ONCE, WOULDN'T DO IT=- -=UNLESS IT WAS MY VERY LAST CHANCE=- -=TO GET MY HEWPIE BACK=- -=YOU ARE THE ONLY ONE RESPONSIBLE FOR YOUR ACTIONS=- . . . here goes: ,remove batteries ,use lots of alcohol (the clear, white kind) on the calc. Use it to dissolve the coffee stuff. Pour it where the coffee is. ,leave to dry for many days ==== Excuse me if this was previously answered, I'm new in HP48... and in this newsgroup too... I'd like to add memory in my 48GX without opening it => I have to buy ramcards. The problem is that ramcards seems not to be sold anymore... Can someone tell me where I can buy some (over the Internet to Paris, France)... Thx Content-Transfer-Encoding: 7bit ==== http://www.cynox.de is one good place for these ramcards. Alan McCornack mccornack@gotmail.com X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 ==== I'm lead to believe that if you contact HP and ask for the cable to enable you to update the flash ROM, they will post it to you free of charge (although I would expect to pay P&P). I bought my (Chinese manufactured)49G in Melbourne and it came with the cable and AUG. ==== Hello, I tried to solve this with a HP49G 1.19-6 ROM INTVX(-k*X-a*sq(x)) This shows only the message Numeric Input. If I write it this way: INTVX(-kX-a*sq(x)) This solution is shown: -(kX*X)-a*(x^3/3) The first part is wrong, must be -(kX^2)/2, second is correct. Is this a bug or am I doing something wrong? Thanks in advance. ********************************** Charly was a chemist, but Charly is no more, for what he thought was H2O was H2SO4. ==== Steen Schmidt wrote: RPN ored The approx mode was on, without it, the result is correct. ********************************** Charly was a chemist, but Charly is no more, for what he thought was H2O was H2SO4. Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 ==== I do not know what you're doing, but I get the correct result in both RPN and algebraic modes. What are your flags? You do not have anything stored in either, 'X', 'a' or 'k'? Regards Steen Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit ==== ==== Raymond Hellstern wrote: And note that the Advanced Users Guide (including the other commands) is available in PDF format there, in case you don't already have it. Also note that pressing the CAT key gives you a choose box of all of the commands, with help for the CAS commands. But seriously, if you're not already familiar with HP calculators and you have an exam in only two days, it may be better to set your 49G aside until after the exam. ==== Wolfgang Rautenberg wrote: I have one more question about the game record: The rotation index and drop column are not sufficient to reproduce the game. sometimes you can put a piece underneath another one. I made a little proggie that records all the keystrokes but that may also not be enough. What if I separate the keystrokes per dropping piece and record the row it settles on? PS how do I get the row? -- This message was written with 100% recycled electrons Pivo ==== Peter Geelhoed wrote: Yes, this is possible and not forbidden. Now I see: Pivo uses the quantum tunnel effect :-) IMHO, the easiest way to solve the arising record problem in this case is to do the following: Activate another key, P say (for Pivo) which adds the 2.5 bytes character P to the global record variable REC before shifting the piece underneath another one. The shifting is characterized by a its own column and rotation index. So in REC may occur a sequence ...-2 3 P 4 1.. to mean the piece was first dropped by first shifting the piece 2 positions to the left, turning it 3x, but before reaching its final position the piece it was again shifted one position to the right and turned 2x, see below. In the indefinite Drop a figure-loop - to be found in the main program after the command sequence GETTOUCH IT - two additional local variables col and rot have to be created which collect the horizontal movements and turnings. P should do the following: :: If P is pressed then append the contents of col and rot *and* character P to REC, and reset col and rot for additional data as long as we are still in the Drop a figure-loop ; It is at any rate easier to verify a single REC list, not two different onces for horizontal movement and rotation extra. The only what one could think of is to create a RECr for each round (=level) extra, r = 1,...,11. You probably mean the column. The initial dropping column depends only on the figure itself, not on a random value (it is written in the ML-code objects I and J ). We should perhaps agree on using the 2.5 byte reals -4,...,+6 instead of the bints 1,...,10 to count the number of horizontal movements of a figure. That makes programming the record and reproducing the game probably easier. - Wolfgang ==== Wolfgang Rautenberg wrote: My favourite is PTR 11320 [48], still wrongly listed as ClrPrgmAnn everywhere, despite being described here: ==== JH> My favourite is PTR 11320 [48], still wrongly listed as ClrPrgmAnn JH> everywhere, despite being described here: JH> http://groups.google.com/groups?hl=es&selm=91cpsj%24hmg%241%40nnrp1.deja.com Thanks for the pointer, I fixed it in my entries list: http://zon.astro.uva.nl/~dominik/hpcalc/entries/hp48g/entries_144.html#IDX32 44 Just so you cannot say everywhere anymore :-) - Carsten ==== Jordi Hidalgo wrote: Sorry, Jordi. I really hadn't noticed it. It's a fine document and I'm sure that it's very useful, although mainly to SysRPL programmers. Now that I've looked at it (well, at least a little), I see that it does tell me what #123h DOERR and #13Eh DOERR are equivalent to. But I don't think that it mentions that these errors evade an IFERR trap. Those errors aren't real errors -- technically they are signals to the system outer loop From In[4]:= abc X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Lines: 13 Message-ID: <3c02995a$0$14653$ba620e4c@news.skynet.be> Organization: -= Skynet Usenet Service =- NNTP-Posting-Host: 62.4.209.151 X-Trace: 1006803291 reader1.news.skynet.be 14653 62.4.209.151 X-Complaints-To: abuse@skynet.be Xref: newsmst01.news.prodigy.com comp.sys.hp48:143884 Hi together, while programing with INFORM I got a problem. After INFORM in an RPN-program the soft-key menu persists: with CANCEL and OK. How can I clear the soft key descriptions? ..Heiko References: <96e03b4b.0111160110.11ecf35a@posting.google.com> <96e03b4b.0111200723.6b6dbaff@posting.google.com> NNTP-Posting-Host: 163.117.12.68 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006773973 20926 127.0.0.1 (26 Nov 2001 11:26:13 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 11:26:13 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143823 nk@imos-consulting.com (Nick Karagiaouroglou) wrote in message news:... Well, I hope that Julian won´t get angry... It was [Humor], you know... Greek means the same in Greece and in Spain? You know, like French Fries in USA and frites in France... NaN5K References: <9tbami$mcp$1@news.kolumbus.fi> <3bfa5747.243015@news.rdc1.az.home.com> <9tjd7u$jvj$1@news.kolumbus.fi> <3BFE763C.278D00A0@iname.com> NNTP-Posting-Host: 212.28.144.137 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006768113 19743 127.0.0.1 (26 Nov 2001 09:48:33 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 09:48:33 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143809 Thomas Rast wrote in message news:<3BFE763C.278D00A0@iname.com>... Sounds like Hercules and the Hydra to me. ;-) Greetings, Nick. X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 X-Original-NNTP-Posting-Host: 210.50.32.9 Message-ID: <3c02e36e$1_1@news.iprimus.com.au> X-Original-NNTP-Posting-Host: 127.0.0.1 Date: Tue, 27 Nov 2001 11:49:25 +1100 NNTP-Posting-Host: 203.134.67.67 X-Trace: news0.optus.net.au 1006822262 203.134.67.67 (Tue, 27 Nov 2001 11:51:02 EST) NNTP-Posting-Date: Tue, 27 Nov 2001 11:51:02 EST Organization: iPrimus Customer - reports relating to abuse should be sent to abuse@iprimus.com.au Xref: newsmst01.news.prodigy.com comp.sys.hp48:143918 Is this all true? :-P http://it.mycareer.com.au/news/2001/11/27/FFX7I564HUC.html Cheers, Gerald. Subject: Re: Automatic unit conversion - not! References: <3BFF2614.4078C9C8@miu.edu> <3BFFEECD.7402A8AF@student.tnw.tudelft.nl> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 213.3.89.178 X-Trace: 26 Nov 2001 12:15:04 +0100, 213.3.89.178 Organization: Bluewin AG Lines: 12 X-Complaints-To: abuse@bluewin.ch Path: newssvr15.news.prodigy.com!newsmst02.news.prodigy.com!newsmst01.news.prodigy. com!prodigy.com!nntp.flash.net!news.tele.dk!small.news.tele.dk!151.189.0.75!n ewsfeed.arcor-online.net!newsfeed01.sul.t-online.de!t-online.de!dns.phoenix-a g.de!news.uni-stuttgart.de!news.belwue.de!news.tesion.net!newsfeed-zh.ip-plus .net!news.ip-plus.net!news.bluewin.ch!stromboli.rastl.ch!nobody Xref: newsmst01.news.prodigy.com comp.sys.hp48:143818 Carsten Dominik wrote: Same here. Perfectly stable. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman References: <0MCK7.698$Lf2.9710@newsfeeds.bigpond.com> <3BFC21D3.F0FCA527@home.com> <9tj7pe$dgr$1@news.kolumbus.fi> <76ea4fd3.0111222106.3936614e@posting.google.com> <7nhM7.72216$gB5.26679@newsfeeds.bigpond.com> Reply-To: Veli-Pekka Nousiainen NNTP-Posting-Host: ui31i8hel.dial.kolumbus.fi X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143885 Jean-Yves Avenard wrote in message news:7nhM7.72216$gB5.26679@newsfeeds.bigpond.com... the AND only now I finally understand this! I never thought that the Saturn would be capable of addressing any single nibble, even in a middle of a register. Well, I surely need to re-read my 71B Internal Documentation... No where did I put in the 80's ??? Veli-Pekka PS: Thanks to all of you for the fun and the final settlement... X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Message-ID: <_VpM7.520869$Lg.21405814@sjcpnn01.usenetserver.com> X-Complaints-To: abuse@usenetserver.com X-Abuse-Info: Please be sure to forward a copy of ALL headers X-Abuse-Info: Otherwise we will be unable to process your complaint properly. NNTP-Posting-Date: Mon, 26 Nov 2001 06:43:54 EST Organization: WebUseNet Corp. http://corp.webusenet.com - ReInventing the UseNet Date: Mon, 26 Nov 2001 05:41:49 -0600 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143825 can one copy the equation library from the 48G or SX card and copy it over to the 49G? References: NNTP-Posting-Host: ppp210-nas1.iperbole.bologna.it X-Trace: pinco.nettuno.it 1006760251 18030 193.207.3.210 (26 Nov 2001 07:37:31 GMT) X-Complaints-To: news@nettuno.it NNTP-Posting-Date: 26 Nov 2001 07:37:31 GMT X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143804 ha scritto nel messaggio news:gl4M7.36680$xS6.60381@www.newsranger.com... Really true! Just a note: as you do later in your post, it's possible to write a symbolic vector directly, e.g.: ['sin(x)' 'cos(y)' 'tan(z)']. I thank you too! ============================================================ Paolo Cavallo I am a teacher at heart, and there are moments in the classroom when I can hardly hold the joy. P. J. Palmer, 1998 paolo.cavallo@iperbole.bologna.it http://www.alberghetti.it/paolo.cavallo/pc.htm ============================================================ EMERGENCY - UNO STRACCIO DI PACE - www.emergency.it References: <9tsrfr$hje$1@pinco.nettuno.it> NNTP-Posting-Host: 212.28.144.137 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006775698 21297 127.0.0.1 (26 Nov 2001 11:54:58 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 11:54:58 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143827 Hi everybody! Fior differentiation of vectors the 49G has many capabilities. For integration of vectors very few. There are many ways to integrate vectors depending on what and how you integrate. 1) Integrate of U(x,y,z)dr where you get a vector out of a scalar. (dr denotes the infinitesimal vector idx+jdy+kdz) I (still) don't know if the 49 has a command that does this, but I use the program: << AXL RISCH AXL >> with the function on stack level 2 and the list of variables on stack level 1. 2) Integrate (i*u(x,y,z)+j*v(x,y,z)+k*w(x,y,z)).dr where you get a scalar out of a vector. ( dr denotes again the infinitesimal vector idx+jdy+kdz ) The result is int(u(x,y,z),dy)+int(v(x,y,z),dy)+int(w(x,y,z),dz). I don't know if the 49 has a built-in command that does this, but I use: << AXL SWAP AXL SWAP RISCH SIGMALIST >> where SIGMA is the greek S used in sums. 3) Integrate (i*u(x,y,z)+j*v(x,y,z)+k*w(x,y,z))xdr where you get a vector out of a vector. We have a cross product of a vector with the infinitesimal vector dr, whic again is idx+jdy+kdz . The result is the vector: i*(int(v(x,y,z),dz)-int(w(x,y,z),dy)) + j*(int(w(x,y,z),dx)-int(u(x,y,z),dz)) + k*(int(u(x,y,z),dy)-int(v(x,y,z),dx)) Again, I think the 49 doesn't do this out of the box. But it can be programed to do so. Greetings, Nick. Paolo Cavallo wrote in message news:<9tsrfr$hje$1@pinco.nettuno.it>... References: NNTP-Posting-Host: 212.28.144.137 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006786283 24351 127.0.0.1 (26 Nov 2001 14:51:23 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 14:51:23 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143836 Stephen Sester wrote in message news:... Only for clarification of the command ->V3 : This command will take three numbers and convert them to a vector. But it doesn't work if the arguments are symbolic. With symbolic arguments you can use either the command ->ARRY or the combination ->LIST AXL. To use ->ARRY, put all components of the vector on the stack, enter the number n of components and then press ->ARRY. This will create an n-dimensional vector. Or put all components of the vector on the stack, enter the number n of components and then press ->LIST and then AXL. The command ->LIST makes a list out of n objects and the command AXL transforms the list to an arry/vector ans vice versa. Vectored greetings, (in x,y,z) ;-) Nick. References: <9sa5me$f32$1@merki.connect.com.au> <56BG7.1595$H82.151909@weber.videotron.net> <9siqec$10q$3@news.kolumbus.fi> <9t05fu$e$1@news.kolumbus.fi> <9t2sco$cgc$1@news.kolumbus.fi> <76ea4fd3.0111161259.6fa892f2@posting.google.com> <9t8i23$mji$1@news.kolumbus.fi> <3BF7E931.B8976B2A@student.tnw.tudelft.nl> <9t8pkl$240$1@news.kolumbus.fi> <3BF7F58A.5806ACA4@iname.com> <9t8tbk$6rn$1@news.kolumbus.fi> <3bff3473.7434925@nntp1.onemain.com> <3BFFC951.8EF08C07@iname.com> Reply-To: Veli-Pekka Nousiainen NNTP-Posting-Host: ui31i8hel.dial.kolumbus.fi X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143892 Bring... VPN, sorry...my english is getting worse Thomas Rast wrote in message news:3BFFC951.8EF08C07@iname.com... wrote: References: <9tea95$qvt$1@wanadoo.fr> NNTP-Posting-Host: 212.22.34.78 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006802316 29485 127.0.0.1 (26 Nov 2001 19:18:36 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 19:18:36 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143877 I've used Erable 3.2 for near two years now, EVERYDAY.(I use the MK version) My free memory, is about 30,000 bytes. Sometimes, I need more space (for some transfers, spliting libraries, for instance) and I must purge it. At the end, I copy the backup from port 2 to port 0 and warmstart. The other Erable libraries, linalg, prep, etc, are in covered ports, running w/o problems. (In fact my 'linalg' has 16 extra commands I've programmed for manipulating symbolic matrix rows and columns, system/matrix conversion, Cramer...) This version of Erable, is faster than the others, and have no many of the bugs I have read about on this posts. It isn't perfect (try to integrate x^(1/3)...)but I think it's absolutly the best for the 48. I'm really certain, that a 48 whith Erable, Alg48 AND MetaKernel (sorry for Java...) is doing its best. Top Level. Regards. Pied Luxi Fere wrote in message news:<9tea95$qvt$1@wanadoo.fr>... X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Lines: 27 Message-ID: <3c0284a4$0$14654$ba620e4c@news.skynet.be> Organization: -= Skynet Usenet Service =- NNTP-Posting-Host: 62.4.209.17 X-Trace: 1006797988 reader1.news.skynet.be 14654 62.4.209.17 X-Complaints-To: abuse@skynet.be Xref: newsmst01.news.prodigy.com comp.sys.hp48:143861 Carsten Dominik schrieb I bought mine for 225$ in germany in 1999. The high price is only due to the vendor. The shop where I bought was not able to give me the AUG, free of charge!! (additional 30 bugs :-( But I get it on a more direct way, free of charge, two years after bying it. This brougt me back a little to good thinking about HP. I have learned: If you want to buy one, first study the prices. If you want to have something more, have the right arguments at the right place. ..Heiko Subject: Re: HP48GX Discontinued References: <65iM7.508652$Lg.20919971@sjcpnn01.usenetserver.com> <99oM7.518247$Lg.21321357@sjcpnn01.usenetserver.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 213.3.89.178 X-Trace: 26 Nov 2001 12:15:05 +0100, 213.3.89.178 Organization: Bluewin AG Lines: 12 X-Complaints-To: abuse@bluewin.ch Path: newssvr15.news.prodigy.com!newsmst02.news.prodigy.com!newsmst01.news.prodigy. com!newscon02.news.prodigy.com!prodigy.com!cpk-news-hub1.bbnplanet.com!nycmny 1-snf1.gtei.net!news.gtei.net!newsfeed-zh.ip-plus.net!news.ip-plus.net!news.b luewin.ch!stromboli.rastl.ch!nobody Xref: newsmst01.news.prodigy.com comp.sys.hp48:143819 Mark S. wrote: I think you mean S&R - Search and Rescue. ;-) Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman References: NNTP-Posting-Host: 212.28.144.137 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006774312 21016 127.0.0.1 (26 Nov 2001 11:31:52 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 11:31:52 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143824 Hi captain JYA! You must understand that all those news about HP are not quite a reason for us (me) to be happy and look at a hopeful future of HP-calculators. Yes, people at HP say that everything will be OK, and yes, they assure that they are not completely (!) abandoning the calcs. But the news (you and others by HP) are telling us that no new HP-calc is taking shape, and that is why we (I at least) are (am) afraid, that we will never again face a new calc made by HP. If no new calc will be made, then the 49 will be too old and obsolete after some years, which means that the line has already taken an end. That's all! :-( How do you expect us to accept this in just a few days? It is not so easy Jean-Yves!!! Greetings, Nick. Jean-Yves Avenard wrote in message news:... References: <3BFD101D.D3AEB201@nokia.com> <4467f306.0111230709.490480ca@posting.google.com> <3BFF3FF2.AB2BE2F8@miu.edu> <7cf00u0gj0vo0lvsqsvl5sn1s7l7cs4072@4ax.com> Reply-To: Veli-Pekka Nousiainen NNTP-Posting-Host: ui31i8hel.dial.kolumbus.fi X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143864 If you want to have a HP made model under warranty in Finland, contact: radix@dlc.fi http://www.dlc.fi/~radix/hp4849nurk.htm Veli-Pekka wrote in message news:7cf00u0gj0vo0lvsqsvl5sn1s7l7cs4072@4ax.com... Subject: Re: HP ROM 1.19+7 release - Was: [VPN] [49] [1.19-6] [BUG] The ROOT Feature, JYA?! References: <9tbami$mcp$1@news.kolumbus.fi> <3bfa5747.243015@news.rdc1.az.home.com> <9tjd7u$jvj$1@news.kolumbus.fi> <9tu10h$rp0$1@news.kolumbus.fi> <3C0289BF.F35D646C@iname.com> <9tua30$arl$1@news.kolumbus.fi> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 62.202.95.245 X-Trace: 26 Nov 2001 22:11:58 +0100, 62.202.95.245 Organization: Bluewin AG Lines: 18 X-Complaints-To: abuse@bluewin.ch Path: newssvr15.news.prodigy.com!newsmst02.news.prodigy.com!newsmst01.news.prodigy. com!newscon06.news.prodigy.com!prodigy.com!enews.sgi.com!diablo.netcom.net.uk !netcom.net.uk!colt.net!newsfeed00.sul.t-online.de!newsfeed01.sul.t-online.de !t-online.de!dns.phoenix-ag.de!news.uni-stuttgart.de!news.belwue.de!news.tesi on.net!newsfeed-zh.ip-plus.net!news.ip-plus.net!news.bluewin.ch!stromboli.ras tl.ch!nobody Xref: newsmst01.news.prodigy.com comp.sys.hp48:143903 Veli-Pekka Nousiainen wrote: Well ... does there have to be a last resort? I'll gladly help and work on the system if it becomes open source or whatever, and I'm sure I'm not the only one. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman References: <9tg5o4$ls9$1@wanadoo.fr> <3BFCF56B.2C9F4677@spam.no> <9tkjai$rge$1@news.wanadoo.es> Reply-To: Veli-Pekka Nousiainen NNTP-Posting-Host: ui31i8hel.dial.kolumbus.fi X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143843 OH -NO 11 It's full of bugs, mosquitos, snakes and alligators !!! :-( VPN, waiting for a fully debugged 1.19-7 Steen Schmidt wrote in message news:dDuL7.3171$TN1.227521@news000.worldonline.dk... 1) References: <96e03b4b.0111210649.4ec6102c@posting.google.com> NNTP-Posting-Host: 163.117.12.68 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006773658 20875 127.0.0.1 (26 Nov 2001 11:20:58 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 11:20:58 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143822 nan5k@yahoo.es (NaN5K) wrote in message news:<96e03b4b.0111210649.4ec6102c@posting.google.com>... Well, the calc must be exactly the same model. rcobo, you are right, batteries lasted 15 years (from mid 80's to late 90's), but it has only 412 bytes RAM. I posted it here because thats the calc i owned before the 48gx and the 49g. References: <9ton2p$gmu$1@pegasus.tiscalinet.it> <9tqs08$doe$1@pegasus.tiscalinet.it> NNTP-Posting-Host: 163.117.12.68 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006775348 21220 127.0.0.1 (26 Nov 2001 11:49:08 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 11:49:08 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143826 rcobo@eng.morgan.edu wrote in message news:... Hey rcobo, don't forget the new guy from Madrid References: <9tp7c8$c6m$1@troll.powertech.no> <3c008bdb.80282939@nntp1.onemain.com> NNTP-Posting-Host: 193.91.152.5 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006773411 20822 127.0.0.1 (26 Nov 2001 11:16:51 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 11:16:51 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143820 cleverjulian@hotmail.com (Julian Fondren) wrote in message news:<3c008bdb.80282939@nntp1.onemain.com>... Sorry for not beeing more spessific in my question. IP Subnet calculators is used to calculate the number of hosts in a Computer IP network. I'm sorry my question sound awkward to people not familiar with the protocols that binds Internet together. I know that someone had the IP Calculator for HP48/28. Reg. Oyvind ps: please respond in English, not in Spanish. Subject: Re: machine language - HELP... References: <262269bf.0111252352.503a9fec@posting.google.com> <9tten0$gdf$1@nreada.inwind.it> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 62.202.90.29 X-Trace: 26 Nov 2001 19:28:26 +0100, 62.202.90.29 Organization: Bluewin AG Lines: 15 X-Complaints-To: abuse@bluewin.ch Path: newssvr15.news.prodigy.com!newsmst02.news.prodigy.com!newsmst01.news.prodigy. com!newscon02.news.prodigy.com!prodigy.com!cpk-news-hub1.bbnplanet.com!news.g tei.net!diablo.netcom.net.uk!netcom.net.uk!newsfeed.icl.net!newsfeed.arcor-on line.net!newsfeed01.sul.t-online.de!newsfeed00.sul.t-online.de!t-online.de!ne ws-ge.switch.ch!news-zh.switch.ch!newsfeed-zh.ip-plus.net!news.ip-plus.net!ne ws.bluewin.ch!stromboli.rastl.ch!nobody Xref: newsmst01.news.prodigy.com comp.sys.hp48:143866 Marco Salvagno wrote: http://www.hpcalc.org/details.php?id=4927 I haven't tried it, but it sounds interesting. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman NNTP-Posting-Host: 209.133.188.60 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006810385 32009 127.0.0.1 (26 Nov 2001 21:33:05 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 21:33:05 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143907 :: GetHeader BINT1 EQ :: COLAITE BINT2 BINT1 ; SetHeader ; bound to 74.21 :: ZINT 1 % -91. DUP xFS? %0<> :: COLAITE :: xCF Matrix ; :: xSF List-of-List ; ; OVER xDISP xFREEZE ; bound to 75.21 and similar on 42.21, except with flag -92. :: ' :: ; xRPLED ; bound to 64.31 :: x->LST ' x>> x+ ' x<< SWAP x+ x->PRG ; which is the inverse of ->XU, and which has an interesting effect on algebraics. References: <3BFC21D3.F0FCA527@home.com> <76ea4fd3.0111222106.3936614e@posting.google.com> <7opsvt8gf6u38art8nm5dimvscn7okgjsl@4ax.com> <3c009181.81729348@nntp1.onemain.com> <9tu4l0$2tm$1@news.kolumbus.fi> NNTP-Posting-Host: 172.147.215.18 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006820069 2510 127.0.0.1 (27 Nov 2001 00:14:29 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 27 Nov 2001 00:14:29 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143917 Veli-Pekka Nousiainen wrote : Ok. And what about an OS with Forth virtual machine and fast just-in-time compiler ? May be MUCH faster than RPL... (inlining code, direct threaded et ceteta...) This mean portability, and possible reuse of the quite big work done on this language. See you gentlemen Stephane Cocquereaumont References: <9tm24a$h0d$1@pinco.nettuno.it> Reply-To: Veli-Pekka Nousiainen NNTP-Posting-Host: ui31i8hel.dial.kolumbus.fi X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143882 HP Jornada 720 with a build-in 56K modem and a scandinavian keyboard. 206MHz StrongARM gives enough fuel to the 49G emulation and I just love the Pocket Excel Veli-Pekka PS: I still regret selling my HP 200LX ;-( Paolo Cavallo wrote in message news:9tm24a$h0d$1@pinco.nettuno.it... NNTP-Posting-Host: 34-coru-x13.libre.retevision.es (62.83.48.34) X-Trace: fu-berlin.de 1006806927 5093534 62.83.48.34 (16 [94894]) X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143895 I have a HP 620LX with a 'large' screen (640x240x256) and 33kbps PCMCIA Modem for Ericsson mobyle phone. It's an old machine but I can use Emu48 running like an original hp48 (it's a Hitachi (SH3) at 75mhz). Pepe Paolo Cavallo escribi227 en el mensaje news:9tm24a$h0d$1@pinco.nettuno.it... References: <3bff7bf5$1@netnews.web.de> <9tqibe$kq1$1@reader07.wxs.nl> Reply-To: Veli-Pekka Nousiainen NNTP-Posting-Host: ui31i8hel.dial.kolumbus.fi X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143894 FROM http://mrityunjay.tripod.com/OS/HPUX/hp48/part1/faq.html I found that http://stolte-edv.com/ but there is no access to the site Try something else from the link after FROM VPN Ensar wrote in message news:9tqibe$kq1$1@reader07.wxs.nl... X Date: Tue, 27 Nov 2001 03:29:19 GMT NNTP-Posting-Host: 63.14.73.247 X-Complaints-To: abuse@onemain.com X-Trace: nntp1.onemain.com 1006831365 63.14.73.247 (Mon, 26 Nov 2001 22:22:45 EST) NNTP-Posting-Date: Mon, 26 Nov 2001 22:22:45 EST Xref: newsmst01.news.prodigy.com comp.sys.hp48:143921 On 26 Nov 2001 13:58:55 -0800, mithrandir@terra.com (Mithrandir) wrote: Um... with no knowledge of ROOT at all, I *promise* you that it accepts local vars. References: <9trr65$15q$1@namru.matavnet.hu> NNTP-Posting-Host: 212.28.144.137 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006773110 20772 127.0.0.1 (26 Nov 2001 11:11:50 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 11:11:50 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143816 vince wrote in message news:<9trr65$15q$1@namru.matavnet.hu>... Hi Vince! You can solve such equations by entering (using your example): COS(2*X+PI/3)=-SQRT(2)/2 then entering X and then using the command SOLVE. The result is: { X=-(24*PI*n1+13*PI)/24 X=(24*PI*n1+5*PI)/24 } The HP49G finds here two families of solutions. Each one has an arbitrary parameter n1 which is integer. That means that all the following expressions are solutions: X=-(24*PI*1+13*PI)/24 X=-(24*PI*2+13*PI)/24 X=-(24*PI*3+13*PI)/24 X=-(24*PI*4+13*PI)/24 ...... You want only solutions in the interval from -PI to PI. You can use OBJ-> on the solution list, to get the two contained solutions as distinct objects on the stack. DROP the 2 (number of list elements) and you have: 2: X=-(24*PI*n1+13*PI)/24 1: X=(24*PI*n1+5*PI)/24 on the stack. Now use EQ-> and then NIP to get the right hand side of the equation on stack level 1. To get a numeric result with different values for n1, you can enter n1, then -3 , then 3 , then 1 and use the command SEQ. This command uses the equation on stack level 5 to create a list of expressions, with the variable on stack level 4 (here n1) going from the number on stack level 3 (-3) to number on stack level 2 (3) stepping by number on stack level 1 (1). When the list is there, use XNUM to get a nummeric result. Now you can see thet the third and the forth number in the list are bettween -PI and PI. The corresponding values for n1 are therefor -1 and 0. You can do the same for the other solution. The corresponding n1 values are again -1 and 0. Note that you can also solve by first setting flag -1 (principal solutions, entering COS(2*X+PI/3)=-SQRT(2)/2 and X , and then using the command ISOL. When flag -1 is set, then ISOL returns only 1 solution, the principal solution, which in this case is 5*PI/24, corresponding to X=(24*PI*n1+5*PI)/24 with n1=0. Generally speaking, such equations have more than one solutions because the trigonometric functions are periodic. There are more than one values for x, where the cosine of x is for example 1 or 0 or any other number between -1 and 1. Hope it helped a little, Nick. References: <3BF460EF.445178A8@home.com> <3BF623FB.5060202@t-online.de> <3BF6821D.2F742B18@home.com> <3BF8575D.5040709@t-online.de> <3BF985EB.37EE261D@home.com> <3BFB22B6.AC5254EB@home.com> <3BFC3777.2040909@t-online.de> <3BFC4CA9.53821A72@home.com> <3BFC5A6A.9090001@t-online.de> <3BFD3582.F6BCA74C@home.com> <3BFFFD75.31A883D2@home.com> <3C0163BB.791611EA@home.com> <3C01705F.95F044E2@home.com> <5NeM7.7213$TN1.726077@news000.worldonline.dk> <3C01C993.3030808@t-online.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Trace: news.t-online.com 1006750350 04 32470 T+XOT-7XSVA3yL 011126 04:52:30 X-Complaints-To: abuse@t-online.com X-Sender: 320094103318-0001@t-dialin.net User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.5) Gecko/20011012 X-Accept-Language: en-us Xref: newsmst01.news.prodigy.com comp.sys.hp48:143800 Albert Graef wrote: Oops, just realized that you're talking about your own custom representation, I should've read the other posts first. :) -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag References: <3BF460EF.445178A8@home.com> <3BF623FB.5060202@t-online.de> <3BF6821D.2F742B18@home.com> <3BF8575D.5040709@t-online.de> <3BF985EB.37EE261D@home.com> <3BFB22B6.AC5254EB@home.com> <3BFC3777.2040909@t-online.de> <3BFC4CA9.53821A72@home.com> <3BFC5A6A.9090001@t-online.de> <3BFD3582.F6BCA74C@home.com> <3BFFFD75.31A883D2@home.com> <3C0163BB.791611EA@home.com> <3C01705F.95F044E2@home.com> <5NeM7.7213$TN1.726077@news000.worldonline.dk> <3C01C993.3030808@t-online.de> <3C01CB35.1090602@t-online.de> Reply-To: Veli-Pekka Nousiainen NNTP-Posting-Host: ui31i8hel.dial.kolumbus.fi X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143850 Albert Graef wrote in message news:3C01CB35.1090602@t-online.de... X Yeah! Steen has binaries with decimal exponents in packed binary Go Steen, Go! When do we have the 49G ROM revised to use Steenary ??? ;-) VPN References: <9thbti$9c0$06$1@news.t-online.com> NNTP-Posting-Host: 212.22.34.78 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1006800329 28920 127.0.0.1 (26 Nov 2001 18:45:29 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 26 Nov 2001 18:45:29 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:143870 I didn't see your answer... Well, I've found the UISTUFF and I have test it on Emu48. I've installed only the BF and CF libraries. All run well if I install both on Port 0. Because Erable and Metakernel, I prefer install B0 in a covered port and then I get a pretty TTRM? when trying use RS 8 (Plot Input form) Help, please. Raymond Hellstern wrote in message news:<9thbti$9c0$06$1@news.t-online.com>... NNTP-Posting-Host: pm589-26.dialip.mich.net Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: cwis-1.wayne.edu 1007007308 20839 207.74.111.132 (29 Nov 2001 04:15:08 GMT) X-Complaints-To: usenet@REMOVE_THIS_TO_E-MAIL.wayne.edu NNTP-Posting-Date: 29 Nov 2001 04:15:08 GMT X-Mailer: Mozilla 4.77 [en] (Windows NT 5.0; U) X-Accept-Language: en Xref: newsmst01.news.prodigy.com comp.sys.hp48:144108 where are the current 49g being made? what is the diffrence if calc is made in china or indonesia? is it the same calc just different cables/manuals? which is better to get china or indo model? where to get a particular model Thank you M17 charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.00.2919.6700 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700 Message-ID: Date: Thu, 29 Nov 2001 02:36:55 +0100 NNTP-Posting-Host: 213.237.13.235 X-Complaints-To: news-abuse@wol.dk X-Trace: news010.worldonline.dk 1006997749 213.237.13.235 (Thu, 29 Nov 2001 02:35:49 MET) NNTP-Posting-Date: Thu, 29 Nov 2001 02:35:49 MET Organization: Customer of Tiscali A/S (World Online) Xref: newsmst01.news.prodigy.com comp.sys.hp48:144090 that Then you're not using polynomials I gather? That's odd when you're an engineer... That depends. Examples please. Regards Steen X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.00.2314.1300 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2314.1300 X-Trace: 377a0a846d648088c77804501686303455654d008306044a062300863c05e6b1 NNTP-Posting-Date: Thu, 29 Nov 2001 07:41:37 GMT Message-ID: <1007019697.27751.1@eurus> Lines: 14 Xref: newsmst01.news.prodigy.com comp.sys.hp48:144134 I'm an aeronautical engineer and rely heavily on differential equations. but its not just that its general algebra. I deal with equations that are so complex I use variables and substitute them in. with the TI one example took about 5seconds yesterday. HP49G 1min30 and the clock stops. I'll find it, I can't remember it off hand but i'll dig it out. that one wasn't even that complex. -- You should quit while I'm ahead. X-Newsreader: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 Message-ID: <8TgN7.53237$e5.14295@newsfeeds.bigpond.com> Date: Thu, 29 Nov 2001 13:19:05 +1100 NNTP-Posting-Host: 144.138.17.253 X-Trace: newsfeeds.bigpond.com 1007000132 144.138.17.253 (Thu, 29 Nov 2001 13:15:32 EST) NNTP-Posting-Date: Thu, 29 Nov 2001 13:15:32 EST Organization: Telstra BigPond Internet Services (http://www.bigpond.com) Xref: newsmst01.news.prodigy.com comp.sys.hp48:144097 Hello Raymond Hellstern wrote in message news:9u3t0l$hud$04$1@news.t-online.com... Funny, that's one of the reason I like MASD, I can put one function (or generally) per line, so it's easy to remember later what it does. Like filling the screen for example, I prefer to put everything on one line rather than using 10 lines. But overall, I don't really mind. Jean-Yves References: <9u3iak$flq$1@wanadoo.fr> <9u3m08$vj5$06$1@news.t-online.com> <3C055CB6.FC9B51B8@iname.com> <9u3t0l$hud$04$1@news.t-online.com> NNTP-Posting-Host: 217.11.167.237 Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: norfair.nerim.net 1007021743 33070 217.11.167.237 (29 Nov 2001 08:15:43 GMT) X-Complaints-To: abuse@nerim.net NNTP-Posting-Date: Thu, 29 Nov 2001 08:15:43 +0000 (UTC) X-Newsreader: Sylpheed version 0.6.1 (GTK+ 1.2.10; i586-mandrake-linux-gnu) Xref: newsmst01.news.prodigy.com comp.sys.hp48:144137 Le Wed, 28 Nov 2001 23:42:58 +0100, Raymond Hellstern a 216crit : You can also put only one instruction per line in MASD, and only use extended instructions like A=A+C A instead of A+C.A. If you want to make readable code, you can also do it with MASD. But remember that the use of blocs in MASD allow to write less useless labels: this is a good way to do a more readable code that you can't use with SASM syntax. Hmmm... When you do cross development, it can be usefull to be able to write ASM files on both plateforms... nearly each line is too much... I'm sure you've got a lot of useless comments like these ones : A=R3 A % We put value X in Aa C=R4 A % We put value Y in Ca C=C+A A % We add value X with value Y ;-p Do you think you can't write comments in MASD ? :-p Bye. NNTP-Posting-Host: pcuevas.vianetworks.es X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Xref: newsmst01.news.prodigy.com comp.sys.hp48:144147 Hello I'm and engineer student from Spain, And I want to buy a Hp49g, and I want to ask if some of us know where can I purchase one HP49G cheap, this site has to do the shipping in my country? THANK YOU... X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Message-ID: X-Complaints-To: abuse@usenetserver.com X-Abuse-Info: Please be sure to forward a copy of ALL headers X-Abuse-Info: Otherwise we will be unable to process your complaint properly. NNTP-Posting-Date: Tue, 27 Nov 2001 22:10:01 EST Organization: WebUseNet Corp. http://corp.webusenet.com - ReInventing the UseNet Date: Tue, 27 Nov 2001 21:07:54 -0600 Xref: newsmst01.news.prodigy.com comp.sys.hp48:144094 well, I hope someone releases a PDA with HP49G functionality!! talk about killer web-user wrote in message news:dIWM7.589981$Lg.23659280@sjcpnn01.usenetserver.com... then Distribution: world Message-ID: References: <76ea4fd3.0111261333.743ff721@posting.google.com> <76ea4fd3.0111261921.5e608244@posting.google.com> <3C038E01.1A5B5EE6@iname.com> <76ea4fd3.0111271612.24218e25@posting.google.com> <3C04AFE0.3AC6FEE9@student.tnw.tudelft.nl> <9u2dpa$ql8$1@news.kolumbus.fi> <76ea4fd3.0111281252.47607dd@posting.google.com> NNTP-Posting-Host: sand.science.uva.nl Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII X-Trace: info.science.uva.nl 1007018583 20948 146.50.10.74 (29 Nov 2001 07:23:03 GMT) X-Complaints-To: usenet@science.uva.nl NNTP-Posting-Date: Thu, 29 Nov 2001 07:23:03 +0000 (UTC) User-Agent: Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Cuyahoga Valley) Xref: newsmst01.news.prodigy.com comp.sys.hp48:144135 JF> Carsten Dominik wrote in message news:... JF> Huh? Maybe he's asked several times; one I read a long time ago is at JF> http://groups.google.com [...] OK, one response. JF> But each new command further restricts the set of valid strings, JF> as they use marker characters. Absolutely. - Carsten X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Message-ID: Date: Thu, 29 Nov 2001 01:29:03 GMT NNTP-Posting-Host: 24.160.90.79 X-Complaints-To: abuse@rr.com X-Trace: typhoon.austin.rr.com 1006997343 24.160.90.79 (Wed, 28 Nov 2001 19:29:03 CST) NNTP-Posting-Date: Wed, 28 Nov 2001 19:29:03 CST Organization: Road Runner - Texas Xref: newsmst01.news.prodigy.com comp.sys.hp48:144093 I just need to know why my HP49 doesn't factor certain problems using the the FACTOR function. The TI-89 my friend has can do many more factoring problems, such as ones with fractional exponents and exponents with variables in them, example: (x^2+2)^(1/2)*(x^2-2)^(-2/3)-(x^2+3)^(1/3) or x^(2a+1)+5x-9 (these don't necessisarily work, but just to give the idea), along with other random ones that just don't seem to work. I really need to be able to use this for my class and the 49 doesn't seem to deliver. Thanks a lot for the help. charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.00.2919.6700 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6700 Message-ID: Date: Thu, 29 Nov 2001 03:01:57 +0100 NNTP-Posting-Host: 213.237.13.235 X-Complaints-To: news-abuse@wol.dk X-Trace: news010.worldonline.dk 1006999253 213.237.13.235 (Thu, 29 Nov 2001 03:00:53 MET) NNTP-Posting-Date: Thu, 29 Nov 2001 03:00:53 MET Organization: Customer of Tiscali A/S (World Online) Xref: newsmst01.news.prodigy.com comp.sys.hp48:144095 to The HP49G should be at least as good as the TI89 for factoring - please provide a real example that don't seem to work on the '49, but works on the TI89. Regards Steen References: <92e02401.0111281138.7949638@posting.google.com> NNTP-Posting-Host: 63.14.73.2 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1007006393 19974 127.0.0.1 (29 Nov 2001 03:59:53 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 29 Nov 2001 03:59:53 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:144106 mithrandir@terra.com (Mithrandir) wrote in message news:<92e02401.0111281138.7949638@posting.google.com>... Neither have I =) But I haven't tried the latter at all. (this following paragraph isn't at all helpful to you; it's just a meme that interests me; you might want to skip it.) OK. I'm aware that some people prefer to do all their programming for their highly portable and mostly self-contained calculator on a computer that is either A) fairly well rooted in one spot, or at least heavy to carry everywhere they have their calculator, or B) far more expensive to use and carry around than their HP48G. I am understanding of this; some people have normal lives and can actually use their computer very frequently, without worrying about niddly things like having a computer actually accessible or near. I myself have never programmed for my calculator on a PC, and couldn't easily tell you how to do so. If you're open to an alternative, here's what I and some others do: we program the calculator on the calculator, using assemblers on the calculator -- and not just UserRPL! we also have SysRPL and ML. See the built-in library 257 for the basic ML and SysRPL capabilities that come with your calculator. If you plan to do a lot of SysRPL programming, I would suggest that you find off hpcalc.org and download (to your calculator!) the following libraries: extable Emacs OT49 Keyman extable is important because, without it or a mockery of it, SysRPL is really no fun at all. Emacs is important for a lot of reasons, but for now just realize that it eases the writing of SysRPL programming. OT49 makes library creation and manipulation and investigation very easy, and has a number of other interesting functions. Keyman, besides being a convenient keyboard-managing device, also has a very useful function that gives you what keys have natively assigned to them (ever want to put the MODE button somewhere more convenient for you? Get Keyman). You can also get functionality for doubleclicking keys and long-pressing keys, which is both interesting in implementation and sometimes useful. If you don't know how to program SysRPL, see www.hpcalc.org and *both* Programming documentation sections (that is, the 48GX section too) -- just be aware of important differences like entry points, or else be TTRM. Here's a simple SysRPL program that you might use to write further SysRPL programs, once you have Emacs: :: NULL:: xRPLED ; It will start emacs with you editing a secondary, which will be recompiled upon exit (and you'll get a chance to catch and deal with compilation errors 'as they happen', so at least you'll know immediately that your program *compiles* (but whether it's safe... that's another story)). Note that the next version of Emacs will have a long-press enter that will compile-and-exit in the manner that you get with editing a secondary -- just use long-RPLED, I think it is, to start out with a nice template for writing SysRPL programs. Oh! and also be *very sure* to get Carsten's entries document. It's a good reference for SysRPL commands, and also one way to learn a bit about SysRPL in the grande olde experimental sort of way. Watch out for the TTRMs. References: <92e02401.0111281138.7949638@posting.google.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: news.t-online.com 1007013340 03 11327 lLrlTWwTSbAB53 011129 05:55:40 X-Complaints-To: abuse@t-online.com X-Sender: 320094103318-0001@t-dialin.net X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Xref: newsmst01.news.prodigy.com comp.sys.hp48:144122 Hi, Mithrandir, on the 49G itself I also use Emacs from Carsten Dominik, which is very convenient. You can use it to disassemble and edit existing SysRPL code, and create your own. As Julian has pointed out you need to install the extable library. Moreover, the SDiag library included in the Emacs distribution allows you to view stack diagrams of entry points in the editor. I'd also suggest to install Nosy or CQIF, which allows you to explore the ROM inside Emacs (very instructive!). All this stuff is available on www.hpcalc.org. But don't give up on the HPTools just yet, they're great for doing PC development. I use them together with XEmacs and GNU make, which makes managing bigger projects a lot easier. (There is yet another alternative for the 49G. If you don't like command line utilities, take a look at the Debug2 SDK. I prefer the HPTools, though.) There's one problem getting started with the HPTools, namely that the necessary stuff is scattered out all over hpcalc.org. ;-) First make sure you get the latest version of the executables (search for hptools-3.0.6-win32.zip on hpcalc.org) and install them on your PATH. For linking your program, you'll also need a suprom file (e.g., there's a suprom49.a for the 49G). I don't know from the top of my head where I got mine, but I think it is included in the Debug2 package mentioned above. For the docs you have to look in other (older) HPTools packages. You can find all the text docs in HORN4.ZIP, and some docs in Word format in HP4W4DOC.ZIP. Once you collected all the stuff pointed out above, compiling a simple SysRPL program from a single source file is really not that difficult. First you need to assemble your suprom file, e.g.: sasm suprom49.a (You can skip this step if you already have the suprom49.o file.) Then you'll just have to invoke the RPL compiler, the assembler and the loader (a.k.a. linker), in that order. E.g.: rplcomp example.s example.a sasm example.a sload -H -o example example.m Note that the loader is driven by a loader file (example.m in this example), this is a simple text file which lists some options, the object file to be linked (example.o) and the suprom file to be used. In this example, the loader file might look as follows: SU XR SU SE suprom49.o RE example.o Of course the whole build process can be automated using, e.g., GNU make. Once you have the example program file, you can simply drag it to the emulator (emu48) and run it. I prefer this to downloading to the real calc directly, because I can always quickly recover everything on the emulator if something goes wrong (TTRMs ;-). If you're still interested now, I can send you a working example and a zipball with the necessary stuff I collected from hpcalc.org, so you can try it for yourself. ;-) If there's enough interest, I could also upload that to hpcalc.org. Hope this helps, Albert -- Dr. Albert Graef, Dept. of Music-Informatics Johannes Gutenberg University Mainz, Germany Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de http://www.musikwissenschaft.uni-mainz.de/~ag References: <92e02401.0111281138.7949638@posting.google.com> <20011128154039.09433.00002057@mb-fy.aol.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: news.t-online.com 1007017261 00 30072 hrJlT8FXSAY1l+ 011129 07:01:01 X-Complaints-To: abuse@t-online.com X-Sender: 320044523859-0001@t-dialin.net X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.50.4522.1200 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4522.1200 Xref: newsmst01.news.prodigy.com comp.sys.hp48:144130 Hi, source You could automate even the copy process and more. To transfer a file to Emu48, you could use DDE48 by Christoph Giesselink or just drag the file with the mouse and drop it into Emu48;-) Raymond References: NNTP-Posting-Host: 199.46.200.230 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit X-Trace: posting.google.com 1007001595 18793 127.0.0.1 (29 Nov 2001 02:39:55 GMT) X-Complaints-To: groups-abuse@google.com NNTP-Posting-Date: 29 Nov 2001 02:39:55 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:144098 rcobo@eng.morgan.edu wrote in message news:... laplace of f(x - a) = exp(-a*s)*F(x) <- correct? laplace of sin(x) = 1 / (s^2 + 1) so... laplace of sin(x + pi) -> exp(pi*s) * L{sin(x)} -> exp(pi*s) * (1 / (s^2 + 1)) ? or am I taking the Laplace incorrectly? exp(pi * s) /= -1 <- right? please correct me if I'm doing this wrong. Dave References: <5264ac2d.0111281839.1f33a9f6@posting.google.com> NNTP-Posting-Host: mix-grenoble-103-2-86.abo.wanadoo.fr Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: wanadoo.fr 1007015451 4071 193.250.90.86 (29 Nov 2001 06:30:51 GMT) X-Complaints-To: abuse@wanadoo.fr NNTP-Posting-Date: 29 Nov 2001 06:30:51 GMT X-Mailer: Mozilla 4.72 [fr] (X11; I; Linux 2.2.16-9mdk i686) X-Accept-Language: en Xref: newsmst01.news.prodigy.com comp.sys.hp48:144125 No, this is not correct. More precisely for the formula above to be correct you need to multiply f(x-a) by an appropriate Heaviside function, if I'm not wrong: f(x-a).1_{x>a} X-Abuse-Info: forward the entire news article including headers or X-Abuse-Info: else we will not be able to process your request X-Complaints-To: abuse@newsranger.com NNTP-Posting-Date: Thu, 29 Nov 2001 01:58:16 EST Organization: http://www.newsranger.com Date: Thu, 29 Nov 2001 06:58:16 GMT Xref: newsmst01.news.prodigy.com comp.sys.hp48:144128 hello Dave, sin( x + Pi ) = -1 * sin ( x ) because it is a periodic funtion ( and the shift is Pi ), it will mirror with respect to the origin, making it the negative of the sin( x )) also, cos( x + Pi ) = - cox( x ) now, when you do the integration by parts of the inproper integral, the -1 comes out as a constant. int(o, infinity, exp( -s*t ) * sin ( t ) * ( -1 ), t) where u = sin( t ) dv = exp( -s*t )dt you can also obtain the Lap of sin(x) from a table of transforms and multiply your result by -1 (comes out of the integral) cheers In article <5264ac2d.0111281839.1f33a9f6@posting.google.com>, davidj says... Subject: Re: Well, it's over then...Good Bye. References: <3BE89812.AB352522@home.com> <9sa6bd$8gd$1@news.kolumbus.fi> <8hLG7.28712$8x1.5582@newsfeeds.bigpond.com> <7osoutkted5698d1ici0n2lbpotqk8mdme@4ax.com> <3BF6EF64.7A04165C@home.com> <2953da04.0111240234.1d84794f@posting.google.com> <3C0001B9.3DBB8A43@home.com> <3C00480A.ED7335D2@miu.edu> <3C0065DE.2010506@t-online.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 208.145.205.253 X-Trace: corp.newsgroups.com 1007004732 208.145.205.253 (28 Nov 2001 21:32:12 -0600) Lines: 65 X-Comments: This message was posted through Newsfeeds.com X-Comments2: IMPORTANT: Newsfeeds.com does not condone, nor support, spam or any illegal or copyrighted postings. X-Comments3: IMPORTANT: Under NO circumstances will postings containing illegal or copyrighted material through this service be tolerated!! X-Report: Please report illegal or inappropriate use to X-Abuse-Info: Please be sure to forward a copy of ALL headers, INCLUDING the body (DO NOT SEND ATTACHMENTS) Organization: Newsfeeds.com http://www.newsfeeds.com 80,000+ UNCENSORED Newsgroups. Path: newssvr16.news.prodigy.com!newsdst01.news.prodigy.com!newsmst01.news.prodigy. com!newscon02.news.prodigy.com!prodigy.com!cpk-news-hub1.bbnplanet.com!news.g tei.net!newsfeed1.cidera.com!Cidera!local-out2.newsfeeds.com!corp.newsgroups. com Xref: newsmst01.news.prodigy.com comp.sys.hp48:144105 Steen Schmidt wrote: Only in the sense that, *on*the*average* (as A.G. says), the [normalized] binary mantissa has only a 2:1 largest:smallest ratio range (if exponent is true binary and not the sometimes-used base 16, as on old IBM mainframes, which is worse than base 10), whereas decimal floating-point has a 10:1 ratio, so the average percentage error for one count in the last decimal digit is less than for one count in the last binary bit :) But still, calculations which are *exact* in decimal are those which have *zero*error*, and if these values are not representable *exactly* as binary fractions, then *any* amount of error in binary is greater than in decimal! Another perspective on the whole thing (which I may be repeating from another prior answer which I'm going to post later on :) is that all the representable normalized floating-point numbers, given a fixed base and a fixed number of mantissa digits, form a lattice, a subset of all real values; if one carries on transactions over a certain other lattice field (e.g. metric, decimalized currency transactions), such quantized transactions fit some of those lattices a little more perfectly than others -- especially, decimally quantized processes like currency transactions are always *absolutely*exact* in a decimal FP calculation system, provided that they don't need more *significant*digits* than the mantissa. Conveniently for all calculators of the past, the use of decimal FP also eliminated the need for base conversion for every data entry and display, which must have been a great ROM saver, especially on those Sinclair models :) By the way, various 10-digit HP calculators of not so many series ago used to get some imperfect answers (9.9999 etc.) for, say 10^44 (or some higher powers, using the x^y function rather than ALOG), whereas old Casio calculators of the same era *never* embarrassed themselves in this way. Why was this so? Because HP did all of its x^y calculations using *natural* logarithms (base e), whereas Casio did them using *base-10* logarithms -- that is, HP calculated base-10 logarithms indirectly from fundamental natural logarithms, while Casio elected to do exactly the reverse, calculating natural logarithms indirectly from fundamental base-10 logarithms, and it was a no-brainer for Casio to correctly calculate the base-10 logarithm of 10 as *exactly* 1, and the base-10 antilogarithm of 44, say, as *exactly* 1E44! So that was much the same thing -- a certain lattice which better fit no-brainer problems gave perfect answers all the time for Casio, while HP fans had to defend their brand by pointing out other, more subtle things which it could do better :) And that little bit of PR might still sway calc makers for some while, I'd guess :) [r->] [OFF] -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 80,000 Newsgroups - 16 Different Servers! =----- Subject: Re: Well, it's over then...Good Bye. References: <3BE89812.AB352522@home.com> <9sa6bd$8gd$1@news.kolumbus.fi> <8hLG7.28712$8x1.5582@newsfeeds.bigpond.com> <7osoutkted5698d1ici0n2lbpotqk8mdme@4ax.com> <3BF6EF64.7A04165C@home.com> <2953da04.0111240234.1d84794f@posting.google.com> <3C0001B9.3DBB8A43@home.com> <3C00480A.ED7335D2@miu.edu> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit NNTP-Posting-Host: 208.145.205.253 X-Trace: corp.newsgroups.com 1007024877 208.145.205.253 (29 Nov 2001 03:07:57 -0600) Lines: 183 X-Comments: This message was posted through Newsfeeds.com X-Comments2: IMPORTANT: Newsfeeds.com does not condone, nor support, spam or any illegal or copyrighted postings. X-Comments3: IMPORTANT: Under NO circumstances will postings containing illegal or copyrighted material through this service be tolerated!! X-Report: Please report illegal or inappropriate use to X-Abuse-Info: Please be sure to forward a copy of ALL headers, INCLUDING the body (DO NOT SEND ATTACHMENTS) Organization: Newsfeeds.com http://www.newsfeeds.com 80,000+ UNCENSORED Newsgroups. Path: newssvr16.news.prodigy.com!newsmst02.news.prodigy.com!newsmst01.news.prodigy. com!newscon02.news.prodigy.com!prodigy.com!newsfeed.cwix.com!sjc-peer.news.ve rio.net!news.verio.net!newspeer.cwnet.com!local-out2.newsfeeds.com!corp.newsg roups.com Xref: newsmst01.news.prodigy.com comp.sys.hp48:144146 Steen Schmidt wrote: Say, what are the first 12 significant *decimal* mantissa digits of the very simple-looking in binary floating point value 2^32768, anyway? Or even 2^2048 ? The time to calculate even the leading decimal digits increases with the magnitude of the exponent; simple direct conversion algorithms (even those retaining only the most significant digits) might take time roughtly proportional to the logarithm of the value, which means proportional to the magnitude of the exponent; if you convert using logarithms/exponentials instead (at expense of more internal tables and program), then the time could be more uniform and bounded, but larger exponents still require much more accurate logarithms etc.) I wouldn't know; I thought that Jacek was implying that Windows Calc [WC] uses the internal [Intel?] FP, but I certainly don't know myself! Try watching its process' memory usage with WinNT/Win2K Task Manager while you compute -- it grows quite a bit, but who knows for what? If the HP48/49 had used that much scratch memory (or even ROM for the program, for such a ridiculously simple calculator), they'd have lost money very rapidly ;-) Anyway, with Windows Calc, it looks very superficially as if it might take time proportional to the exponent, perhaps doing all its conversion calculation using BCD in memory strings (but I have neither timed nor plotted any regression fit to this data :) What, by the way, is the upper limit to the results of WC? If it turns out that it *isn't*bounded* at all, then you have a case for it actually being in exact mode all the time ;) BTW, what's the internal range of the Intel FPU? Eh? Starting from any idea of how accurate one wants to be, one can then choose a mantissa size to provide that accuracy, but once one has done this in a fixed-size system, then one is committed, and the next model of the competitor can then leapfrog whatever you've got :) How many everyday engineers (chemical, mechanical, electrical, structural, actuarial, social, ...) would ever get significantly different results for continuous-function answers to practical problems on a 10-digit calc than on a 100-digit calc? If they do, it's amazing how they ever got by with slide rules and log tables (plus those mechanical desk clunkers for accountants only :) back in the stone ages (well into the 1970s?) HP used to make the best-designed calcs with very stingy memory, while TI used to stuff theirs with more registers than you could ever use, as I recall. Anyway, on the sole criterion of how many digits it keeps, my very best calculator is my OZ-707 Basic card, plugged into my Sharp Wizard OZ-7000 [20-digit display from 24-digit internal calculations, IIRC] And hey, it has much better Basic than today's HP Basic ;-) One more thing -- note that HP has an internal long FP (15-digit mantissa, 5-digit exponent) and that only final function results are *rounded* back to (12,3) again. Both the internal expanded-range calculation and the final unbiased rounding are important, and will remain important even if binary FP is substituted for decimal FP; there are also some remarkable extra lengths that HP has gone to, to avoid carelessly losing significant digits: http://groups.google.com/groups?selm=5ruis3%2480v%241%40news.iastate.edu http://groups.google.com/groups?th=c5ffacd653fa02d2 [entire thread] HP's unbiased rounding is also the necessary basis for DIGI24 (double-precision HP4x arithmetic, found on Goodies Disk #10/Math): http://www.hpcalc.org/hp48/compilations/horn/horn10.zip Truncating FP calcs (practically everybody else) tend to wander monotonically downward from correct results in 3 out of 4 basic arithmetic operations, whereas rounding (of originally longer internal results) helps greatly reduce or eliminate the magnitude of such wanderings, per operation, in the long run. I'll bid two bits! And gee whiz, my old HP calcs held one whole command per *byte* (rather than 2.5 bytes); weren't those old calcs much better, then? And Gates was right, 640K was once plenty for me, with MS-DOS, but now I get System process out of memory almost as soon as I boot up NT on this crummy 80MB desktop I'm typing on :) As far as that 20% extra packing density, however, FWIW, I recently tried out Cab Pack, which makes MS-compatible cab[inet] files [readable by WinZip] rather than zip files (Stuffit's Drop Stuff may also be able to make cab files), and I found that, sure enough, on a few samples of my favorite archives, the best-compressed cab file was about 20% smaller than the best-compressed zip file! So, are we all going to switch to cab and give up zip ? If so, get Cab Pack here (note the version 1.4a update): http://home.t-online.de/home/lars.hederer/english.htm Caution -- may use a hidden MS-DOS window, invoking Microsoft's own traditional cabinet-file builder program, so will that be DOA on XP? (anybody want to get new Linux without a shell?) I think you refer to the mantissa range effect, as described in my previous post (later in this thread :) You threw in the shorter sign effect as well, but I have already acknowledged, two posts ago, that HP calcs use up 4 bits for a sign (because of memory alignment), just as the stupid HP4x series wastes 20 bits on every object type prefix, even though there are darned fewer than 2^20 object types :) Then we also discussed that with calculations which happen to be *exact* with decimal mantissas, you will never get in trouble using those same decimal mantissas, but you could in binary (and likewise, certain large powers of 2 would be perfectly precise in binary but not with decimal mantissa, which is excellent for civilizations that have two fingers :) Let me add one more tidbit here, which is the heretofore unexplored slide rule inaccuracy effect: Positioning (and dimensional) inaccuracy causes errors when using a slide rule. Now, you can, by similar logic to what you have been using, show that a slide rule marked for the *binary* floating-point number system will be that much more accurate than the more familiar slide rule based on decimal mantissas, for the same reason: its mantissa range, end to end, would be 2:1 rather than 10:1, and the exact same linear slippage (on same-length slide rules) would therefore represent less of a percentage error! So why have I never seen a binary-based slide rule? More ideas for greater efficiency (a very old, copied article): http://www.lesbell.com.au/Home.nsf/3d8a910f57b615d44a2565ea0006041e/27605611 4efe6c834a256799007eceb4?OpenDocument Merry Christmas [shopping] to all, and to all a good night [r->] [OFF] -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 80,000 Newsgroups - 16 Different Servers! =-----