6344 === Subject: Re: Think Lithium! posting-account=aJhvMAoAAABqICGF80eSUn6d3D9SANQE Gecko/20080201 Firefox/2.0.0.12,gzip(gfe),gzip(gfe) > I fear the humour is not so easy to see if your mother language > is German. .-) Sorry, I should have explained. > But i'll try to explane it: torch is one word for the german Taschenlampe - but only in british > english. > In american english torch ist a person who lights something in order to > destroy it, and to torch means just that: to light s.th. to destroy. Teh correct word fpr Taschenlampe in american english is flashlight Volker > who is, like you, NOT an english person ;) This is correct, but for americans, the most common use of torch as a noun always refers to something that has an open flame, such as the Olympic Torch, or what you would see being carried through the jungle, or a gas torch used in welding. Your usage was perfectly correct, it's just that with all the news reports, the thought of trying to create a flaming torch from your laptop battery struck me as being funny. -Wes Loewer (or L.9awer as my Gro§papa spelled it, uh, spelt it) === Subject: Re: Think Lithium! > This is correct, but for americans, the most common use of torch as > a noun always refers to something that has an open flame, such as the > Olympic Torch, or what you would see being carried through the jungle, > or a gas torch used in welding. I know, but i just forgot to mention this. > Your usage was perfectly correct, it's just that with all the news > reports, the thought of _trying_ to create a flaming torch from your > laptop battery struck me as being funny. ;) That's why i referred to flashlight instead of torch. >Wes Loewer (or L.9awer as my Gro§papa spelled it, uh, spelt it) German too? Volker -- Stop TCPA/TCG now! === Subject: Re: Think Lithium! >Wes Loewer (or L.9awer as my Gro§papa spelled it, uh, spelt > German too? In heritage, yes, but the only German I heard growing up was when my parents used it as the secret parent language. :) -wes === Subject: Re: HPGCC related ... posting-account=aJhvMAoAAABqICGF80eSUn6d3D9SANQE Gecko/20080201 Firefox/2.0.0.12,gzip(gfe),gzip(gfe) > Even MS abandoned the long double years ago (at least on the x86 > platform, don't know on others), and now long double is mapped to > double, to add more confusion. I seem to recall that this happened when they moved to the 32-bit Windows compiler, probably with Windows-95. It seemed a shame not to take advantage of the 80-bit floating point hardware that was there for the using. At the time, I was doing a lot of fractal programming. Using 80-bit floats instead of 64-bit allowed a few more zooms before having to resort to slower software code. Your comment got me curious about which x86 compilers still support the 80-bit long doubles. Of the current compilers that I could get me hands on: gcc (and variants like Cygwin, MinGW), Intel, Digital Mars, and Turbo C++ (it's back) all support 80-bit long doubles in hardware; MS and Open Watcom do not. I gather that ARM gcc only goes up to 64-bit floats, so it makes sense for hpgcc to do the same. This brings to mind what may be a silly question, but please pardon my ignorance. I know there is a _lot_ of code in hpgcc that is specific to the 49g+/50g. But there is also a lot that is standard to any C compiler (math libraries, std i/o, etc.). How much of ARM gcc's standard library had to be rewritten for hpgcc? For instance, the scanf() function that we've been discussing. Was there already a glibc scanf() for ARM, or did this have to be written from scratch? I see from the source code that some of the fast IEEE math routines are present while others are not (such as hyperbolics, expm1, log1p). Would these have to be written from scratch, or do they exist in an ARM glibc? Can you just recompile glibc (with a few tweaks for the mixed endian floats), or am I oversimplifying the process? -wes === Subject: Re: HPGCC related ... > Your comment got me curious about which x86 compilers still support > the 80-bit long doubles. ÊOf the current compilers that I could get me > hands on: gcc (and variants like Cygwin, MinGW), Intel, Digital Mars, > and Turbo C++ (it's back) all support 80-bit long doubles in hardware; > MS and Open Watcom do not. Since the math coprocessor (or FPU unit for Pentium and up) works internally in 80-bit, it really makes sense to use the full precision for user types as well. I don't know why they dropped it, but they did. I gather that ARM gcc only goes up to 64-bit floats, so it makes sense > for hpgcc to do the same. This brings to mind what may be a silly question, but please pardon my > ignorance. ÊI know there is a lot of code in hpgcc that is specific > to the 49g+/50g. ÊBut there is also a lot that is standard to any C > compiler (math libraries, std i/o, etc.). ÊHow much of ARM gcc's > standard library had to be rewritten for hpgcc? How much? Everything. We didn't take anything from the standard libraries, except the floating point emulation routines. Problem is, that functions in the standard libraries that come with gcc are heavily interconnected. For example, the math functions throw exceptions, wich pulls the entire exception machinery from the libraries, taking a lot of space. Executables were really big doing it that way. We had to rewrite routines to avoid linking in those extra modules. If you need certain specific routines, you can still link against newlib and it will retrieve the needed modules, provided you link newlib *after* you link against hpgcc libraries. This way it will first take our routines, then only take what's really needed from newlib. To give you an example, when I was trying to get C++ running, there was no 'operator new' defined. Since it's just one routine, I decided to link against libsupc++ and use the standard... big mistake! It worked fine, except that Hello world ended up taking 34 kbytes of normal size (less than 1 kbyte). The particular example of scanf, if I'm not mistaken it uses streams, so it brings several other modules from newlib. It had to be 'stripped down' in order to make it fit in the calc. It is unfortunate that those generic libraries are designed for much bigger devices than our modest calculator, and they are not written as a set of independent routines, but thinking on the library as a whole. That's just not what we needed for HPGCC. Claudio === Subject: Re: HPGCC related ... > How much of ARM gcc's standard library had to be rewritten for hpgcc? How much? Everything. Hang on a second while I bump you guys up a few pegs on my admiration scale. All the more impressive what the hpgcc team has accomplished. me. -wes === > If you neglect to hold the shift key while trying to access > a special top-row graphing function in RPN mode... > [becomes an accidental STO into variables with VAR menu] Various defenses against accidentally overwriting variables were discussed in this thread: how i lose programs [Feb 2007] More ancient: Safer top-row App keys [March 2001] No doubt the ideal way is to just put the calculator in ALGebraic mode, as the Gods of HP management had originally intended it to be used :) === > If you neglect to hold the shift key while trying to access > a special top-row graphing function in RPN mode... > [becomes an accidental STO into variables with VAR menu] > I had forgot to hold the shift key. guilippo === Subject: Re: input command > DIR/FILE.TXT Use either a ZINT or a string for tagging. 3.00 > R->Z > ->TAG or 3 The extra zeros are what is messing it up. TW I choosed the second way. It works. === Subject: (hp50g) HPGCC and User RPL I discovered HPGCC yesterday and so i though I can speed my project using HPGCC. The project ist a mechanical engineering program to calculate the ISO tolererances of holes and rods. For example input: 100. H7/g6 output: Hole 0 -7 um; Rod -5 +9 um (the values are from fantasy) also in C. So I would like to reuse the C code. My question is: How can I manage the input/output on the HP50g with a C programm? Should I mix C compiled functions in a User RPL program? Someone can give me hints or suggestions? guilippo === Subject: Re: (hp50g) HPGCC and User RPL > I discovered HPGCC yesterday and so i though I can speed my project > using HPGCC. > The project ist a mechanical engineering program to calculate the ISO > tolererances of holes and rods. For example input: 100. H7/g6 output: Hole 0 -7 um; Rod -5 +9 um (the values are from fantasy) also in C. So I would like to reuse the C code. My question is: > How can I manage the input/output on the HP50g with a C programm? > Should I mix C compiled functions in a User RPL program? Someone can give me hints or suggestions? > guilippo Check out http://sense.net/~egan/hpgcc/. There are a lot of examples. === Subject: Re: (hp50g) HPGCC and User RPL OpenPGP: id=14137F7D; url=http://hpgcc.org/pgp_key_14137F7D.txt > I discovered HPGCC yesterday and so i though I can speed my project > using HPGCC. > The project ist a mechanical engineering program to calculate the ISO > tolererances of holes and rods. For example input: 100. H7/g6 output: Hole 0 -7 um; Rod -5 +9 um (the values are from fantasy) also in C. So I would like to reuse the C code. My question is: > How can I manage the input/output on the HP50g with a C programm? HPGCC has an extensive interface to the HP 50's stack. Please refer to the sat_* functions in the documentation. You could also read variable values via console style functions, but I strongly recommend using the stack interface. > Should I mix C compiled functions in a User RPL program? While this is technically well possible, it's probably not advisable. Simply stay with a pure C program, unless you have good reasons for a mix. > Someone can give me hints or suggestions? -- Ingo Blank http://hpgcc.org http://blog.hpgcc.org Public Key ID: 14137F7D http://hpgcc.org/pgp_key_14137F7D.txt Fingerprint: FEA5 B074 1EB2 29B3 1FAB CFD3 AFD6 E2D0 1413 7F7D 'Reply-To' address. === Subject: Re: (hp50g) HPGCC and User RPL Just to add a second concurring opinion. > HPGCC has an extensive interface to the HP 50's stack. > Please refer to the sat_* functions in the documentation. Like Ingo said, most things you'll want done are already implemented. To take a set of three real numbers from the stack, you'd use something like: double stack1 = sat_pop_real(), stack2 = sat_pop_real(), stack3 = sat_pop_real(); And to push a result: sat_push_real(result); //returns real to stack > Should I mix C compiled functions in a User RPL program? While this is technically well possible, it's probably not advisable. > Simply stay with a pure C program, unless you have good reasons for a mix. An example of a good reason would be that you have a really complicated set of inputs and want to use a custom input form to get data. Making one in C would not have the same flexibility, and be much larger. Then your interface would be a few hundred bytes, and once calculations start it would be essentially instantaneous. TW === Subject: Re: new ROM? > > At ftp://ftp-fourier.ujf-grenoble.fr/ (with anonymous login) > both my browser and another FTP client show linux as a directory (empty), > and all the other entries (including pub) as files (not accessible). > The ftp server had a disk crash, should work again now. I have added a HTTP mirror for the latest 49/50 ROM === Subject: Re: new ROM? > Ê Ê> Atftp://ftp-fourier.ujf-grenoble.fr/(with anonymous login) > Ê> both my browser and another FTP client show linux as a directory > (empty), > Ê> and all the other entries (including pub) as files (not accessible). > Ê > The ftp server had a disk crash, should work again now. I have added a > HTTP mirror for the latest 49/50 ROM of the CAS recently (in you spare time), or have other projects grabbed your attention? Also, Is it just me or is the source code available in CASrelease out of date? (The more recent fixes not being included). Oh, yes, a sort-of bug: the lack of help text for Geo causes a problem with Emac's autocomplete. It seems to think there is CAS help available, but pushing the button simply causes to calc to hang or reset. === Subject: Re: new ROM? [CASrelease] > is the source code available in CASrelease out of date? > (The more recent fixes not being included). The object files (all the rest of the ROM, with which you must link any compiled CAS) are way out of date, too, so any ROM which you make yourself from that specific kit, even if you have completely new CAS source, is as of a very old ROM in all other parts. It's therefore a bit difficult for anyone to make a personal, up-to-date, complete ROM at home, except for those who possess the entire source and the entire system which produces ROMs. One might be able to break apart a more recent ROM to get the other object (vger) parts, but you would also need to have the updated complete list of both supported (unchanging) and non-supported ROM addresses, or else a newly linked CAS part might not work with it. Come to think of it, casrelease is as of ROM 1.24 for the original HP49G -- I think you can not even create a ROM for the 49G+/50G series from it, although it does also output a ROM suitable for the emulator (emu48/49). Another sort of Catch-22 -- you can have it, but you can't do much with it: http://en.wikipedia.org/wiki/Catch-22 http://en.wikipedia.org/wiki/Catch-22_(logic) http://www.imdb.com/title/tt0065528/ [film] http://www.amazon.com/Catch-22-Joseph-Heller/dp/0684833395 http://www.bookrags.com/notes/c22/ http://www.nytimes.com/books/98/02/15/home/heller.html === Subject: Re: new ROM? [CASrelease] >>is the source code available in CASrelease out of date? >>(The more recent fixes not being included). > The object files (all the rest of the ROM, with which > you must link any compiled CAS) are way out of date, too, > so any ROM which you make yourself from that specific kit, > even if you have completely new CAS source, > is as of a very old ROM in all other parts. It's therefore a bit difficult for anyone > to make a personal, up-to-date, complete ROM at home, > except for those who possess the entire source > and the entire system which produces ROMs. > That's correct. I do not have the right to update object files for the casrelease myself. And that would be only helpfull for the 49G anyway (patient people may wait until the copyright expires, then we will be able to publish the whole source code...). The main purpose of the release of the CAS under the GPL was for me to be sure not to have any problems with giac/xcas (I don't think I would have had problems since the languages are different, but it's better to be sure). === Subject: Re: new ROM? Ê Ê> Atftp://ftp-fourier.ujf-grenoble.fr/(withanonymous login) > Ê> both my browser and another FTP client show linux as a directory > (empty), > Ê> and all the other entries (including pub) as files (not accessible). > Ê > The ftp server had a disk crash, should work again now. I have added a > HTTP mirror for the latest 49/50 ROM of the CAS recently (in you spare time), or have other projects > grabbed your attention? > Also, Is it just me or is the source code available in CASrelease out > of date? (The more recent fixes not being included). Oh, yes, a sort-of bug: the lack of help text for Geo causes a problem > with Emac's autocomplete. It seems to think there is CAS help > available, but pushing the button simply causes to calc to hang or > reset. Correction, it seems that Emac's tab completion attempting to open any of the CAS help windows causes a crash under 2.10-7. I'm not much of a fan of that. Any idea's what the problem is? CD, are you still around? Can you fix this? === Subject: Re: new ROM? (was: In 50g, how to compute Fourier transform??) Cc: parisse@fourier.ujf-grenoble.fr posting-account=82unwgoAAABKn0foN-teZr4DLYjqF_TL Gecko/2008030714 Firefox/3.0b4,gzip(gfe),gzip(gfe) > Speaking of Fourier transform: ftp> open ftp-fourier.ujf-grenoble.fr > Connected to malherbe.ujf-grenoble.fr. > 220 (vsFTPd 2.0.5) > User (malherbe.ujf-grenoble.fr:(none)): anonymous > 331 Please specify the password. > Password: > 230 Login successful. > ftp> dir > 200 PORT command successful. Consider using PASV. > 150 Here comes the directory listing. > lrwxrwxrwx Ê Ê1 0 Ê Ê Ê Ê65534 Ê Ê Ê Ê Ê 5 Feb 05 12:19 debian -> linux > drwxr-sr-x Ê Ê2 0 Ê Ê Ê Ê50 Ê Ê Ê Ê Ê 4096 Feb 05 12:20 linux > lrwxrwxrwx Ê Ê1 0 Ê Ê Ê Ê65534 Ê Ê Ê Ê Ê 9 Feb 05 12:20 pub -> linux/pub > lrwxrwxrwx Ê Ê1 0 Ê Ê Ê Ê65534 Ê Ê Ê Ê Ê10 Feb 05 12:20 xcas -> linux/xcas > 226 Directory send OK. > ftp: 286 bytes received in 0.02Seconds 17.88Kbytes/sec. > ftp> cd linux > 250 Directory successfully changed. > ftp> dir > 200 PORT command successful. Consider using PASV. > 150 Here comes the directory listing. > 226 Directory send OK. > ftp> cd .. > 250 Directory successfully changed. > ftp> dir > 200 PORT command successful. Consider using PASV. > 150 Here comes the directory listing. > lrwxrwxrwx Ê Ê1 0 Ê Ê Ê Ê65534 Ê Ê Ê Ê Ê 5 Feb 05 12:19 debian -> linux > drwxr-sr-x Ê Ê2 0 Ê Ê Ê Ê50 Ê Ê Ê Ê Ê 4096 Feb 05 12:20 linux > lrwxrwxrwx Ê Ê1 0 Ê Ê Ê Ê65534 Ê Ê Ê Ê Ê 9 Feb 05 12:20 pub -> linux/pub > lrwxrwxrwx Ê Ê1 0 Ê Ê Ê Ê65534 Ê Ê Ê Ê Ê10 Feb 05 12:20 xcas -> linux/xcas > 226 Directory send OK. > ftp: 286 bytes received in 0.02Seconds 17.88Kbytes/sec. > ftp> cd xcas > 550 Failed to change directory. > ftp> cd pub > 550 Failed to change directory. > ftp> bye . That matches my experience exactly. Firefox 3 shows 1 directory and 3 symbolic links. I also did that exact same test with the command line FTP client. Either all the files somehow got removed from that ftp server, except the top level folder and symbolic links, or somebody over there really messed the server configuration. Unfortunately, I have not found any mirrors containing the files, nor do those files look to be available via http. BP had a few files stored on the http server, but that was not one of them. Since Bernard Parisse, would not only know these things better than anybody else, but might also be interested to know his files are not available, but has not yet joined the conversation, I am CCing him. To Bernard Parisse: apparently your files stored on ftp-fourier.ujf- grenoble.fr are no longer accessable. This is a shame, as your 2.10-7 rom is the current latest and greatest, but is no longer available with the server offline. username@localhost === Subject: Re: What HP calculators you currently have? >> Veli-Pekka has: >> HP-75C >> HP-71B * 2 >> HP 50G >> HP 49g+ >> HP 49G * 2 (one is overclocked to 5MHz) >> HP 48GII >> HP-48GX >> HP-48SX >> HP-28S * 2 >> HP-41CX * 3 (all somewhat broken) >> HP-42S >> HP 33S >> HP-32S >> HP-18C >> HP 17BII+ >> HP-16C >> Perhaps a few others, no Xpander :-( >> I could sell any of those to raise money because I'm between jobs >> Later I will regret (like selling my HP-15C) >> but that's another story >> AND >> I'm definately not a standard nerd...I just like gadgets I renew my sales effort, I still need the cash (4 my mc) Veli-Pekka Nousiainen from Finland === Subject: Re: What HP calculators you currently have? > Veli-Pekka has: > HP-75C > HP-71B * 2 > HP 50G > HP 49g+ > HP 49G * 2 (one is overclocked to 5MHz) > HP 48GII > HP-48GX > HP-48SX > HP-28S * 2 > HP-41CX * 3 (all somewhat broken) > HP-42S > HP 33S > HP-32S > HP-18C > HP 17BII+ > HP-16C Perhaps a few others, no Xpander :-( I could sell any of those to raise money because I'm between jobs > Later I will regret (like selling my HP-15C) > but that's another story > AND > I'm definately not a standard nerd...I just like gadgets I renew my sales effort, I still need the cash (4 my mc) Veli-Pekka Nousiainen from Finland > Welcome back!!! === Subject: Re: What HP calculators you currently have? posting-account=QSfVcQoAAADkgn5Et4hWcvOa13b-lJy1 Gecko/20080201 Firefox/2.0.0.12,gzip(gfe),gzip(gfe) On 23 mar, 09:05, Veli-Pekka Nousiainen >> HP-41CX * 3 (all somewhat broken) >> HP-42S >> I'm definately not a standard nerd...I just like gadgets I renew my sales effort, I still need the cash (4 my mc) Veli-Pekka Nousiainen from Finland Nice to have you again in the group ! I am interested in the HP42S ! Please instruct me how to aquire it. Daniel. p.s. What is a mc ? Master Card? === Subject: Re: What HP calculators you currently have? >> Veli-Pekka has: >> HP-75C >> HP-71B * 2 >> HP 50G >> HP 49g+ >> HP 49G * 2 (one is overclocked to 5MHz) >> HP 48GII >> HP-48GX >> HP-48SX >> HP-28S * 2 >> HP-41CX * 3 (all somewhat broken) >> HP-42S >> HP 33S >> HP-32S >> HP-18C >> HP 17BII+ >> HP-16C > Perhaps a few others, no Xpander :-( > I could sell any of those to raise money because I'm between jobs >> Later I will regret (like selling my HP-15C) >> but that's another story >> AND >> I'm definately not a standard nerd...I just like gadgets I renew my sales effort, I still need the cash (4 my mc) Veli-Pekka Nousiainen from Finland Where were you all that time, VPN? Please explain! Great to see you back ;) === Subject: Re: What HP calculators you currently have? > Welcome back, VPN! And JKH! (last heard from in August 2007) What brings both brothers back at the same time? And on Easter sunday, too :) === Subject: Re: What HP calculators you currently have? > And JKH! (last heard from in August 2007) > What brings both brothers back at the same time? > And on Easter sunday, too :) last half year in a D.C. hospital recuperating from a badly botched surgery in November 2006 here in California. While I was in D.C., they discovered multifocal high-grade in situ carcinoma in my bladder. Luckily they caught it early enough to eradicate it (apparently). Good folks out there at Washington Hospital Center! -Joe- === Subject: Re: Why new calculators will be not as good as old ones?... A.L. schrieb im Newsbeitrag >> X > But the COS() Bug was known from the 33s ... > Even hp/Kinpo should have been aware of it, > at least if Cyrille forwarded the bug. > Maybe someone at 'hp' remembered the extremely embarrassing > first years of the 49g, and the useless first batch of the 48gII, > and pulled the emergency brake... Maybe someone recognized that some severe bug(s) > on the anniversary model would not shed a good light onto it. > And I'm not necessarily talking about the COS() bug, > because there are others, like the wrong index offset... The functions and features are less than those of an > HP-41 (12-24K ROM) or HP-42S (64K ROM) . > So it should be possible to debug the beast > to a certain level before release. Raymond > COS() bug? I've been away 2 long... :-( === Subject: Re: HP 50g: Help with quadratic regression > I have just gone against the will of allof my teachers and bought a >> new HP 50g calculator. I greatly enjoy rpn. However to my horror, the >> calc does not appear to be capable of preforming quadratic regression. >> Now my teacher is saying I told you so and is making me use one of >> her ti 83's. I can't stand it after having gotten used to rpn. Does >> anyone know of a way to make it do quadratic regression? If you mean fitting an equation of the form > >y = a*x^2 + b*x + c to a set of data points > >[ x1, y1], [x2, y2], ..., [xn, yn] where n >= 3, by least squares, it is not difficult. > Yep. They should be selling Turing machine. Everything can be done. A.L. === Subject: Re: The HP 35S has arrived... X > You can also see what the demand is like on ebay!! The last auction for > a 42S I followed went for over $400!!! In my again never humble opinion > it HP released an updated version of the 42S like they did with the 32SII > there would take six monthes to produce enough machines to cover the > backlog in orders. I'll sell my 42S at 400 euros (+shipping) right away!!! === Subject: Program to parse HEX into data words posting-account=6obVmQoAAADodZS4NHJIHe-diqGjQuBD AppleWebKit/523.15.1 (KHTML, like Gecko) Version/3.0.4 Safari/523.15,gzip(gfe),gzip(gfe) I have a 35s and 48g. Can they--or any currently available HP--parse data words? What I want to do is write a simple program that will take a transfer off of a 64-bit bus in HEX, convert it to binary, slice it into 18-bit words (don't ask, not my decision), then convert it back to binary and place it on the stack. So, for a 64-bit bus it would be three 18-bit words and one partial word all on the stack in X,Y,Z, and === Subject: Re: Program to parse HEX into data words posting-account=tK2cUgkAAABPl7vxXuLS8-X63eEhN1C3 1.1.4322; InfoPath.1; .NET CLR 2.0.50727; .NET CLR 3.0.04506.30; .NET CLR 3.0.04506.648),gzip(gfe),gzip(gfe) > I have a 35s and 48g. Can they--or any currently available HP--parse > data words? What I want to do is write a simple program that will take > a transfer off of a 64-bit bus in HEX, convert it to binary, slice it > into 18-bit words (don't ask, not my decision), then convert it back > to binary and place it on the stack. So, for a 64-bit bus it would be > three 18-bit words and one partial word all on the stack in X,Y,Z, and Both machines could do that, but my guess is that it'd be a lot faster and easier to do it on a 48g. If you could give us an example input and the expected outputs, we could give you several workable solutions. -Joe- === Subject: Re: Program to parse HEX into data words posting-account=6obVmQoAAADodZS4NHJIHe-diqGjQuBD AppleWebKit/525.13 (KHTML, like Gecko) Version/3.1 Safari/525.13,gzip(gfe),gzip(gfe) > I have a 35s and 48g. Can they--or any currently available HP--parse > data words? What I want to do is write a simple program that will take > a transfer off of a 64-bit bus in HEX, convert it to binary, slice it > into 18-bit words (don't ask, not my decision), then convert it back > to binary and place it on the stack. So, for a 64-bit bus it would be > three 18-bit words and one partial word all on the stack in X,Y,Z, and Both machines could do that, but my guess is that it'd be a lot faster > and easier to do it on a 48g. ÊIf you could give us an example input > and the expected outputs, we could give you several workable > solutions. -Joe- 0xB5D94F2C4E65C8E2 That would be equal to the BIN number split into 18-bit words... 1011010111 011001010011110010 110001001110011001 011100100011100010 Broken into words on the stack would be (least significant word in X register)... 1011010111 (T) (this is the chunk left over) 011001010011110010 (Z) 110001001110011001 (Y) 011100100011100010 (X) And then back to HEX... 2D7 (T) 194F2 (Z) 31399 (Y) 1C8E2 (X) I hope that was clear enough. I apologize if messed up somewhere along the way in that example. It's a chore doing one parse by hand let alone thousands over the course of troubleshooting a chip design. === Subject: Re: Program to parse HEX into data words > 0xB5D94F2C4E65C8E2 That would be equal to the BIN number split into 18-bit words... 1011010111 011001010011110010 110001001110011001 011100100011100010 Broken into words on the stack would be (least significant word in X > register)... 1011010111 (T) (this is the chunk left over) > 011001010011110010 (Z) > 110001001110011001 (Y) > 011100100011100010 (X) And then back to HEX... 2D7 (T) > 194F2 (Z) > 31399 (Y) > 1C8E2 (X) I hope that was clear enough. I apologize if messed up somewhere along > the way in that example. It's a chore doing one parse by hand let > alone thousands over the course of troubleshooting a chip design. Here a 48 program that I think will do what you want. << -> x << @ most sig chunk (10 bits) @ shift right 54 bits x 1. 6. START SRB SR NEXT @ 2nd most sig chunk (18 bits) @ shift right 36 bits x 1. 4. START SRB SR NEXT @ keep 18 bits # 3FFFFh AND @ 3rd most sig chunk (18 bits) @ shift right 18 bits x SRB SR SRB SR @ keep 18 bits # 3FFFFh AND @ least sig chunk (18 bits) @ keep 18 bits x # 3FFFFh AND >> Or if you prefer a smaller stack based program << @ chunks @ 4321 DUP @ 4321 4321 RLB RL RL @ 4321 3214 DUP @ 4321 3214 3214 #3FFh AND @ 4321 3214 4 SWAP @ 4321 4 3214 RLB RLB RL RL @ 4321 4 2143 #3FFFFh AND @ 4321 4 3 ROT @ 4 3 4321 DUP @ 4 3 4321 4321 SRB SRB SR SR @ 4 3 4321 432 #3FFFFh AND @ 4 3 4321 2 SWAP @ 4 3 2 4321 #3FFFFh AND @ 4 3 2 1 >> Both programs assume a word size of 64. -wes === Subject: Re: Program to parse HEX into data words posting-account=aJhvMAoAAABqICGF80eSUn6d3D9SANQE Gecko/20080201 Firefox/2.0.0.12,gzip(gfe),gzip(gfe) or taking the best ideas from both programs, weighing in at 83 bytes on the 48 emulator: %%HP: T(3)A(D)F(.); << RLB RL RL # 3FFh 1. 3. START OVER AND SWAP RLB RLB RL RL # 3FFFFh NEXT AND >> -wes === Subject: Re: Program to parse HEX into data words > or taking the best ideas from both programs, weighing in at 83 bytes > on the 48 emulator: %%HP: T(3)A(D)F(.); > << > RLB RL RL # 3FFh > 1. 3. START > OVER AND SWAP > RLB RLB RL RL > # 3FFFFh > NEXT > AND > > > -wes I can somewhat decipher most of the lines, but the first is really throwing me. Could someone explain %%HP: T(3)A(D)F(.);? Does that line === Subject: Re: Program to parse HEX into data words posting-account=aJhvMAoAAABqICGF80eSUn6d3D9SANQE Gecko/20080201 Firefox/2.0.0.12,gzip(gfe),gzip(gfe) > %%HP: T(3)A(D)F(.); I can somewhat decipher most of the lines, but the first is really > throwing me. Could someone explain %%HP: T(3)A(D)F(.);? Does that line Sorry about that. If you're keying in the program by hand, just ignore that line. It's a header for when you are transferring the program to the calculator. On my 50g, I can still transfer without this header, but then it asks me if the file is text or binary. I confused the issue even more by using a 50g header, which I'm guessing is probably different from the 48 header. Just leave it out. Here's a line by line description of what the program is doing. << @ starts off with binary in the form: 4321 @ most significant chunk has only 10 bits RLB RL RL @ 3214 (rotates 8+1+1=10 bits) # 3FFh @ 3214 mask 1 3 START @ repeat loop 3 times OVER @ 3214 mask 3214 AND @ 3214 4 SWAP @ 4 3214 RLB RLB RL RL @ 4 2143 (rotates 8+8+1+1=18 bits) # 3FFFFh @ 4 2143 mask (ready to repeat loop) NEXT @ 4 3 2 4321 mask (after all 3 loops) AND @ 4 3 2 1 >> -wes === Subject: Re: File transfer headers Re: %%HP: T(3)A(D)F(.); > 50g header, which I'm guessing > is probably different from the 48 header. Nothing different about ascii mode (program as text) transfer header, since program text language is the same across entire 48/49/50 series. T(3) Translation mode 3 (<< and >> as program delimiters, etc.) A(D) Angles in degrees (affects only vectors expressed as polar coordinates) F(.) Fraction mark (a/k/a the decimal point display character in real numbers) For F(.) the argument separator in algebraic/complex expressions is , For F(,) the argument separator in algebraic/complex expressions is ; The [.] and [,] keystrokes automatically generate the correct pair of characters, according to the Fraction mark flag (-51) setting. Exact vs Approximate mode for HP49/50 series was never included. Binary object headers are HPHP48-x (48S[X]/48G[X][+]) vs. HPHP49-x (49/50/etc. series) indicating non-compatibility of binary objects between series. === Subject: Re: File transfer headers > Re: %%HP: T(3)A(D)F(.); 50g header, which I'm guessing > is probably different from the 48 header. Nothing different about ascii mode (program as text) transfer header, > since program text language is the same across entire 48/49/50 series. T(3) Translation mode 3 (<< and >> as program delimiters, etc.) > A(D) Angles in degrees (affects only vectors expressed as polar coordinates) > F(.) Fraction mark (a/k/a the decimal point display character in real numbers) something much more complex. I guess I should stop guessing in public. > Exact vs Approximate mode for HP49/50 series was never included. That, and the fact that different models have different commands available (as per Bill Markwick's list). Two previous attempts at the small program I posted were rejected when I tried it on 48 emulator because I had used NIP and UNROT. By the way, back on May 17, 2001, John had a detailed post explaining this header issue even further. Worth the read. -wes === Subject: Re: File transfer headers and 49G/50G new commands for HP48 > different models have different commands available... > Two previous attempts at the small program I posted were rejected > when I tried it on 48 emulator because I had used NIP and UNROT. NIP and UNROT will compile to global names, on HP48, so that if you have small programs named 'NIP' and 'UNROT' in your HOME directory, doing what those commands would do in the 49/50 series, then programs using those newer UserRPL commands will work on the HP48 the same as they would on later models. A set of small programs which handle various common new 49/50 series commands is posted here: UserRPL commands from 49/50 for 48[S/G] Commands included, with UserRPL versions for all, and alternate SysRPL versions for the six stack commands: NIP, DUPDUP, UNROT, PICK3, UNPICK, NDUPN, I->R, R->I, AXM, AXL Once these are installed in an HP48, many more HP49/50 series UserRPL programs in text form will then function identically on the HP48, without modification. === Subject: Re: Program to parse HEX into data words posting-account=6obVmQoAAADodZS4NHJIHe-diqGjQuBD AppleWebKit/525.13 (KHTML, like Gecko) Version/3.1 Safari/525.13,gzip(gfe),gzip(gfe) > 0xB5D94F2C4E65C8E2 That would be equal to the BIN number split into 18-bit words... 1011010111 011001010011110010 110001001110011001 011100100011100010 Broken into words on the stack would be (least significant word in X > register)... 1011010111 (T) (this is the chunk left over) > 011001010011110010 (Z) > 110001001110011001 (Y) > 011100100011100010 (X) And then back to HEX... 2D7 (T) > 194F2 (Z) > 31399 (Y) > 1C8E2 (X) I hope that was clear enough. I apologize if messed up somewhere along > the way in that example. It's a chore doing one parse by hand let > alone thousands over the course of troubleshooting a chip design. Here a 48 program that I think will do what you want. << > Ê -> x > Ê << > Ê @ most sig chunk (10 bits) > Ê @ shift right 54 bits > Ê x 1. 6. START SRB SR NEXT Ê @ 2nd most sig chunk (18 bits) > Ê @ shift right 36 bits > Ê x 1. 4. START SRB SR NEXT > Ê @ keep 18 bits > Ê # 3FFFFh AND Ê @ 3rd most sig chunk (18 bits) > Ê @ shift right 18 bits > Ê x SRB SR SRB SR > Ê @ keep 18 bits > Ê # 3FFFFh AND Ê @ least sig chunk (18 bits) > Ê @ keep 18 bits > Ê x > Ê # 3FFFFh AND > Ê > > Or if you prefer a smaller stack based program << > Ê Ê Ê Ê Ê Ê Ê Ê Ê Ê Ê @ chunks > Ê Ê Ê Ê Ê Ê Ê Ê Ê Ê Ê @ 4321 > Ê DUP Ê Ê Ê Ê Ê Ê Ê Ê @ 4321 4321 > Ê RLB RL RL Ê Ê Ê Ê Ê @ 4321 3214 > Ê DUP Ê Ê Ê Ê Ê Ê Ê Ê @ 4321 3214 3214 > Ê #3FFh AND Ê Ê Ê Ê Ê @ 4321 3214 4 > Ê SWAP Ê Ê Ê Ê Ê Ê Ê Ê@ 4321 4 3214 > Ê RLB RLB RL RL Ê Ê Ê @ 4321 4 2143 > Ê #3FFFFh AND Ê Ê Ê Ê @ 4321 4 3 > Ê ROT Ê Ê Ê Ê Ê Ê Ê Ê @ 4 3 4321 > Ê DUP Ê Ê Ê Ê Ê Ê Ê Ê @ 4 3 4321 4321 > Ê SRB SRB SR SR Ê Ê Ê @ 4 3 4321 432 > Ê #3FFFFh AND Ê Ê Ê Ê @ 4 3 4321 2 > Ê SWAP Ê Ê Ê Ê Ê Ê Ê Ê@ 4 3 2 4321 > Ê #3FFFFh AND Ê Ê Ê Ê @ 4 3 2 1 Both programs assume a word size of 64. > -wes You all have been a great help. I really appreciate it. I believe I've underestimated the complexity and usefulness of programming HP === Subject: Re: Program to parse HEX into data words posting-account=zYTuBQoAAAC_bXzGjGVT5rxv8bOnpefP Gecko/20080201 Firefox/2.0.0.12,gzip(gfe),gzip(gfe) 0xB5D94F2C4E65C8E2 That would be equal to the BIN number split into 18-bit words... 1011010111 011001010011110010 110001001110011001 011100100011100010 Broken into words on the stack would be (least significant word in X > register)... 1011010111 (T) (this is the chunk left over) > 011001010011110010 (Z) > 110001001110011001 (Y) > 011100100011100010 (X) And then back to HEX... 2D7 (T) > 194F2 (Z) > 31399 (Y) > 1C8E2 (X) I hope that was clear enough. I apologize if messed up somewhere along > the way in that example. It's a chore doing one parse by hand let > alone thousands over the course of troubleshooting a chip design. Here is a way that could be ported to the 35s, but the 35s cannot take very long inputs: %%HP: T(3)A(R)F(.); << 0 -> n e << WHILE n B->R REPEAT n #3FFFFh AND n #40000h / 'n' STO 'e' INCR DROP END e ->LIST REVLIST OBJ-> DROP >> === Subject: Re: Program to parse HEX into data words posting-account=zYTuBQoAAAC_bXzGjGVT5rxv8bOnpefP Gecko/20080201 Firefox/2.0.0.12,gzip(gfe),gzip(gfe) Here is a better version: %%HP: T(3)A(R)F(.); << 0 -> n e << WHILE n B->R REPEAT n #3FFFFh AND n #40000h / 'n' STO 'e' INCR ROLLD END >> === Subject: Re: Overcome the STO-bug X Now, having the above said in a program, this will circumvent the infamous STO-bug. But writing this program will be a little bit tricky, especially for running in a port > 0 because the debugger of debug4x only works with port 0. X Avenard once helped me to create FinLib49 which translates error messages into Finnish it should always be at port 0 Veli-Pekka PS: Sorry for late posts, but be patient I'm catching up AND rejecting about 99% of the interesting stuff... === Subject: Re: Overcome the STO-bug posting-account=57eBeAkAAABkNMhsZnDBzGQsstJRmvKD Gecko/20080201 Firefox/2.0.0.12,gzip(gfe),gzip(gfe) > Avenard once helped me to create FinLib49 > which translates error messages into Finnish > it should always be at port 0 And I told you at that time that your FinLib49 is affected by the infamous STO-bug as well. You can easily check it again by storing something in a port. After that all alternative messages above lib6 are zeroed out. The STO-bug is still present in the latest ROM and it does not look like as it would be fixed at any time. Also my language packs work in any port. Replacing all messages needs port2. Now, feel free to translate your *whole* 49 to Finnish. After that we can use my engine with your translation for a *complete* Finnish talking 49G. Andreas http://www.software49g.gmxhome.de === Subject: Re: Overcome the STO-bug Avenard once helped me to create FinLib49 > which translates error messages into Finnish > it should always be at port 0 > And I told you at that time that your FinLib49 is affected by the > infamous STO-bug as well. You can easily check it again by storing > something in a port. After that all alternative messages above lib6 > are zeroed out. > The STO-bug is still present in the latest ROM and it does not look > like as it would be fixed at any time. /&(/!&Û)=(/&=% port2. Now, feel free to translate your *whole* 49 to Finnish. After that we > can use my engine with your translation for a *complete* Finnish > talking 49G. Some help please nousiainen dot velipekka (at) gmail.... === Subject: Re: Overcome the STO-bug > nousiainen dot velipekka (at) gmail.... I dropped a mail at this address, otherwise you can reach me through the email-address of my website. === X > * There's no more room in the ROM to add anything. So whatever you > add, you need to take something else out. There would be room, if the unfinshed and half-hearted translation to french and spanish would be removed and fully localized versions for port2 were offered. Also the hardcoded strings in IfMain and the statistic choose-boxes, etc. could be moved to a message array finaly. Every mobile phone comes with a dozen languages today and HP who makes money of a product basically developed 20-30 years ago doesn«t take the effort of translating roughly 2000 message strings with the majority less than four to five words. Not even to mention the quality of there german manuals. I guess it«s called globalisation. The whole started because I wanted *my* 49G im my mother tongue. Everything would work flawless if it wouldn«t be for this infamous sto- bug! X I whish they would fix the STO-bug and put the languages into an external library like extable(2) one could then compile their own language and replace the lib with whatever language (Klingon?/Elf language from LOTR?) they please Please? Veli-Pekka BTW: There would be even more room in ROM after that... === table of a lib is stored ? posting-account=57eBeAkAAABkNMhsZnDBzGQsstJRmvKD Gecko/20080201 Firefox/2.0.0.12,gzip(gfe),gzip(gfe) > I whish they would fix the STO-bug and put the languages > into an external library like extable(2) That is just what I did. As I can not work on the ROM I programmed a work-around for the bug and all messages reside in one lib which can be stored anywhere. Andreas http://www.software49g.gmxhome.de === Subject: Re: Poll for other languages: Your HP in Spanish, French, Italian, Portuguese ... X > However, I have developed an engine for allowing translation of all > built-in messages arrays (around 2100 to 2200). The engine is written > loaded to a protected memory area at start and circumvents the STO > bug. The messages can reside in any port. Once loaded, exchanging the > message is fast, it slows down each key-press at around 0,2 > milliseconds and has gone through extensive testing and proven to be > stable. X > An example of a 49G running completely in German can be found on my > website > www.software49g.gmxhome.de. Andreas YES, PLEASE! === posting-account=IwmXhAoAAACdWE9MO2XMyVrYqeYlhQ9X CLR 1.1.4322),gzip(gfe),gzip(gfe) not find enough information to build up PCB. I did search hpcalc.org where only mention about 128kbyte card mostly or I overlook something. Does anybody direct me where can I find such information, I know there is some logic IC to switch the port selection, how? please help. Roy === >Does anybody direct me where can I find such information, I know there >is some logic IC to switch the port selection, how? please help. I never made 1Mbyte card for GX, only 128k for SX, but I think there is no banking on this card. If I remember correctly the GX has more addresslines, but only for second port. Many years ago I bought a 1MByte Card for GX, cut three adress-line from the bus and glue three switches on the PCB, so I can use now 1Mbyte card for SX. (bankswitching by hand :-) But did you realy need so big card? My card has now eight banks with 128kB. I installed realy everything on this card I found and I think three banks are still empty. I bought my 1Meg card in cafe of my university very cheap, because someone run out of money. :-D For normal live 512kb should be enought and it is easier to build because you need only one memory chip. Do you think about making the PCB yourself? It is important to made gold on the copper at the connector! Without gold-connector you can use a card for 2-3month, than you will enjoy some problem. On my homepage there is a picture of one of my 128 card: http://www.criseis.ruhr.de/hp48/hp48b.html I think the aditional adress lines are from left to right at the unused connector-pin. I store two picture of my 1MByte card at my homepage, perhaps they are useful for you, or someone who like to make a 8x128kb card for the SX. http://www.criseis.ruhr.de/sany0279.jpg http://www.criseis.ruhr.de/sany0281.jpg Olaf === Subject: Re: Does xSTO handles storing in Port 2 properly ? X > Does Port 1 have the same restriction, that objects > must be less than 128K, and can not straddle the two banks? I can't C an answer to this... An educated guess would be...yes and that only The Blue Lightning aka HP-49G (no plus here) had two banks @ Port 1 The (overlying) Calc System has it's limitations How is that handled? Like...no object is in the between-banks?? ??? === Subject: Re: ML: Parameters for ACCESSBank0 > Hi On 2008-01-26 04:33:05 +1100, Andreas M.9aller >> I have been reporting over and over that if you change the hash-tables >> and the message-tables of the build-in libs and then store something >> in a port all hash-tables and all message-tables from lib6 upwards are >> simply zeroed out and that«s what I call the infamous STO-bug. This is not a bug, you are using non supported entries to perform > non-supported function. What if they could be fixed AND release as supported entry points?! Seems like a lot of developers really want it? Maybe it is too difficult? (not that I myself have anything to STOre, my money goes like Flash) === Subject: Re: Strange splash screen on hp49g Wasn't that in one phase also GPL code? MetaKernel & Parisse CAS... _________________________________________________ > [MDG is] Maubert Development Group (from Maubert Electronic) [HP retailer in France, also sold original MK card] BTW, the HP49 Filer is entirely under copyright by MDG and is fully part of the MetaKernel operating system. HP is licensing it. Therefore, you could say that it's one of the best things of MDG (Maubert Development Group) - Jean-Yves Avenard (May 21 2002) . === Subject: Re: Upgrading the boot sector of an old HP 49g+ >> I know that this is dangerous but: Could anyone tell how to upgrade >> the boot sector of HP 49g+? The old boot loader needs paper clip >> instead of just pressing Enter. >> I understand that there is risk of killing the calculator, but... > There is no way to flash the first 16KB of the ROM via HP provided > methods. I'm also looking for an app to flash the whole thing... or maybe I'll > write one. Matt There is, but only 4 the old 49G and you need a special ROM and a special script (and luck: check batteries, etc..) === Subject: Re: Is the 50g the end of the line for advanced calculators? X > On the other hand HP50G can be considered more of a computer, less > calculator. > Limited to stay in calculator department, on the edge with PDA > just enough not to be banned from schools and importand exams. X a new PDA style calculator like Qonos (using Linux) or perhaps use Nokia N810 Internet Tablet (using Linux) or a brand new HP X-1 (using Linux) which could use Parisse Xcas/Giac http://en.wikipedia.org/wiki/Xcas http://www.hpcc.org/datafile/V21N4/giac.html http://www-fourier.ujf-grenoble.fr/~parisse/giac.html and RPL/2 http://www.systella.fr/~bertrand/rpl2/english.html http://www.systella.fr/~bertrand/rpl2/download/download.html Get more power from USB and recharge it from your PC By integrating all this into a new PDA: HP X-1 one can have a bigger screen than just a calc, Naturally PIM, spreadsheet, whatever Linux apps you need The professional model HP X-1+ has integrated a 8GB Flash so that you have full Linux source with you and still have both SDHC slots free more fiction..err..Vision! later... === Subject: Re: Is the 50g the end of the line for advanced calculators? >>And oh, please god! Do not waste the utility and precious resources of >>our devices by bolting bags on the side such as playing music (or >>movies) or adding spreadsheets or PIM functions as built-in >>capabilities. >> Too late.... TI has spreadsheet and PIM, Casio has spreadsheet... MXeval for HP http://www.hpcalc.org/details.php?id=6516 I store the spreadsheets in variables and then recall them to stack for evaluation although I usually write a small wrapper for input GUI >> A.L. IMHO, a spreadsheet on a calculator is useless. The limited screen > space of a handheld calculator allows you to see only about 4x5 cells, > depending on the size of the text. Why not use something like Excel > for that, which has hundreds of formulas to boot? If you need a spreadsheet, then use a computer, not a calculator. I > had a TI-89 Titanium (no longer) and it had the spreadsheet and > calendar. I never used either of those features because they were so > limiting. To be honest, I am probably most satisfied by the Casio fx-260 that I > bought for $9 several years ago. It is a basic one-line scientific > that does everything that I don't want to turn on a computer for. > Being a one-line display, it also forces the user to think a bit more > rather than just dumping the whole equation onto the calculator. This > is good because it keeps your brain in check. I don't know what PIM stands for (is it some messenging program?) but > it sure doesn't sound like it belongs on a calculator. PIM = Personal Information Manager http://www.hpcalc.org/hp49/apps/pims/ === Subject: Re: Few questions about HP 50g > I'm a newbie with this calculator but I noticed some strange > behaviour : 1. ILAP ain't working properly. When searching for ILAP(Kc/(S+1)^2) I > get a wrong result.If I do LAP(result) I get Kc/(S+1), and than ILAP > again the same result. Why?! X need that function for my Control System exam. I'll just avoid using it and stick with the tables. Anyway, I'm just starting to use and program calculator,and I'm also new to RPN so I might have some more questions for dummies. If 'S' is your variable name then do 'S' STOVX 1st If 'MyVar' is the LAPlace, Integrate, REVX variable then do 'Test' STOVX and try again The Header shows typically RAD XYZ DEC R= 'X' or DEG R<< HEX C~ 'Test' RAD DEG is Angle unit XYZ R<< is Coord type DEC HEX is binary base R C is Real Complex = ~ is Exact (integers) vs. Approx 'X' 'Text' is VX or Variable used in eXecution name maybe you guess the rest... === === === Subject: Re: HP49/50 Commands not in the HP48 May you should look for an excel worksheet I made http://www.hpcalc.org/details.php?id=5265 ______________________________________________________ > All of the commands (and any synonyms) are in certain libraries, > though I don't quite know how to find them there :) The menus inherited from the 48 can be a hunt, but the CAS menu is available from MENUXY, which displays 141 commands from 0 to 140. To show all of them, which is how I made the list: 0 140 MENUXY. I found that SORT wouldn't work unless I used DOLIST to make each command a string. > Now that we have all these lists, what do we do with them? > (why climb the mountain? because it's there :) We're list freaks. Having a good list is an end in itself. :-) Bill === Subject: HP 35s posting-account=7OS_CAoAAABr2aaw7eKrVJfoiJ8Rom3f AppleWebKit/523.15.1 (KHTML, like Gecko) Version/3.0.4 Safari/523.15,gzip(gfe),gzip(gfe) Hello everyone. I've used a 48sx for 18 years. I just bought a 35s just too see if I can rest my 48sx. I noticed a couple of things and was wondering if someone can tell me if there is a way to change the settings on the 35s. 1. On the 48sx when you start typing out a number the number is on the left and when you enter the number, the number is on the right. So you know that the number has been entered. It is annoying in the 35s that when you enter a number the number appears twice. I guess this tells me that the number has been entered in the stack, but is there a way to show me a zero on the first stack instead of the same number after I enter a number? When the 35s does this, you can do arithmetic on two numbers, even though I entered only one number (remember I'm coming from a 48sx perspective). 2. On the 48sx I can enter many numbers to the stack. Then I can do any operation I want. On the 35s, It seems there are only 4 stacks, then it repeats the number that was on stack 4. What the heck is going on! It feels like it is algebraic calculator! 3. On the 48sx, when you clear, it was to clear the stack. With the 35s, I have to clear, then clear all, then clear all 'yes'. Is there a faster way of doing this? I hope HP one day they will put a 4 roll screen so we can see more than two stacks at one time! Other than that, === Subject: Re: HP 35s > 3. On the 48sx, when you clear, it was to clear the stack. With the > 35s, I have to clear, then clear all, then clear all 'yes'. Is there > a faster way of doing this? [right-shift] [clear] [5] (clear the 4 register stack) WD -- === Subject: Re: HP 35s >> 3. On the 48sx, when you clear, it was to clear the stack. With the >> 35s, I have to clear, then clear all, then clear all 'yes'. Is there >> a faster way of doing this? [right-shift] [clear] [5] (clear the 4 register stack) You can also type 0 enter enter enter. While it is one more keystroke than using the clear menu, it is just as fast becase only two separate buttons are required, one of them is really big, and you don't have to remember which off-screen menu option is the one you need. More generally, I think the fixed stack requires a different way of thinking than the infinite stack of RPL - generally, one *doesn't* clear it, and it doesn't matter. Just let the stuff you don't care about anymore scroll off the top. I came to the 35s from the 48 as well, and it took a little getting used to, but now I'm just as happy that I don't have to bother clearing the stack all the time. (I do wish undo worked for all operations, though) === Subject: Re: HP 35s posting-account=7OS_CAoAAABr2aaw7eKrVJfoiJ8Rom3f AppleWebKit/523.15.1 (KHTML, like Gecko) Version/3.0.4 Safari/523.15,gzip(gfe),gzip(gfe) You are right. You have to think differently with an RPN calculator than a RPL. On an RPN calculator it seems you are forced to make sure you follow orders of operation, in an RPL you don't. You can mix it up as you go. On Mar 25, 12:37Êpm, Terran Melconian > 3. ÊOn the 48sx, when you clear, it was to clear the stack. ÊWith the >> 35s, I have to clear, then clear all, then clear all 'yes'. ÊIs there >> a faster way of doing this? Ê Ê[right-shift] [clear] [5] Ê Ê(clear the 4 register stack) You can also type 0 enter enter enter. ÊWhile it is one more keystroke > than using the clear menu, it is just as fast becase only two separate > buttons are required, one of them is really big, and you don't have to > remember which off-screen menu option is the one you need. More generally, I think the fixed stack requires a different way of > thinking than the infinite stack of RPL - generally, one *doesn't* clear > it, and it doesn't matter. ÊJust let the stuff you don't care about > anymore scroll off the top. ÊI came to the 35s from the 48 as well, and > it took a little getting used to, but now I'm just as happy that I don't > have to bother clearing the stack all the time. (I do wish undo worked > for all operations, though) === Subject: Re: HP 35s posting-account=2G4L-woAAABoWmw_YYmxjw-bOIfoIAqk Gecko/20080201 Firefox/2.0.0.12,gzip(gfe),gzip(gfe) > You are right. You have to think differently with an RPN calculator > than a RPL. On an RPN calculator it seems you are forced to make sure > you follow orders of operation, in an RPL you don't. You can mix it > up as you go. On Mar 25, 12:37 pm, Terran Melconian >> 3. On the 48sx, when you clear, it was to clear the stack. With the >> 35s, I have to clear, then clear all, then clear all 'yes'. Is there >> a faster way of doing this? [right-shift] [clear] [5] (clear the 4 register stack) You can also type 0 enter enter enter. While it is one more keystroke > than using the clear menu, it is just as fast becase only two separate > buttons are required, one of them is really big, and you don't have to > remember which off-screen menu option is the one you need. More generally, I think the fixed stack requires a different way of > thinking than the infinite stack of RPL - generally, one *doesn't* clear > it, and it doesn't matter. Just let the stuff you don't care about > anymore scroll off the top. I came to the 35s from the 48 as well, and > it took a little getting used to, but now I'm just as happy that I don't > have to bother clearing the stack all the time. (I do wish undo worked > for all operations, though) I haven't been around for a while, did they fix all the 35s bugs? eg. the decimals outside of the screen... === Subject: Re: HP 35s > I haven't been around for a while, did they fix all the 35s bugs? Even newer 35s (build date CNA8020....) have some spurious behaviour like: 2 [+/-] [sqrt] -> error: sqrt negative 2 [+/-] [enter] 0.5 [y^x] -> error: invalid y^x 2 [+/-] [i] 0 [enter] 0.5 [y^x] -> correct result. (0i1.141...) very strange. Seems that the automatic conversion from a real input type to a complex output type doesnt work. there are some more functions which are not able to handle complex inputs and/or results. (like asin(-2.5) results in error invalid data. the trick doesnt work here: asin(-2.5i0) results also in an invalid data error) If you are looking for a calculator which is able to handle smoothly complex numbers, the HP35s is not the best choice.... WD -- === Subject: Re: HP 35s <0t47ebe710i7808n1f4@news.msgid.medusa.priv.at> posting-account=6obVmQoAAADodZS4NHJIHe-diqGjQuBD AppleWebKit/525.13 (KHTML, like Gecko) Version/3.1 Safari/525.13,gzip(gfe),gzip(gfe) This is a very tiny issue, but does anyone else have a squeaky C/Off/ On button? It works fine and doesn't miss any keystrokes, but it squeaks. When I first got the calculator I thought there was an actual beep associated with turning the calculator on and off. It almost sounds like the 48g's beep. === Subject: Re: HP 35s > I haven't been around for a while, did they fix all the 35s bugs? eg. I have not found a single bug that has been fixed in my recent hp35s, and I have found several which are not fixed (I can confirm 2,3,10,12, at least as still present). I suspect there simply has not been a firmware update at all. Disappointing. === Subject: Re: HP 35s posting-account=sOAX1QkAAAC-FcySTSbz29Uk8huUtFRz CLR 1.1.4322; .NET CLR 2.0.50727),gzip(gfe),gzip(gfe) > 1. ÊOn the 48sx when you start typing out a number the number is on > the left and when you enter the number, the number is on the right. > So you know that the number has been entered. ÊIt is annoying in the > 35s that when you enter a number the number appears twice. ÊI guess > this tells me that the number has been entered in the stack, but is > there a way to show me a zero on the first stack instead of the same > number after I enter a number? When the 35s does this, you can do > arithmetic on two numbers, even though I entered only one number > (remember I'm coming from a 48sx perspective). 2. ÊOn the 48sx I can enter many numbers to the stack. ÊThen I can do > any operation I want. ÊOn the 35s, It seems there are only 4 stacks, > then it repeats the number that was on stack 4. ÊWhat the heck is > going on! ÊIt feels like it is algebraic calculator! > This is because of the fundamental difference between RPL calculators and RPN calculators. The HP's that use RPL include the 28/48/49 series while the RPN models include the Classic, Voyager, Pioneer, and the 33/35s, to name a few. RPL calculators feature an infinite stack and when a number is entered it appears on level 1 only. RPN calculators work slightly differently, as you have noticed. Hence to square a number in RPL, you need X ENTER ENTER * while using RPN you just type X ENTER *. Since the original HP35 in 1972, the stack size in RPN calculators has been limited to 4 stack levels. Even though the newest models have much more memory, the 4-level stack was probably kept because (a) it promoted program compatability, and (b) without being able to see all the stack levels at once, 4 seems like a good number such that the human mind is capable of keeping track of all the levels. Remember, on the 48SX, you can scroll through the stack (and see 4 levels at once). On the pocket scientifics, you must roll through the stack to do this. With more than 4 levels, this might get inconvenient. S.C. === Subject: Re: HP 35s This makes much more sense. It also makes sense that my 11c acts different from my 48sx. Has there been an HP calculator which have more then 2 rows and is a RPL that is not a 48/49/50? === Subject: Re: HP 35s schrieb im Newsbeitrag > [..] > Has there been an HP calculator which have > more then 2 rows and is a RPL that is not a 48/49/50? > Yes: HP-28C , HP-28S, and internally, also the 18C and 19B/II . HTH Raymond === Subject: Re: HP 35s > 1. ÊOn the 48sx when you start typing out a number the number is on > the left and when you enter the number, the number is on the right. > So you know that the number has been entered. ÊIt is annoying in the > 35s that when you enter a number the number appears twice. ÊI guess > this tells me that the number has been entered in the stack, but is > there a way to show me a zero on the first stack instead of the same > number after I enter a number? When the 35s does this, you can do > arithmetic on two numbers, even though I entered only one number > (remember I'm coming from a 48sx perspective). 2. ÊOn the 48sx I can enter many numbers to the stack. ÊThen I can do > any operation I want. ÊOn the 35s, It seems there are only 4 stacks, > then it repeats the number that was on stack 4. ÊWhat the heck is > going on! ÊIt feels like it is algebraic calculator! This is because of the fundamental difference between RPL calculators > and RPN calculators. The HP's that use RPL include the 28/48/49 series > while the RPN models include the Classic, Voyager, Pioneer, and the > 33/35s, to name a few. RPL calculators feature an infinite stack and when a number is > entered it appears on level 1 only. RPN calculators work slightly > differently, as you have noticed. Hence to square a number in RPL, you > need X ENTER ENTER * while using RPN you just type X ENTER *. Since the original HP35 in 1972, the stack size in RPN calculators has > been limited to 4 stack levels. Even though the newest models have > much more memory, the 4-level stack was probably kept because (a) it > promoted program compatability, and (b) without being able to see all > the stack levels at once, 4 seems like a good number such that the > human mind is capable of keeping track of all the levels. Remember, on > the 48SX, you can scroll through the stack (and see 4 levels at once). > On the pocket scientifics, you must roll through the stack to do this. > With more than 4 levels, this might get inconvenient. S.C. don't remember anything about it. Then I got a 48g in '95 and have been using it ever since. When I got the 35s I was confused in the same way the original poster was. I assume my 42s worked the same way and I'm just not remembering it. It is taking some time to get used to. I've gotten used to the way the stack works for the most part. However, I still can't get used to everything taking place on the left side of the screen. It's a small thing, but it just seems so odd to me. === Subject: 128K/64K banks was: FlashTools 1.0 released > <...> I suspect some form of corruption, but what do you mean with not >> enough memory: memory in a bank, or whole memory? I mean main memory. The process of repacking a bank requires copying > all non-deleted libraries in the bank to ram, then erase the flash > memory and rewrite all your libs back. If your bank is full, you would > need up to 128 kbytes of free ram to perform a repack. X This is why I think that OS could be made to use 64K banks e.g.. the same size as Flash Physical banks Thus RMA/ROM banks could be doubled and user giving a configuration choice ROM 256KB divided in 4 banks .87 64KB while using heavy bank-switching and slower speed on ROM 192KB visible (3 banks only) 3) Boost ROM to 5 banks=320KB, causing less bank-switching thus speed is increased 4) Maximum Boost ROM 6 banks=384KB but almost no bank-switching due to LRU ROM bank witching algorithm All of this will never happen because of the resources of HP calc special task force (used to be a division) that's why I suggest rather the PDA/Linux Xcas/Giac RPL/2 approach === Subject: Re: my first programming experience on the hp50g X > some simple programs that all evaluate to 4 are > << 4 > << 2 2 + > << << 1 1 + >> EVAL << 1 1 + >> EVAL + > > to actually run the program on the last line of the stack just press > the [EVAL] key > something i didn't understand yet is if i evaluate this program on > line 1 of the stack << 2 2 + >> then then number 4 replaces it. now > if i hit [left-shift] [enter] (or [ANS]) i get a three line stack that > looks like: 4 > 2 > 2 instead of 4 > 4 ANS gives you back all your previous arguments of the last performed function see you manual about command LASTARG === Subject: Re: New release of Orpie rpn calculator program > Version 1.5.0 of the Orpie full-screen rpn calculator is now available > athttp://www.eecs.umich.edu/~pelzlpj/orpie/ Orpie is a fullscreen RPN calculator for the console. Its operation is > similar > to that of modern HP calculators, but data entry has been optimized for > efficiency on a PC keyboard. Features include extensive scientific > calculator > functionality, command completion, and a visible interactive stack. -- >> Ack! What does 'For the Console mean...??? This program runs in text mode only (no graphics). > It can run in an xterm window X windows. > i was looking at the web page a little, and i noticed that it was >> utterly missing just about any information about this product, which >> is apparently... an HPxx calculator for a desk top computer... > But WHAT computer... LINUX ( et al... ) The program can be compiled on any computer with a gcc compiler > (any linux, *bsd, etc). have anyone tried it on the HPGCC compiler - for fun :-D >> Then there was a bit about compiling the source code for Mac's, but if >> this were possible...???... why hasn't someone that knew what they >> were doing, already done that...??? I have no idea, other than that the code was released on Friday. > ( gawd! i hate computers...!!! ) >> Thanx. > > -- === Subject: Re: Programming >> Finally have time to really learn my calculators. I want to learn to >> program them since one of them is always with me and my computer is >> not. I want to program for the math problems in some of the math >> problem books. I don't know weather to use sys RPL, user RPL, or >> C on the computer and transfer over to calculator. I intend to >> Don On the other hand, if you > wanted to write an efficient function to generate, say, the plasma > dispersion function from series approximations, where the basic > operations are 4-function but there are a lot of them, then it would > be a lot faster in SysRPL, I think. C a lot faster in C === Subject: Re: SysRPL a useful reference guide. X > Omigosh, that stuff's getting pretty stale. Time to add some new > goodies! where are the pictures?? Father? Mathew 23:9 === Subject: Re: SysRPL a useful reference guide. <6i4a039lmmhnudpbj1s55of6nakj3h2g76@4ax.com> <20070325123303.00c68a86@aiguebelle.pamplemousse.homeunix.org> posting-account=wcRVPwkAAACV9qEmJuiPuuML0GXuIOZ6 4.90),gzip(gfe),gzip(gfe) On Mar 25, 4:54Êpm, Veli-Pekka Nousiainen X Omigosh, that stuff's getting pretty stale. ÊTime to add some new > goodies! where are the pictures?? Father? > Mathew 23:9 Well, this is a very comprehensive list that the OP has listed: 1. rplman (hp) 2. Programming in System Rpl - Kalinowski Dominik 3. SysRPL Programming - Heiskanen 4. Syseval Entries - (hp? Dominik?) 5. Advanced programming information - Tim Wessman Could/would someone be able to post the exact locations/linkd where we (who are interested) can obtain these? TomCee === Subject: Re: SysRPL a useful reference guide. <6i4a039lmmhnudpbj1s55of6nakj3h2g76@4ax.com> <20070325123303.00c68a86@aiguebelle.pamplemousse.homeunix.org> posting-account=9rwnLQoAAAC_FoizTwvxTD73niS5ww5R Gecko/20080207 Ubuntu/7.10 (gutsy) Firefox/2.0.0.12,gzip(gfe),gzip(gfe) > On Mar 25, 4:54 pm, Veli-Pekka Nousiainen > X Omigosh, that stuff's getting pretty stale. Time to add some new > goodies! where are the pictures?? Father? > Mathew 23:9 Well, this is a very comprehensive list that the OP has listed: 1. rplman (hp) > 2. Programming in System Rpl - Kalinowski Dominik > 3. SysRPL Programming - Heiskanen > 4. Syseval Entries - (hp? Dominik?) > 5. Advanced programming information - Tim Wessman Could/would someone be able to post the exact locations/linkd where we > (who are interested) can obtain these? TomCee Hi TomCee. Here's the single links to the references: > 1. rplman (hp) http://www.hpcalc.org/hp48/docs/programming/rpl-pdf.zip RPL Programming Guide from Goodies Disk 4 in Adobe Acrobat PDF format. By Hewlett-Packard (http://www.hp.com/calculators/). > 2. Programming in System Rpl - Kalinowski Dominik http://www.hpcalc.org/hp49/docs/programming/progsysrpl_pdf.zip Second edition of the free on-line book Programming in System RPL, a 640-page tutorial and reference for System RPL programmers. This second edition has many improvements and describes new features present in the HP49G calculator. In Adobe PDF format. By Eduardo M. Kalinowski (http://move.to/hpkb) and Carsten Dominik (http://staff.science.uva.nl/~dominik/hpcalc/). http://www.hpcalc.org/hp49/docs/programming/progsysrpl_examples.zip Example programs from the second edition of the free on-line book Programming in System. By Eduardo M. Kalinowski (http://move.to/hpkb) and Carsten Dominik (http://staff.science.uva.nl/~dominik/hpcalc/). > 3. SysRPL Programming - Heiskanen ??? can't find :-( > 4. Syseval Entries - (hp? Dominik?) http://www.hpcalc.org/hp49/programming/entries/edb.zip ASCII database with more than 7500 entries points for HP48, HP49, hp49g +, hp48gII, HP38, HP39, HP40, complete with stack diagrams and descriptions. Includes a program to extract PDF and ASCII listings of selected entries. The reference sections of the book Programming in System RPL, 2nd edition are based on an earlier version of this database. Extracts for individual calculators are available from this site, in PDF format. By Carsten Dominik (http://staff.science.uva.nl/~dominik/hpcalc/), Thomas Rast, and Eduardo M. Kalinowski (http://move.to/hpkb). > 5. Advanced programming information - Tim Wessman http://www.timandkatie.com/calc/index.html (see the Docs page for more references). Hope this helps. Giancarlo === Subject: Re: SysRPL a useful reference guide. > X Omigosh, that stuff's getting pretty stale. ÊTime to add some new > goodies! where are the pictures?? Father? > Mathew 23:9 Well, this is a very comprehensive list that the OP has listed: Ê Ê Ê1. Êrplman Ê (hp) > Ê Ê Ê2. ÊProgramming in System Rpl - Kalinowski ÊDominik > Ê Ê Ê3. SysRPL Programming - Heiskanen > Ê Ê Ê4. Syseval Entries - (hp? Dominik?) > Ê Ê Ê5. ÊAdvanced programming information - Tim Wessman Could/would someone be able to post the exact locations/linkd where we > (who are interested) can obtain these? TomCee Hi TomCee. > Here's the single links to the references: Ê Ê Ê1. Êrplman Ê (hp) http://www.hpcalc.org/hp48/docs/programming/rpl-pdf.zip RPL Programming Guide from Goodies Disk 4 in Adobe Acrobat PDF format. > By Hewlett-Packard (http://www.hp.com/calculators/). Ê Ê Ê2. ÊProgramming in System Rpl - Kalinowski ÊDominik http://www.hpcalc.org/hp49/docs/programming/progsysrpl pdf.zip Second edition of the free on-line book Programming in System RPL, a > 640-page tutorial and reference for System RPL programmers. This > second edition has many improvements and describes new features > present in the HP49G calculator. In Adobe PDF format. > By Eduardo M. Kalinowski (http://move.to/hpkb) and Carsten Dominik > (http://staff.science.uva.nl/~dominik/hpcalc/). http://www.hpcalc.org/hp49/docs/programming/progsysrpl examples.zip Example programs from the second edition of the free on-line book > Programming in System. > By Eduardo M. Kalinowski (http://move.to/hpkb) and Carsten Dominik > (http://staff.science.uva.nl/~dominik/hpcalc/). Ê Ê Ê3. SysRPL Programming - Heiskanen ??? can't find :-( Ê Ê Ê4. Syseval Entries - (hp? Dominik?) http://www.hpcalc.org/hp49/programming/entries/edb.zip ASCII database with more than 7500 entries points for HP48, HP49, hp49g > +, hp48gII, HP38, HP39, HP40, complete with stack diagrams and > descriptions. Includes a program to extract PDF and ASCII listings of > selected entries. The reference sections of the book Programming in > System RPL, 2nd edition are based on an earlier version of this > database. Extracts for individual calculators are available from this > site, in PDF format. > By Carsten Dominik (http://staff.science.uva.nl/~dominik/hpcalc/), > Thomas Rast, and Eduardo M. Kalinowski (http://move.to/hpkb). Ê Ê Ê5. ÊAdvanced programming information - Tim Wessman http://www.timandkatie.com/calc/index.html(see the Docs page for > more references). Hope this helps. > Giancarlo- Hide quoted text - - Show quoted text - Giancarlo: TomCee === Subject: Re: Need a sysRPL CRC16 routine. Have you tried changing the wordsize? 4; 8; ... 60; 64 _____________________________________________ Note that user binary objects are also hex strings, but are fixed-length (16d nibbles of data): #123456789ABCDEF0h ->H ==> E4A20 51000 0FEDCBA987654321 #1234567h ->H ==> E4A20 51000 7654321000000000 === Subject: Why Mail Foundry Quarantined a comp.sys.hp48 post On down the page you will find a post starting with harvard abstinence...... from anitadv...@gmail.com whose entire content is pornographic. This message is why my ISPs MailFoundry quarantined the entire digest of comp.sys.hp48 messages. MF must peer down into a post to find 'naughty' content. Too bad we don't have a moderator. === Subject: sqrt(4) -> 2 and such I don't know what's happened but lately my HP 50g doesn't simplify expressions with square roots such as: sqrt(4) -> 2 sqrt(8) -> 2*sqrt(2) 1/sqrt(2) -> sqrt(2)/2 etc. I guess some flag may have changed accidentally but I can't find out which one it is. === Subject: Re: sqrt(4) -> 2 and such >I don't know what's happened but lately my HP 50g doesn't simplify >expressions with square roots such as: >sqrt(4) -> 2 >sqrt(8) -> 2*sqrt(2) >1/sqrt(2) -> sqrt(2)/2 >etc. >I guess some flag may have changed accidentally but I can't find out >which one it is. [RightShift]+[Enter] to toggle between exact (=) and approximate (~) mode. Damir === Subject: Re: sqrt(4) -> 2 and such <3igku31r1uum34re2cruu6hvjdvfd6f71a@4ax.com> posting-account=7XHRvQkAAAC4QU_AaK7Q-xQQ_wk7BFnr Gecko/20080311 Firefox/2.0.0.13,gzip(gfe),gzip(gfe) But sqrt(4) is exactly 2, why don't I get that result while in exact mode? What about the others, because I don't want a decimal approx. === Subject: Re: sqrt(4) -> 2 and such posting-account=Q2CEjQoAAACue5ZDDhUeRzv1w0u8hxZE Gecko/2008030714 Firefox/3.0b4,gzip(gfe),gzip(gfe) > But sqrt(4) is exactly 2, why don't I get that result while in exact > mode? What about the others, because I don't want a decimal approx. Run the command CASCFG. That resets back to defaults. TW === Subject: [35s] Bizarre program bugs when running under solver I'm getting bizarre and wrong behaviour from programs which execute as a subprocess to the solver. Here's the most trivial case I was able to come up with to reproduce the problem: B001 LBL B B002 INPUT T B003 INPUT X B004 CF 1 B005 FS? 1 B006 XEQ C001 B007 0.2xX-T B008 RTN C001 LBL C C002 STOP C003 RTN When program B is executed normally, the XEQ C is never called, as ought to be the case. When I set FN=B and solve, it calls C, even though it should not, and hits the STOP. All kinds of conditionals seem to be broken when running under the solver, not just FS?. I looked through the manual for some statement about functionality which is not available in programs being solved instead of executed, but I could not find anything. WTF is going on? Do other people have this problem, or do I have a calculator with a subtle defect that needs to be returned? Any thoughts or insight would be appreciated. === Subject: Re: [35s] Bizarre program bugs when running under solver Replying to myself: > I'm getting bizarre and wrong behaviour from programs which execute as a > subprocess to the solver. Here's the most trivial case I was able to > come up with to reproduce the problem: B001 LBL B > B002 INPUT T > B003 INPUT X > B004 CF 1 > B005 FS? 1 > B006 XEQ C001 > B007 0.2xX-T > B008 RTN > C001 LBL C > C002 STOP > C003 RTN When program B is executed normally, the XEQ C is never called, as ought > to be the case. When I set FN=B and solve, it calls C, even though it > should not, and hits the STOP. All kinds of conditionals seem to be I beat my head against this problem for several more hours after posting, and I now have a hypothesis for what's going on. The solver needs to find all the INPUT statements when it first starts up so that it can prompt the user for the variables as necessary. In order to find ALL the relevant INPUT statements, it needs to examine all the code lines which are going to execute. To do this, it employs a simple heuristic to step through the program in an initial pass: * Do not store anything to variables * Do not put anything on the stack * Evaluate all conditionals as if they were true * Prompt for any variable in an INPUT statement * Obey any XEQ, GTO or RTN statement * Obey any STOP statement (I don't know why) * End upon reaching the RTN from the original program After making one initial pass in this manner, presumably to find the INPUT statements, it then starts executing normally to actually evaluate the function. However, this initial INPUT finding heuristic is braindamaged, because it causes most loops written the usual way to become infinite loops. Whoever came up with this can't have tested it on a nontrivial program, and to make matters worse, this behavior appears to be undocumented. On the bright side, once the issue is understood, it is possible to work around it by inserting an always-false conditional followed by a return right after your input statements. This RTN will execute in the initial modified-behavior pass, and then never again. Example: B001 LBL B B002 INPUT T B003 INPUT X B004 CF 1 B005 FS? 1 B006 RTN B007 FS? 1 B008 XEQ C001 B009 0.2xX-T B010 RTN C001 LBL C C002 STOP C003 RTN I still wonder if I'm missing something, since I searched the newsgroup surprised if I were actually the first person to run into this issue, but perhaps running the solver on arbitrary programs is just not commonly-used functionality? Perhaps I just didn't find the right terms to search on. === Subject: A bug and a related mini-challenge I recently came across an unexpected bug in the HP48G (and descendants). The LSQ command is supposed to return the solution with minimum Euclidean length when given an underdetermined system (with an infinite number of solutions). Given system A*x=B, with A: [[ 1 1 0 0 0 -1 ] [ 1 0 -1 0 1 0 ] [ 0 1 1 1 0 0 ] [ 0 0 0 1 1 1 ]] and B: [[ 4 ] [ 8 ] [ 12 ] [ 16 ]] LSQ gives a solution: [[6.99541284404 ] [ 6.15137614679 ] [ 2.84403669725 ] [ 3.00458715596 ] [ 3.84862385321 ] [ 9.14678999083 ]] with a Euclidean length of 14.2255745922. But this is not the minimum length solution. For example, [ 6 2 1 9 3 4 ]T is a solution and has Euclidean length 12.124355653, but there IS a solution with minimum possible length. The mini-challenge is to find it, using your HP48G, HP49G+ or HP50G. Don't use MATLAB or something similar until you give it a shot on your HP calculator. I was very surprised by this bug. I've used LSQ a lot, and never before seen such behavior, and I have no idea what particular circumstance leads to the error. === Subject: Re: A bug and a related mini-challenge re LSQ Rodger the Phantom Rosenbaum observed: [[ 1 1 0 0 0 -1 ] [ 1 0 -1 0 1 0 ] [ 0 1 1 1 0 0 ] [ 0 0 0 1 1 1 ]] 'A' STO [[ 4 ] [ 8 ] [ 12 ] [ 16 ]] 'B' STO B A LSQ The desired theoretical answer [[ 3 ] [ 4 ] [ 1 ] [ 7 ] [ 6 ] [ 3 ]] is not obtained, although it would be if the last row of each matrix were dropped, e.g. [[ 4 ] [ 8 ] [ 12 ]] [[ 1 1 0 0 0 -1 ] [ 1 0 -1 0 1 0 ] [ 0 1 1 1 0 0 ]] LSQ Discussion: Change 16 to 16.0000000001 (in which case there exists no _exactly_ correct theoretical solution at all) and see how the result of LSQ is affected; what does this say about the given system, and about the potential difficulty of arriving at a precisely correct answer to the _theoretical_ problem which LSQ is supposed to solve, using real arithmetic? > I have no idea what particular circumstance leads to the error. If A has less than full row rank (the system is underdetermined), an infinite number of solutions exist. LSQ returns the solution with the minimum Euclidean length. But there is not an infinite number of solutions in cases like this (use 17 to make it worse); in fact, there are _no_ solutions (perhaps one can say that it is overdetermined, after all), so what is the criterion now, and has it been satisfied or not? -=-=-=- === Subject: Re: A bug and a related mini-challenge re LSQ no _exactly_ correct theoretical solution at all) >and see how the result of LSQ is affected; >what does this say about the given system, >and about the potential difficulty of arriving >at a precisely correct answer to the _theoretical_ problem >which LSQ is supposed to solve, using real arithmetic? I have no idea what particular circumstance leads to the error. If A has less than full row rank (the system is underdetermined), >an infinite number of solutions exist. >LSQ returns the solution with the minimum Euclidean length. But there is not an infinite number of solutions in cases like this >(use 17 to make it worse); in fact, there are _no_ solutions >(perhaps one can say that it is overdetermined, after all), >so what is the criterion now, and has it been satisfied or not? -=-=-=- Another way to see that there really shouldn't be numerical problems with this system and perturbations thereof is like this: Given system A*x=B, with A: [[ 1 1 0 0 0 -1 ] [ 1 0 -1 0 1 0 ] [ 0 1 1 1 0 0 ] [ 0 0 0 1 1 1 ]] and B: [[ 4 ] [ 8 ] [ 12 ] [ 16 ]] A classic solution for a linear system is x = Inverse(A)*B But if the system matrix A is not square and non-singular, this solution won't exist, since the inverse doesn't exist. In that case, replace Inverse(A) with PseudoInverse(A). The pseudoinverse of A can be found with the aid of the SVD; it is: [[ .25 .25 0 0 ] [ .25 0 .25 0 ] [ 0 -.25 .25 0 ] [ 0 0 .25 .25 ] [ 0 .25 0 .25 ] [ -.25 0 0 .25 ]] The solution of the system we've been considering is given by: x = PseudoInverse(A)*B, where B can be the original B, or a perturbed version. If B is the perturbed version: [[ 4 ] [ 8 ] [ 12 ] [ 16.0000000001 ]] then it's easy to see that there is no numerical difficulty whatsoever getting the correct minimum norm solution (no matter what B is), and such a solution is what LSQ ought to get if it's working properly. === Subject: Re: A bug and a related mini-challenge re LSQ >Rodger the Phantom Rosenbaum observed: [[ 1 1 0 0 0 -1 ] > [ 1 0 -1 0 1 0 ] > [ 0 1 1 1 0 0 ] > [ 0 0 0 1 1 1 ]] 'A' STO [[ 4 ] > [ 8 ] > [ 12 ] > [ 16 ]] 'B' STO B A LSQ The desired theoretical answer [[ 3 ] > [ 4 ] > [ 1 ] > [ 7 ] > [ 6 ] > [ 3 ]] is not obtained, although it would be >if the last row of each matrix were dropped, e.g. Dropping any row from each matrix gives that solution. [[ 4 ] > [ 8 ] > [ 12 ]] [[ 1 1 0 0 0 -1 ] > [ 1 0 -1 0 1 0 ] > [ 0 1 1 1 0 0 ]] LSQ Discussion: Change 16 to 16.0000000001 (in which case there exists >no _exactly_ correct theoretical solution at all) >and see how the result of LSQ is affected; >what does this say about the given system, >and about the potential difficulty of arriving >at a precisely correct answer to the _theoretical_ problem >which LSQ is supposed to solve, using real arithmetic? It doesn't say much. First, some terminology. The AUR uses the word solution in places where it should add an adjective in front to be precise. I'll use the terms exact solution, approximate solution, minimum norm solution and minimum residual solution. By exact solution, I mean a solution x (a vector) that gives a residual norm, NORM(A*x-b), of exactly zero. By approximate solution, I mean a solution x that gives a non-zero residual. There are almost always many approximate solutions, even if there is also an exact solution (he said tautologically). By minimum residual solution, I mean a solution which gives a minimum residual norm. Typically this term will be used when there are many approximate solutions and no exact solution, but over the set of approximate solutions, only one has the smallest residual norm; this would be the common overdetermined least squares solution. By minimum norm solution, I mean a solution vector which has the smallest Euclidean length (norm) compared to all other solutions. In some underdetermined systems. there an infinite number of exact solutions, but only one solution vector is of minimum length. In the early days of linear algebra (Gauss' time), elimination methods were used to solve linear systems. Such methods have a hard time with systems whose A matrix is singular, but such systems can often be solved approximately with modern methods, giving a relatively small residual (we hope). Sometimes systems with a singular A matrix even have an exact solution. Modern methods of solution using orthogonal decompositions are able to provide solutions for almost any system. The AUR doesn't mention it, but a third party book I have mentions that LSQ uses QR and LQ decompositions to obtain solutions. For an example system A*x=B, let A be: [[ 1 1 1 1 ] [ 2 2 2 2 ] [ 3 3 3 3 ] [ 4 4 4 4 ]] and B: [[ 1 ] [ 2 ] [ 3 ] [ 4 ]] If singular could be quantified, we would have to say that this A matrix is very singular. Put B and A on the stack of your HP50 and execute LSQ. You will see [ .25 .25 .25 .25 ]T, which is an exact solution. Many people think that if the A matrix is singular, the system has no solution. This is not necessarily so, as this example demonstrates. Now, if we go back to the system in my first post in this thread, we have an A matrix: [[ 1 1 0 0 0 -1 ] [ 1 0 -1 0 1 0 ] [ 0 1 1 1 0 0 ] [ 0 0 0 1 1 1 ]] Create another matrix, and call it A' (the first two columns are swapped): [[ 1 1 0 0 0 -1 ] [ 0 1 -1 0 1 0 ] [ 1 0 1 1 0 0 ] [ 0 0 0 1 1 1 ]] The B matrix is: [[ 4 ] [ 8 ] [ 12 ] [ 16 ]] Let B' be: [[ 4 ] [ 8 ] [ 12 ] [ 16.0000000001 ]] If we solve the system A*x=B' with LSQ, we get: [[ -2974.6559633 ] [ -1599.35321101 ] [ -1373.30275229 ] [ 2984.6559633 ] [ 1609.35321101 ] [ -4578.00917431 ]] Apparently you got this result and then asked: ...what does this say about the given system, and about the potential difficulty of arriving at a precisely correct answer to the _theoretical_ problem which LSQ is supposed to solve, using real arithmetic? believing that the large difference in the solution compared to what it was before the last element of the b matrix was perturbed to 16.0000000001 indicates difficulty in solving the problem. But now solve the system A'*x=B' with LSQ and get: [[ 4 ] [ 3 ] [ 1 ] [ 7.00000000002 ] [ 6.00000000002 ] [ 3.00000000002 ]] which has a residual of A'*x-B' =: [[ .00000000002 ] [ .00000000002 ] [ 0 ] [ 0 ]] When you're using modern orthogonal decomposition methods for solving the system, a small perturbation which moves the system away from the case of an infinite number of exact solutions doesn't make it any harder to solve. The perturbed system has an infinite number of approximate solutions, one of which is the minimum norm approximate solution. If the perturbation is small, the minimum norm approximate solution will be nearly the same as the minimum norm exact solution to the unperturbed system. It is LSQ's job to find approximate solutions to systems like this. The reason you got that very different LSQ solution to the A*x=B' system is that LSQ is not working right with the original A matrix. The A' matrix should give the same LSQ solution, and in fact it behaves properly. LSQ shouldn't have the same kind of problems with ill conditioned systems that elimination methods have; that's why one should use modern orthogonal decomposition methods. Mathematica 6, which just came out, has a new function, LeastSquares{}. If I give it the A*x=B' system, it gives as a (minimum norm) solution: [[ 4 ] [ 3 ] [ 1 ] [ 7.000000000025 ] [ 6.000000000025 ] [ 3.000000000025 ]] I used exact numbers and the last three elements of the solution vector are to be considered exact though expressed as decimals. Compare to what LSQ got for the solution to A'*x=B'. As the original system is perturbed, the minimum norm approximate solution should gradually move away from the minimum norm exact unperturbed solution. There is no sudden, large change, and a modern orthogonal decomposition method (which LSQ is) should not suddenly have a problem solving the problem just because of a slight perturbation (assuming that the system is not so ill-conditioned that it was having trouble before the perturbation.). LSQ has a bug which shows with the original A matrix, but not the A' matrix. Try this experiment. Perturb the last element of the B matrix more and more; let it become 16.000000001, 16.00000001, 16.0000001, 16.000001, etc. Repeatedly solve the system A'*x=B' with LSQ, where B' is incrementally perturbed. Look at what the solutions are doing. I have no idea what particular circumstance leads to the error. If A has less than full row rank (the system is underdetermined), >an infinite number of solutions exist. >LSQ returns the solution with the minimum Euclidean length. But there is not an infinite number of solutions in cases like this >(use 17 to make it worse); in fact, there are _no_ solutions Certainly there are no exact solutions, but there are approximate solutions, and there is only one minimum norm approximate solution. And that is what LSQ is supposed to find. >(perhaps one can say that it is overdetermined, after all), >so what is the criterion now, and has it been satisfied or not? It hasn't been satisfied by the LSQ solution to the A*x=B' system, but it is satisfied by the LSQ solution to the A'*x=B' system. There is no reason for LSQ to have trouble with this system. You can also get the same solution as the LSQ solution to the A'*x=B' system by using the SVD with no hint of numerical problems. And this is the result that Mathematica 6 gives to both the A*x=B' system and the A'*x=B' system, as I would expect. LSQ has a bug. -=-=-=- === Subject: Re: A bug and a related mini-challenge > I recently came across an unexpected bug in the HP48G (and descendants). The LSQ command is supposed to return the solution with minimum Euclidean > length when given an underdetermined system (with an infinite number of > solutions). Given system A*x=B, with A: [[ 1 1 0 0 0 -1 ] > [ 1 0 -1 0 1 0 ] > [ 0 1 1 1 0 0 ] > [ 0 0 0 1 1 1 ]] and B: [[ 4 ] > [ 8 ] > [ 12 ] > [ 16 ]] LSQ gives a solution: [[6.99541284404 ] > [ 6.15137614679 ] > [ 2.84403669725 ] > [ 3.00458715596 ] > [ 3.84862385321 ] > [ 9.14678999083 ]] with a Euclidean length of 14.2255745922. But this is not the minimum > length solution. For example, [ 6 2 1 9 3 4 ]T is a solution and has > Euclidean length 12.124355653, but there IS a solution with minimum > possible length. The mini-challenge is to find it, using your HP48G, HP49G+ > or HP50G. Don't use MATLAB or something similar until you give it a shot on > your HP calculator. I was very surprised by this bug. I've used LSQ a lot, and never before > seen such behavior, and I have no idea what particular circumstance leads > to the error. This is not necessarily a bug. Given A and B, the LSQ command is supposed to find a vector x for which the vector (A*x - B) is of minimal length. Actually least squares refers to minimizing the dot product of (A*x - B) with itself which involves the sum of squared components. This does not usually coincide with x being of minimal length. === Subject: Re: A bug and a related mini-challenge > I recently came across an unexpected bug in the HP48G (and descendants). >> >> The LSQ command is supposed to return the solution with minimum Euclidean >> length when given an underdetermined system (with an infinite number of >> solutions). >> >> Given system A*x=B, with A: >> >> [[ 1 1 0 0 0 -1 ] >> [ 1 0 -1 0 1 0 ] >> [ 0 1 1 1 0 0 ] >> [ 0 0 0 1 1 1 ]] >> >> and B: >> >> [[ 4 ] >> [ 8 ] >> [ 12 ] >> [ 16 ]] >> >> LSQ gives a solution: >> >> [[6.99541284404 ] >> [ 6.15137614679 ] >> [ 2.84403669725 ] >> [ 3.00458715596 ] >> [ 3.84862385321 ] >> [ 9.14678999083 ]] > >> with a Euclidean length of 14.2255745922. But this is not the minimum >> length solution. For example, [ 6 2 1 9 3 4 ]T is a solution and has >> Euclidean length 12.124355653, but there IS a solution with minimum >> possible length. The mini-challenge is to find it, using your HP48G, HP49G+ >> or HP50G. Don't use MATLAB or something similar until you give it a shot on >> your HP calculator. >> >> I was very surprised by this bug. I've used LSQ a lot, and never before >> seen such behavior, and I have no idea what particular circumstance leads >> to the error. This is not necessarily a bug. It's a bug alright. Let me quote from the AUR: If A has less than full row rank (the system is underdetermined), an infinite number of solutions exist. LSQ returns the solution with the minimum Euclidean length. Given A and B, the LSQ command is supposed to find a vector x for which >the vector (A*x - B) is of minimal length. For the underdetermined case when there are an infinite number of solutions, LSQ does more than just finding a vector x which minimizes the residual. Actually least squares refers to minimizing the dot product of >(A*x - B) with itself which involves the sum of squared components. This does not usually coincide with x being of minimal length. In this case there are an infinite number of solutions which give the same minimum for the vector (A*x - B), namely zero. Even though all those solutions give the same zero residual, the solution vectors themselves are of various Euclidean lengths. LSQ is supposed to return the minimum length solution vector. === Subject: Re: A bug and a related mini-challenge > I recently came across an unexpected bug in the HP48G (and descendants). > The LSQ command is supposed to return the solution with minimum Euclidean >> length when given an underdetermined system (with an infinite number of >> solutions). > Given system A*x=B, with A: > [[ 1 1 0 0 0 -1 ] >> [ 1 0 -1 0 1 0 ] >> [ 0 1 1 1 0 0 ] >> [ 0 0 0 1 1 1 ]] > and B: > [[ 4 ] >> [ 8 ] >> [ 12 ] >> [ 16 ]] > LSQ gives a solution: > [[6.99541284404 ] >> [ 6.15137614679 ] >> [ 2.84403669725 ] >> [ 3.00458715596 ] >> [ 3.84862385321 ] >> [ 9.14678999083 ]] > with a Euclidean length of 14.2255745922. But this is not the minimum >> length solution. For example, [ 6 2 1 9 3 4 ]T is a solution and has >> Euclidean length 12.124355653, but there IS a solution with minimum >> possible length. The mini-challenge is to find it, using your HP48G, HP49G+ >> or HP50G. Don't use MATLAB or something similar until you give it a shot on >> your HP calculator. > I was very surprised by this bug. I've used LSQ a lot, and never before >> seen such behavior, and I have no idea what particular circumstance leads >> to the error. This is not necessarily a bug. It's a bug alright. Let me quote from the AUR: >> and I have no idea what particular circumstance leads to the error. I don't have a definitive answer, but the following may help point in right direction. The original system was A*X=B, where A= [[ 1 1 0 0 0 -1 ] [ 1 0 -1 0 1 0 ] [ 0 1 1 1 0 0 ] [ 0 0 0 1 1 1 ]] B = [[ 4 ] [ 8 ] [ 12 ] [ 16 ]] Notice that the last row is linearly dependent on the first three rows. -R1 + R2 + R3 = R4 As a result, the last row provides no new information about the system. Leaving out the last row, A'= [[ 1 1 0 0 0 -1 ] [ 1 0 -1 0 1 0 ] [ 0 1 1 1 0 0 ]] B' = [[ 4 ] [ 8 ] [ 12 ]] Then LSQ gives [[ 3. ] [ 4. ] [ 1. ] [ 7. ] [ 6. ] [ 3. ]] Which is consistent with the original A and B and has a distance of sqrt(120)= 10.9544511501 It's not clear to me why it fails. If I make up other linearly dependent matrices, it seems to find the correct answer, even when I make up a random [4x6] with -R1 + R2 + R3 = R4. Since we have only one example of failure to work with, I can't say whether the following will always work, but putting the system in Row Echelon Form does seem to fix the problem at least for this case. << -> vec mat @ save the vector and matrix << mat SIZE 2. GET 1. + -> n @ get the width+1 << mat vec n COL+ @ augment the column REF @ Row Echelon Form (should not change answer) n COL- SWAP @ split back into vector and matrix LSQ @ run LSQ on reduced vector and matrix >> >> -wes === Subject: Re: A bug and a related mini-challenge posting-account=aJhvMAoAAABqICGF80eSUn6d3D9SANQE Gecko/20080311 Firefox/2.0.0.13,gzip(gfe),gzip(gfe) It gets curiouser and curiouser. It seems just about any benign thing you do to the matrices fixes the problem. << 2 * SWAP 2 * SWAP LSQ >> -wes === Subject: Re: A bug and a related mini-challenge >It gets curiouser and curiouser. It seems just about any benign >thing you do to the matrices fixes the problem. << 2 * SWAP 2 * SWAP LSQ > >-wes Interesting! I find that simply swapping rows 1&3, or 2&3 of both A and B fixes the problem. Also, swapping columns 1&2, 1&3 and 1&6 of A fixes the problem. === Subject: Re: Rechargeables vs Alkalines. Just to remind you folks: 1) Alkalines: Voltage starts from ~1,5V and drops steadily => easy to predict battery life 2) NiMH: ~1,2V, immediately drops a little then STAYS at ~1,2V until at the end sharply drops down the voltage Keep replacements with you! VPN PS: You may consider new Lion 1,5V batteries... Anyone to test those? >>oh I have NiMH batteries AAA/HR03/DC2400/1.2V >> OK, I took fresh NiMH battery, set my battery tester to alkaline and >> chedked the recharegable. Tester was on boundary beween red (good) >> and yellow (80% discharged) >> A.L. So does that mean it will function almost as properly as a regular non- > rechargeable alkaline battery? However, doesn't the Voltage drop fast? > It is 1.2V , 0.3 V lower than regular alkalines. Mabye that is the > only problem? I'm guessing from all of the posts I've searched through > that the calculator only detects the voltage level? > === Subject: Re: Rechargeables vs Alkalines. <6lLGj.316331$yF7.196107@reader1.news.saunalahti.fi> posting-account=zwF5mgoAAABtN7MLMYmZfSdAoZ2In_RC rv:1.8.1.13) Gecko/20080311 Firefox/2.0.0.13,gzip(gfe),gzip(gfe) On Mar 27, 6:18 am, Veli-Pekka Nousiainen > You may consider new Lion 1,5V batteries... > Anyone to test those? http://data.energizer.com/PDFs/l92.pdf It looks like it drops off faster than an alkaline. Comparing to their NiMH, though, the drop is not so dramatic at the end. (energizer nimh: http://data.energizer.com/PDFs/nh12-900.pdf ) Keep in mind though, that at the maximum current draw of a calculator, the lithium will only last ~30% longer than an alkaline. The gain diminishes at lower (typical) current draw (to about zero gain during normal calculations). The benefit from the lithium is their lower internal resistance, which is important for high drain devices. -Jonathan === Subject: Re: Rechargeables vs Alkalines. X The benefit from the lithium is their lower > internal resistance, which is important for high drain devices. > What is high drain (if not the 50g)?? === Subject: Re: Rechargeables vs Alkalines. posting-account=zwF5mgoAAABtN7MLMYmZfSdAoZ2In_RC rv:1.8.1.13) Gecko/20080311 Firefox/2.0.0.13,gzip(gfe),gzip(gfe) On Mar 27, 8:54 pm, Veli-Pekka Nousiainen > What is high drain (if not the 50g)?? Digital cameras, especially the flash, are notoriously high drain. Reading out the CCD constantly doesn't help either. :-) -Jonathan === Subject: Re: Rechargeables vs Alkalines. I mean >100mA or what? Is it just ~high drain (another hard fact down the drain...) ??? > On Mar 27, 8:54 pm, Veli-Pekka Nousiainen >> What is high drain (if not the 50g)?? Digital cameras, especially the flash, are notoriously high drain. > Reading out the CCD constantly doesn't help either. > :-) -Jonathan === Subject: Re: Rechargeables vs Alkalines. > I mean >100mA or what? high or low drain usually refers to a load relative to the battery's capacity. A common value is a load > 1 CA, i.e. the battery's capacity divided by 1 hour. Example: 1500 mAh battery => 1500 mA load. Some figures: a high power flashgun (e.g. Metz 45-series) is able to draw more than 10 Amperes from a decent NiCd-battery-pack for a short moment right after the flash is fired. AC-adaptors for digital cameras are usually rated for currents around 2 Amperes. Finally you might want ask someone who's building RC-cars and similar stuff about what he would consider a high drain application - he probably will add another zero to the previous figures. Yes, *left* of the decimal point. <8) In other words: if the battery is able to run a device longer than about one hour, it's not a high drain device. <8) Dieter === Subject: Re: Rechargeables vs Alkalines. BTW: have anyone tested the battery drain on a 203MHz/ overclocked 50G / 49g+ > I mean >100mA or what? high or low drain usually refers to a load relative to the battery's > capacity. A common value is a load > 1 CA, i.e. the battery's capacity > divided by 1 hour. Example: 1500 mAh battery => 1500 mA load. Some figures: a high power flashgun (e.g. Metz 45-series) is able to > draw more than 10 Amperes from a decent NiCd-battery-pack for a short > moment right after the flash is fired. AC-adaptors for digital cameras > are usually rated for currents around 2 Amperes. Finally you might want > ask someone who's building RC-cars and similar stuff about what he would > consider a high drain application - he probably will add another zero > to the previous figures. Yes, *left* of the decimal point. <8) In other words: if the battery is able to run a device longer than about > one hour, it's not a high drain device. <8) Dieter > === === === Subject: Re: Hp48GX vs HP50G Look deeper try Author= Nousiainen >I have checked www.hpcalc.org but I didn't find a clear comparison between >these calculators. Anyone else know where I can find a clear comparison? > ET === Subject: hpcalc.org website problem? posting-account=XOfZ9woAAACnHbd7tyI_QRfVm9UtK_HP CLR 1.1.4322; .NET CLR 2.0.50727; InfoPath.1; .NET CLR 3.0.04506.30),gzip(gfe),gzip(gfe) The hpcalc website is not fully functional, since several days. Does anybody have any information? === Subject: Re: hpcalc.org website problem? > The hpcalc website... is http://www.hpcalc.org [must include www] === Subject: Re: hpcalc.org website problem? posting-account=XOfZ9woAAACnHbd7tyI_QRfVm9UtK_HP CLR 1.1.4322; .NET CLR 2.0.50727; InfoPath.1; .NET CLR 3.0.04506.30),gzip(gfe),gzip(gfe) > The hpcalc website... ishttp://www.hpcalc.org[must include www] Do you get access to the site? === Subject: Re: hpcalc.org website problem? > The hpcalc website... >> ishttp://www.hpcalc.org[must include www] >> Do you get access to the site? Yes, works just fine. === Subject: Re: hpcalc.org website problem? <47edf4b1$1_6@news.bluewin.ch> posting-account=XOfZ9woAAACnHbd7tyI_QRfVm9UtK_HP CLR 1.1.4322; .NET CLR 2.0.50727; InfoPath.1; .NET CLR 3.0.04506.30),gzip(gfe),gzip(gfe) The hpcalc website... >> ishttp://www.hpcalc.org[mustinclude www] > Do you get access to the site? Yes, works just fine. both hpcalc and ibiblio was alive and kicking. And remains so this morning. So the problem was local, both in time and === Subject: Re: hpcalc.org website problem? > The hpcalc website... ishttp://www.hpcalc.org[must include www] Do you get access to the site? Yes. No problem at all. === Subject: Re: hpcalc.org website problem? posting-account=9rwnLQoAAAC_FoizTwvxTD73niS5ww5R Gecko/20080207 Ubuntu/7.10 (gutsy) Firefox/2.0.0.12,gzip(gfe),gzip(gfe) The hpcalc website... ishttp://www.hpcalc.org[mustinclude www] Do you get access to the site? All is regularly operating, here in Italy :) Site is accessible, can login, can do everything. Giancarlo === Subject: Re: hpcalc.org website problem? posting-account=QSfVcQoAAADkgn5Et4hWcvOa13b-lJy1 Gecko/20080311 Firefox/2.0.0.13,gzip(gfe),gzip(gfe) > Do you get access to the site? > All is regularly operating, here in Italy :) > Site is accessible, can login, can do everything. > Giancarlo I do have normal access to the site, from Costa Rica (Central Am.8erica) Daniel === Subject: Re: hpcalc.org website problem? posting-account=XOfZ9woAAACnHbd7tyI_QRfVm9UtK_HP CLR 1.1.4322; .NET CLR 2.0.50727; InfoPath.1; .NET CLR 3.0.04506.30),gzip(gfe),gzip(gfe) > Do you get access to the site? > All is regularly operating, here in Italy :) > Site is accessible, can login, can do everything. > Giancarlo I do have normal access to the site, from Costa Rica (Central Am.8erica) Daniel Once I had some great times in Puntarenas. It seems that the problem experienced here is not isolated to hpcalc.org, but affects other sites related to ibiblio.org. === Subject: Re: hpcalc.org website problem? > It seems that the problem experienced here is not isolated > to hpcalc.org, but affects other sites related to ibiblio.org Can you display www.ibiblio.org ? Is your network being filtered by blocking software/hardware? Come to think of it, there are HP48 grayscale supermodel images at www.hpcalc.org, aren't there? And what about Tanya in 49G/Emu49 Boot ROM? (:) === Subject: Re: hpcalc.org website problem? > The hpcalc website... ishttp://www.hpcalc.org[mustincludewww] Do you get access to the site? All is regularly operating, here in Italy :) > Site is accessible, can login, can do everything. > Giancarlo I am located in Sweden, and have tried from several network enviroments, but no go. I noticed the problem 2 days ago. === Subject: Re: Programming: Menus and Display >> isn't [line 11 is last line for font6] fixed in 2.10? What's 2.10? Oh, isn't that odd -- the calc which I tested says: Version HP49-B > Revision #2.10-7 Hmmm, what 2.10 and WHERE! Does it include the EMU ROM ?? Parisse version or something else? [<-l] [CONT] === Subject: Menu Label Builder posting-account=5jQj0AoAAAAGAGJcqkkpunLMBpVi1N5o 2.0.50727),gzip(gfe),gzip(gfe) I own The HP 48 Programmer's ToolKit by James Donnelly but lost the software years ago. Does anyone know where I can obtain a legal copy of that software? If not, is there a similar program out there? === Subject: Re: Menu Label Builder schrieb im Newsbeitrag >I own The HP 48 Programmer's ToolKit by James Donnelly but lost the > software years ago. Does anyone know where I can obtain a legal copy > of that software? If not, is there a similar program out there? > Have you checked at www.hpcalc.org ? Slightly OT: I remember that back in 1992 Detlef (Mueller) and me made kinda late night online contest (actually over the phone, me at W&W HQ, Detlef in Hamburg) where the target was to create the smallest version of a menu label builder. After a few hours of talking and programming simultaneously, we both ended up with a program of about 1KB in size. Detlef's version was a few bytes smaller, and my version was slightly faster... *That* were the times;-) BTW: Detlef's version should be available in our RPL48 package, also on hpcalc.org === Subject: Re: Menu Label Builder > software years ago. ÊDoes anyone know where I can obtain a legal copy > of that software? ÊIf not, is there a similar program out there? Have you checked atwww.hpcalc.org? Slightly OT: > I remember that back in 1992 Detlef (Mueller) and me made > kinda late night online contest > (actually over the phone, me at W&W HQ, Detlef in Hamburg) > where the target was to create the smallest version of a menu label builder. > After a few hours of talking and programming simultaneously, > we both ended up with a program of about 1KB in size. > Detlef's version was a few bytes smaller, and my version was slightly > faster... *That* were the times;-) BTW: Detlef's version should be available in our RPL48 package, > also on hpcalc.org Raymond, HP 49g+ and would like to use a similar program for that calculator. I will check www.hpcalc.org for such a program. Mark === Subject: Re: Menu Label Builder <> schrieb im Newsbeitrag > schrieb im > own The HP 48 Programmer's ToolKit by James Donnelly but lost the > software years ago. Does anyone know where I can obtain a legal copy > of that software? If not, is there a similar program out there? > [..] > BTW: Detlef's version should be available in our RPL48 package, > also on hpcalc.org >Raymond, HP 49g+ and would like to use a similar program for that calculator. >I will check www.hpcalc.org for such a program. Mark > Hi Mark, in case you don't find the Menu Label Builder on hpcalc, I could send you the textual source, which should be compilable right on your calc... The program consists of pure SysRPL, no Saturn assembly this time, so porting it to the doorstop series shouldn't be too difficult;-) HTH Raymond === Subject: Re: Menu Label Builder posting-account=Q2CEjQoAAACue5ZDDhUeRzv1w0u8hxZE Gecko/2008030714 Firefox/3.0b4,gzip(gfe),gzip(gfe) > in case you don't find the Menu Label Builder on hpcalc, So what exactly is a menu label builder? I was 10 years old in 1992. . . :-) TW === Subject: Re: Menu Label Builder TW schrieb im Newsbeitrag >> in case you don't find the Menu Label Builder on hpcalc, So what exactly is a menu label builder? I was 10 years old in > 1992. . . :-) > It's a minimalistic graphics editor, taylored for the 21x8 menu grobs of the HP-48 series. While running the program, you can 'draw' on the work space, and it'll simultaneously show your current drawing in normal and inverse video on menu key locations, as a preview. This program is very handy for creating custom menu grobs. Raymond === Subject: Re: Menu Label Builder posting-account=Q2CEjQoAAACue5ZDDhUeRzv1w0u8hxZE Gecko/2008030714 Firefox/3.0b4,gzip(gfe),gzip(gfe) > It's a minimalistic graphics editor, That is what I thought, but I thought it might be something more complicated. TW === Subject: Re: Problems moving large objects between ports -- HP50+, ROM 2.09 X > Now, for an object to be moved from say the SD card to Port 2, it must > So if you want to copy a 100kB file from Port 3 to Port 2, you must have Jean-Yves AAARGHH Why not use 4K parts to do the transfer? You could check that it fits, then do a piecewise transfer Perhaps even an adaptive one: Naturally one leaves room for alarms etc interrupts unless they are disabled while doing the Move === Subject: Re: Problems moving large objects between ports -- HP50+, ROM 2.09 JYA: > for an object to be moved from say the SD card to Port 2, > So if you want to copy a 100kB file from Port 3 to Port 2, VPN> AAARGHH The leap from HP48 series to HP49/50 series was, however, which could provide all the room needed for object transfers, flash packing, or archiving HOME, provided that the user can resist the temptation to overextend into a much bigger HOME than he can afford, which is the same factor now devastating the economy :) === Subject: Menu Label Builder for HP 49g+ posting-account=5jQj0AoAAAAGAGJcqkkpunLMBpVi1N5o 2.0.50727),gzip(gfe),gzip(gfe) I neglected to mention that I would like a menu label builder program for my HP 49g+. I assume the program Raymond Del Tondo referred to will not work on the HP 49g+ because it uses a different processor. Is that correct? === Subject: Re: Menu Label Builder for HP 49g+ [ARM vs. Saturn?] > I neglected to mention that I would like a menu label builder program > for my HP 49g+. I assume the program Raymond Del Tondo referred to > will not work on the HP 49g+ because it uses a different processor. > Is that correct? The architecture of the HP49G+/50G series calculators emulates the original HP Saturn processor, in a lower-level system which is transparent to the user. Thus, the platform does not give the user a new ARM-based computer to play with, but gives instead a continuation of the same original RPL calculator operating system (with enhancements, of course, in Filer, CAS, and new hardware such as larger screen, USB, and SD cards). The independent HPGCC project is, on the other hand, based on breaking through the existing barrier between the hidden ARM OS and the emulated original Saturn, in effect giving you a new platform. If some of Raymond's programs are not portable to the 49/50 series, it is because of differences between series even in the Saturn level OS, What direction would the series have taken if Raymond had been playing on Bill Wickes' original all-star team? ;-) Just tunnel into any nearby parallel universe to find out. Best wishes from http://www.mum.edu and http://www.maharishischooliowa.org . === Subject: Re: Menu Label Builder for HP 49g+ [ARM vs. Saturn?] posting-account=5jQj0AoAAAAGAGJcqkkpunLMBpVi1N5o 2.0.50727),gzip(gfe),gzip(gfe) I neglected to mention that I would like a menu label builder program > for my HP 49g+. ÊI assume the program Raymond Del Tondo referred to > will not work on the HP 49g+ because it uses a different processor. > Is that correct? The architecture of the HP49G+/50G series calculators > emulates the original HP Saturn processor, > in a lower-level system which is transparent to the user. Thus, the platform does not give the user a new ARM-based computer > to play with, but gives instead a continuation of the same > original RPL calculator operating system > (with enhancements, of course, in Filer, CAS, > and new hardware such as larger screen, USB, and SD cards). The independent HPGCC project is, on the other hand, > based on breaking through the existing barrier > between the hidden ARM OS and the emulated original Saturn, > in effect giving you a new platform. If some of Raymond's programs are not portable to the 49/50 series, > it is because of differences between series > even in the Saturn level OS, What direction would the series have taken if Raymond > had been playing on Bill Wickes' original all-star team? Ê;-) Just tunnel into any nearby parallel universe to find out. Best wishes fromhttp://www.mum.edu > andhttp://www.maharishischooliowa.org . So the ARM OS uses different code to give the user the same functions as the original HP Saturn processor? Why did HP choose to use the ARM OS and why did they deem that it remain hidden? === Subject: Re: Menu Label Builder for HP 49g+ [ARM vs. Saturn?] posting-account=Q2CEjQoAAACue5ZDDhUeRzv1w0u8hxZE Gecko/2008030714 Firefox/3.0b4,gzip(gfe),gzip(gfe) > So the ARM OS uses different code to give the user the same functions > as the original HP Saturn processor? Not exactly. Some of the key functions actually call ARM routines that greatly speed up operation. Things like garbage collection can happen many times a second, so that is a routine that was replaced with an ARM version. The rest of the system is an EMULATOR that runs essentially the same OS that was found in the 49G. As far as code goes, most programs are blissfully unaware that they aren't being run on native hardware. > Why did HP choose to use the ARM OS and why did they deem that it > remain hidden? The reason for this is simple - time and money. It was much quicker and more economical to release an upgraded and slightly faster version then to spend years developing something new. It also meant that essentially all of the stuff on hpcalc.org would run on the new unit. HPGCC launhes out of the emulation, runs directly on the processor, and returns to the emulator. TW === Subject: HP50G Smaller Keys? posting-account=oleuQgoAAAD3rQ9zJAgXXAwIijDVEncA Browser; .NET CLR 1.1.4322; .NET CLR 2.0.50727),gzip(gfe),gzip(gfe) I don't know if it's my illusion, but I noticed while looking at an actual unit the 50G keys seem a little more narrow than the 49G's, leaving more key spacing. The key labels also seem a little bigger, all making the 50G look more old school, and better. Can anyone confirm this? === Subject: Re: Menu Label Builder posting-account=Rr1BpAoAAAC35AO1lUfz6ZegzjxdWL8v Gecko/20080311 Firefox/2.0.0.13,gzip(gfe),gzip(gfe) > It's a minimalistic graphics editor, That is what I thought, but I thought it might be something more > complicated. TW as complicated as your surveying program which I wonder how you miss to mention? === Subject: Re: Menu Label Builder posting-account=Q2CEjQoAAACue5ZDDhUeRzv1w0u8hxZE Gecko/2008030714 Firefox/3.0b4,gzip(gfe),gzip(gfe) > as complicated as your surveying program which I wonder how you miss > to mention? As opposed to someone who constantly says useless, snippy remarks that don't help anyone? You seem to have gotten more bitter over the past month or two. Did someone run over your kitten? I was thinking it might be a complete new menu system that would build menus, control they keys and shift functions, etc. If so, I was wondering how such a thing fit into 1Kb (and then I would bow to Raymond and worship his amazing skill), and also what advantages it would have over the built in menu system which is already so darn flexible. I think you will find when the discussion is on a) surveying, b) connecting the calculator to something serially, or c) HPGCC stuff, it gets mentioned because it is very relevant to the conversation. Not everyone thinks onboard is a good thing. Leica's is the best onboard, but I still think it is limited. Just a few weeks ago I had a guy who has been using Leica's onboard for years tell me how much it sucks and how glad he is to be done with it. This guy is a certified Leica repair tech and works for a large company that has around 100 Leicas with everything on up to the top of the line laser scanners. . . TW PS - just in case you'd like to know, the surveying program does have a simple editor to graphically build menu label graphics. It isn't beautiful as it isn't really a key functionality and so I could care less, but you can stick in text and draw using the cursor keys just fine. :-) === Subject: Re: Programming: Menus and Display <466e3744$0$22461$5a62ac22@per-qv1-newsreader-01.iinet.net.au> posting-account=82unwgoAAABKn0foN-teZr4DLYjqF_TL Gecko/2008030714 Firefox/3.0b4,gzip(gfe),gzip(gfe) On Mar 28, 2:23Êpm, Veli-Pekka Nousiainen > isn't [line 11 is last line for font6] fixed in 2.10? What's 2.10? Oh, isn't that odd -- the calc which I tested says: Version HP49-B > Revision #2.10-7 Hmmm, what 2.10 and WHERE! > Does it include the EMU ROM ?? > Parisse version or something else? > [<-l] [CONT] AFAIK, The only publicly available 2.10 is BP's 2.10, which is available at his web site. You should be able to tell that the rom running is a beta of version 2.10-7. The only such beta is BPs. Other 2.10 ROMs do exist, and I would not be surprised if JYA had some, but AFAICT nobody but HP and apparently BP have the authority to release ROMs. And HP seems to have an official policy to not release any new ROMs. Unless somebody can convince HP to let them take over software development responsibilities for the system (which would also probably require providing full support for the calculators), it seems quite likely that no new official ROMs will over be seen. However more unofficial ROMs might one day be made once all the relevant copyrights expire, if anybody still cares that far into the future. === Subject: Re: Selling HP 50g > I have a HP 50g that I've only used for a short period of time. I am > selling this because I was disappointed by its CAS and want a TI-89 > instead. So, if you would like one, it's up on ebay ! Any reason why you were disappointed with the 50g CAS? Just wondering, since I've never really used an 89 very much. === Subject: Re: Selling HP 50g Any reason why you were disappointed with the 50g CAS? Just wondering, > since I've never really used an 89 very much. The only thing I can think of is the solver. In situations where the 50g returns Not reducible to a rational expression the TI-89 automatically uses the numerical solver. On the TI-89, you do not need to supply an initial guess on the numerical solver. It also tries to return all solutions, though sometimes it chokes. I think the 50g offers much more flexibility in terms of the form of the answer, though. This is especially evident in trig, with the TRIGSIN, TRIGCOS, and TRIGTAN functions that leave the TI-89 in the dust! S.C. === Subject: Serial Port Connectors: HP49, HP49, HP200LX, PC posting-account=wcRVPwkAAACV9qEmJuiPuuML0GXuIOZ6 4.90),gzip(gfe),gzip(gfe) As many diagrams have led to confusion, I collected all of the Serial port pinouts for common devices. At the 'Serial Ports' link here: I've gathered all of the relevant pinouts. Note that these diagrams are looking into the connector on the respective pieces of equipment. Also note that the HP200LX and the HP49 connectors, while physically the same have different pinouts. Hope this information is helpful. TomCee === Subject: Re: Serial Port Connectors: HP49, HP49, HP200LX, PC tomcee schrieb im Newsbeitrag >[..] > Also note that the HP200LX and the HP49 connectors, while physically > the same have different pinouts. > The plugs for the HP-38G, the 200LX, and the 49g are _not_ the same physically. They have a different 'notch' on top of the plugs. HP-38G: none HP-200LX: wide 49g: narrow HTH Raymond === Subject: Re: Serial Port Connectors: HP49, HP49, HP200LX, PC > Also note that the HP200LX and the HP49 connectors, while physically > the same have different pinouts. The plugs for the HP-38G, the 200LX, and the 49g are not the same > physically. They have a different 'notch' on top of the plugs. HP-38G: none > HP-200LX: wide > 49g: narrow HTH Raymond Raymond: TomCee === Subject: Newton's method I am an AP Calculus teacher with many successful years employing the 48G. Now using the 50G and I would like to know how to put Newton's method in the 50G. It was never a problem on the 48G, but what worked for the 48 does not work for the 50. On the 48 you define N(x)= x-f(x)/ f'(x) and also define f(x). Enter a numeric value, press N from the var list and the calc figures out the rest, then you can iterate as desired. When I do this on the 50, all I get is infinity for the answer, which apparently means the 50 doesn't recognize f(x) or if it does, it comes up with zero for its derivative. I don't know much any help. Also, anyone else encountered the 50's bizarre treatment of odd fractional root functions and their graphs? For instance, after many many failed trials, the only way I have found to get the 50 to draw the entire graph of y=x^(5/3) is to graph y=x^2/cube root of x! This is pathetic, and I hope someone knows a much more elegant solution. Srock === Subject: Re: Newton's method posting-account=itRlSAoAAADIJvOzSGM8mjsK7gvlk1wg Gecko/20071127 Firefox/2.0.0.11,gzip(gfe),gzip(gfe) Since nobody of the experts seems to be online, let me give you a first hint in RPN notation, which is the only one I know: 1) store your function to a variable F: e.g. 1-x^2 F STO> 2) calculate the Newton formula and store it to the variable G: F F X delta / +/- X + G STO> 3) enter a starting value and calculate the next value by: starting value X STO> G EVAL and so on Of course you can do all of these statements within a little program and hence get an iteration step by pressing one button Peter === Subject: Program to launch ROM Update? posting-account=Q2CEjQoAAACue5ZDDhUeRzv1w0u8hxZE Gecko/2008020514 Firefox/3.0b3,gzip(gfe),gzip(gfe) Does anyone know if it is/would be possible to have a routine that would jump into the ROM update screen? After a rather frustrating hour trying to help someone do this it got me thinking about a solution. I know that you can do the pop a battery, wait a fraction of a second and then immediately press + and -, but some people just can't get the timing on that. I ask because there are some (essentially) electronic illiterate people using our yellow box which doesn't give them access to the reset button on the back, nor the USB port. I am talking people that have trouble understanding what is meant by press the ON key and drag the file to the other window. I was just thinking about how to make it easier for them to update the ROM should it be necessary. If there was a way I could make a little ROM Update routine that they could select from a menu, that would be perfect. I am thinking that might not be possible however. . . TW === Subject: Re: Program to launch ROM Update? > I know that you can do the pop a battery, wait a fraction of a second > and then immediately press + and - Or ON+F, then ON, then + and - (before calc restarts) Choose 2. Selftest to skip the ROM update :) === Su