A16 I installed Remote V 4.3 by Bjorn Gahm the other day and I successfully setup the codes for my Sony CD and Video recorder. (Impresses my friends (Something about small things and small minds...)) I have a reather oldish (about fifteen years I suppose) NEC TV and I can not seem to find the correct codes. Can anybody advise me? BTW the origonal remote from the TV died a natural death some time ago and has gone to the lounge room in the sky. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143270 Heiko Arnemann escribi227: Wasn't it ALPHA LS ALPHA? Xref: newsmst01.news.prodigy.com comp.sys.hp48:143299 I will try that, but in the edit mode it is ALPHA ALPHA LS ALPHA, when using default flag settings (a flag around the sixties) ..Heiko Xref: newsmst01.news.prodigy.com comp.sys.hp48:143274 In my opinion, the single major advantage of TI 92/89 over HP49 is the built-in spreadsheet. Do you know any comparable program for our best friend? Thank you, ============================================================ Paolo Cavallo I am a teacher at heart, and there are moments in the classroom when I can hardly hold the joy. P. J. Palmer, 1998 paolo.cavallo@iperbole.bologna.it http://www.alberghetti.it/paolo.cavallo/pc.htm ============================================================ EMERGENCY - UNO STRACCIO DI PACE - www.emergency.it Xref: newsmst01.news.prodigy.com comp.sys.hp48:143277 Paolo Cavallo wrote: Soon XCell 2.0 will be released, it is now in the beta testing phase -- This message was written with 100% recycled electrons Pivo OR Hit the HIST key to pick up an element from the stack OR have a huge amount of ready-made accurate Matrix ML-routines as useful programmable user commands OR AAAARRRGGGHHH! *use the TI 89 style spreadsheet matrix writer* Somebody have to do something, TI is winning here!!! VPN in TiBORG-terror !! news:9tb1c9$9r2$1@pinco.nettuno.it... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143294 news:9tbbb1$nuf$1@news.kolumbus.fi... Please don't be so sorry. I owned a 48G for years, then the best thing I could think to do was to add to it a 49G! But I use also TI92, especially for CBL. ??? ============================================================ Paolo Cavallo I am a teacher at heart, and there are moments in the classroom when I can hardly hold the joy. P. J. Palmer, 1998 paolo.cavallo@iperbole.bologna.it http://www.alberghetti.it/paolo.cavallo/pc.htm ============================================================ EMERGENCY - UNO STRACCIO DI PACE - www.emergency.it Xref: newsmst01.news.prodigy.com comp.sys.hp48:143313 Hello, Since I couldn't wait for rom 1.19-7 to come out I decided to upgrade from 1.18 to 1.19-6. I followed all the directions in the html file. The only thing funny I experienced was that the SaveBank0 saved an empty list (not too surprising since I had nothing in bank0). So I skipped the part about restoring the list. The problem now is that any port memory access takes 25 +/- 1 secs to complete. The only memory access that is fast is storing and recalling objects stored in the home directory via the VAR softmenu and the CUSTOM menus. Some of the operations that I tested which took 25 seconds are UNDO, the LIB softmenu, accessing the file manager, most operations within the file manager that require rereading the memory tree, RCL, and STO. My guess is that there is something in my memory that is corrupted. What I'm thinking of doing is: A. Erase all user memory B. Reset calculator and reflash This info from my file manager may be helpful: 0:IRAM 234KB 1:ERAM 255KB 2:FLASH 1021KB Home 234KB |-CASDIR |-VAR |-SI Any help is appreciated-- stuff like how to go about clearing user memory and possible bugs that may cause this. On a sidenote are there any programs for linux that operate similarly to HPComm for windows? I'd hate to have to boot into windows everytime I need to do calculator maintenance. Thanks a lot, -Victor Date: Mon, 19 Nov 2001 22:44:24 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143314 Victor wrote: I use C-Kermit. Not really similiar, but it works fine, I haven't used HPComm for a while now. Check these posts: http://groups.google.com/groups?selm=3BD04D78.309344DE%40iname.com http://groups.google.com/groups?selm=3B9A037E.939A1E43%40iname.com http://groups.google.com/groups?selm=3B53878E.AD355404%40iname.com There are even more. I vote for this as a FAQ, too (and I'm very sorry, I don't know how to put that in good english). Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143265 Here is some documentation about the message handler in libraries and how it can be used. Let me know if you find any mistakes. - Carsten The library message handler =========================== Libraries on the HP49G can contain a message handler. This program is called by the operating system on various occasions, to give the library a chance to modify menus, provide online help for its commands and other actions. When creating a library from a directory, the reserved variable $EXTPRG can contain the name of an ID in the directory which will later become a rompointer in the library. This rompointer must be a program which accepts a BINT on level one (the message) and, depending upon the specific message, other arguments on higher stack levels. The message handler as described here works only from ROM 1.19-6 onwards. Earlier ROMs implemented only part of this functionality. Menu extensions --------------- The majority of messages can be used to extend some builtin menus. Among these are the APPS choose menu, several other choose menus, the SEARCH, GOTO and Tools submenus in the editor menu etc. When the message handler is called to extend a menu, the current menu is on the stack either as a list or as a meta. The program can then modify this menu and return it. So the stack diagram for menu extensions is either The message BINT stays on the stack, so that the message handler of another library can be called immediately to do its work in the same way. The following menus on the HP49G can be extended using library messages. #msg Menu Menu Type ------------------------------------------------------------------ BINT0 APPS list BINT1 Main STAT menu list BINT2 Hypothesis submenu in STAT list BINT3 Confidence Interval submenu in STAT menu list BINT4 Finance menu list BINT5 Numeric Solver menu list BINT6 Time menu list BINT8 Games (in APPS) meta BINT11 Editor SEARCH menu (when flag -117 is clr) list BINT12 Editor TOOLS menu (when flag -117 is clr) list BINT13 Editor GOTO (when flag -117 is clr) list BINT14 Editor SEARCH menu (when flag -117 is set) meta BINT15 Editor TOOLS menu (when flag -117 is set) meta BINT16 Editor GOTO (when flag -117 is set) meta As an example, we show a message handler of a library with ROMID 1234. This handler will add the whole library menu to the APPS menu, and a particular rompointer to the Games menu. When adding to the APPS menu, the example also makes sure that the new item is numbered just like the other items in the APPS menu. :: ZERO OVER#=case :: SWAPINCOMP (Save #msg (ZERO), explode list) #1+DUP #.$ (Make index for my entry) .My Library !append$ (Add name to index number) ' :: % 1234. InitMenu% ; (action: set my menu) TWO{}N (Make list from label and action) SWAP P{}N SWAP (add new entry, get the ZERO back) ; EIGHT OVER#=case :: DROP (drop the message) { PlayMe ROMPTR 4D2 0 } (my new entry for games menu) SWAP#1+ (add to meta) EIGHT (need the message also on exit) ; ; Online Help for library commands -------------------------------- On the HP49G, all the CAS commands have a short help text which can be displayed from the catalog, or with the SDIAG command in the Emacs library. When the catalog choose box highlights a CAS command, the menu under the choose box has an additional button, the HELP button. Pressing this button shows the corresponding help text. External libraries can provide help for their commands in a similar way, using the message handler and messages NINE and TEN. Message NINE is a query if the library provides help for a given rompointer. The stack diagram is where the TRUE/FALSE on stack level two indicates if the library is prepared to provide help for the rompointer level 3. This message is used to determine if the HELP button in the CATalog should be turned on. Message TEN is then used to actually display the help when the user presses the HELP button. The stack diagram here is Before pushing FALSE, the message handler should display the help text. The following example is a message handler which provides a short help string for every visible command in the library. :: NINE #=casedrop :: DROPTRUE NINE ; (we have help for all rompointers) TEN #=casedrop :: DUP DECOMP$ NEWLINE&$ SWAP (save the command name as a string) { (list of help strings) Help text for romptr 0 Help text for romptr 1 ... Help text for romptr N (the last visible rompointer) } SWAP#1+ NTHCOMPDROP (extract the correct help string) &$ (Add the command name) FALSE SWAP ViewStrObject (display the text) ; ; Note that ViewStrObject pushes FALSE on the stack, which is the required return value of message TEN. The message handler gets a bit more complicated if help is only provided for a few rompointers. Then the handler of message NINE must check the rompointer against a list, and message TEN must use Lookup or something similar to extract the help text. You can use the OT49 library in such cases. Instead of simply displaying a string, message TEN can also do more complicated things, like launching a whole application to provide help. The library menu message ------------------------ If the menu of a library is invoked via the LIBS menu (rightshift 2), the romid of this library is sent to the message handler of the library. The library may use this for easter eggs-like stuff (display an icon (see for example the LIBMAN library), do something funny with the menu (e.g. LTool) or play a melody. It can also change the menu settings, for example to provide functionality for the shifted menu buttons (e.g. ConstTools). The stack diagram for this message is The following example is the message handler of a library # 60F and it temporarily displays a copyright notice when the library menu is selected. :: # 60F OVER#=case :: ZEROZERO (c) 2001 Some Author SetDA1Temp ; ; Xref: newsmst01.news.prodigy.com comp.sys.hp48:143272 Carsten Dominik wrote: Great, this had become somewhat of a FAQ, I trust that with this document all the How do I get my library in APPS questions are history -- This message was written with 100% recycled electrons Pivo Xref: newsmst01.news.prodigy.com comp.sys.hp48:143282 news:3BF9042C.4A4F7756@student.tnw.tudelft.nl... all the AND Qs like How do I get rid of the 23 apps in my menu start pouring in. Use APPSMAN from Wolfgang Rautenberg... Just my 3_cEuro... VPN This message was written with 100% fresh positrons Xref: newsmst01.news.prodigy.com comp.sys.hp48:143290 How to substitute dA in integral? A2 / | | p dA | / A1 with: dA=sin(a) * r * b * da (a is angle in radians) giving a2 / | | p*sin(a)*r*b da | / a1 expecting A1 & A2 to correctly change to a1 & a2 I don't know how to do it with SUBST PLEASE HELP! Thanks in advance. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143281 Salut ! 'XROOT(3,1-Sqrt(X))=0' 'X' -1. ROOT Gives: Error: Bad Argument Type BUT answer is correct: 3. (1.,-1.0000005) I wonder if anyone else has this regardless of R/C or Cvars Veli-Pekka Xref: newsmst01.news.prodigy.com comp.sys.hp48:143267 If the result is A*EXP(BX) and you want it in the form D*E^X then just use the following to convert.... D=A (ie your required D value is the same as the calculator's A value) E=e^B (ie natural anti-log of B value) Julian Fondren wrote: -- Colin Croft Old mathematicians never die; they just lose some of their functions. ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ====================================== Xref: newsmst01.news.prodigy.com comp.sys.hp48:143271 As anyone with a calculator can see, this doesn't do what I want. ****************** But is really best, 19-6 because there are some reports in this group about the bugs in the rom!! I am beginning in sysRPL and I am finding the simples ways, what do you think I should begin in the MASD or HPtools? Xref: newsmst01.news.prodigy.com comp.sys.hp48:143288 news:9176bfd9.0111190814.ff9e002@posting.google.com... X about X While I will report some bugs here nowadays quite often the ROM 1.19-5 is the best ! No - WAIT ! The ROM 1.19-6 is even better. I will _never_ go back into older ROMs and I'm waiting for a bug fixed 1.19-7 Veli-Pekka, with great confidence in the freelancer Jean-Yves Avenard Xref: newsmst01.news.prodigy.com comp.sys.hp48:143301 I have a 48 sx (the previous version of the GX). I love the calculator, but the on button frequently fails to work. If I can get the calculator on, then everything else works fine. Has anyone else had this problem and discivered a way to fix it? HP no longer supports it, is there anywhere else I could take it for repair? Xref: newsmst01.news.prodigy.com comp.sys.hp48:143253 I need to find a store or someone that has the connectivity cord for the hp/computer. I live in Des Plaines, IL, so if you know anywhere nearby there, that would be great Xref: newsmst01.news.prodigy.com comp.sys.hp48:143275 Just order from samsoncables.com Xref: newsmst01.news.prodigy.com comp.sys.hp48:143280 Peter Karp wrote: ... I'm getting ready to buy a 49G. Are all of the new ones made in China, or do I need to specifically look for one with a Chinese serial number? Thanks, Glen Henshaw machine@ssl.umd.edu University of Maryland Space Systems Laboratory How knows what are in stores - even 2003... Check *always* the ID for CH...if you're conserned VPN PS: Do you play saxofone? Xref: newsmst01.news.prodigy.com comp.sys.hp48:143296 Veli-Pekka Nousiainen wrote: No, why? Glen Some instrument then? I just have this itschy feeling... :) VPN, usin Nosy, being nosy.. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143298 Veli-Pekka Nousiainen wrote: Took 9 years of piano when I was little, hated it the entire time. Haven't touched it in at least 10 years. The most musical things I play now are CD's. Sorry... Glen Xref: newsmst01.news.prodigy.com comp.sys.hp48:143311 Veli-Pekka Nousiainen wrote: Seriously, though, where are 49g's made now? Are there two plants, one in Indonesia and one in China? How does one find this kind of information? Glen Xref: newsmst01.news.prodigy.com comp.sys.hp48:143312 Hi all, I own a couple of calcs, namely a TI-89 and a HP32S, I am contemplating buying a HP49G but am unsure if it is worth it. I use my HP a lot more than the TI mainly because I now prefer RPN to Algebraic logic. but i don't know if i'll get much more functionality out of the 49G. I know a lot of people in this ng have both a ti-89 and a 49g and was after some opinoins. thankyou sv -- You should quit while I'm ahead. Date: Mon, 19 Nov 2001 22:46:16 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143315 der d212mon wrote: You may want to read Tim's comparison: http://www.hpcalc.org/details.php?id=4738 I'll better not say anything more, or I'd get flamed by Bhuvanesh ;-) Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143276 No, but I'm not sure I would want to. The first three or four builds of my programs crash miserably, and I'd prefer to debug them on a virtual calculator rather than make my dear TI-92+ unstable :-) Of course I can write the source code on the calculator, and I have done so occasionally. Bhuvanesh. Date: Mon, 19 Nov 2001 17:23:54 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143292 Bhuvanesh wrote: If it gets unstable, hit ON-C ... nonono, I'm in the wrong film ;-) Anyway, I think it should be possible (what's impossible, anyway?). I have also crashed my HP49 a lot when I learnt the SysRPL and ML stuff, and it's still alive. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman AND the HP 49G User Flash RAM survives a lot of hacking! How many of you have lost anything after 1.19-5/6 ?? Veli-Pekka Date: Mon, 19 Nov 2001 19:01:28 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143302 Veli-Pekka Nousiainen wrote: I've lost port 1 once. My fault though - a *BAD* bug in my own ML code. I've never lost a single byte of Flash. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143308 Yep. The six list variables do have numbers at the end (L1, L2, ..., L6), but other variables can't. This (and the lack of a CAS) is why I haven't even considered getting a TI-83 series calculator. Bhuvanesh. For hp49, have you tried the comand AXL ?? J.Manrique CdU de la ETSIG Xref: newsmst01.news.prodigy.com comp.sys.hp48:143286 Hi, where can I download the two manuals in .pdf-format (basic/advanced) [I would prefer the german version if available] Xref: newsmst01.news.prodigy.com comp.sys.hp48:143289 http://www.hp.com/calculators/techsupport/graphing/49g_userguide/ VPN news:9tbbfo$94f$01$1@news.t-online.com... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143273 use RCLMENU and MENU Xref: newsmst01.news.prodigy.com comp.sys.hp48:143278 Julian Fondren wrote: Not a chance, the menu is build on the program itself. Due to the annoying BZ compression it is hard to reach. I'd say that it is impossible without some hacking. -- This message was written with 100% recycled electrons Pivo Xref: newsmst01.news.prodigy.com comp.sys.hp48:143285 news:3BF91456.384BA778@student.tnw.tudelft.nl... being some hacking. X Hey, Pivo - the Great Hacker - can you give us a documented example VPN Xref: newsmst01.news.prodigy.com comp.sys.hp48:143303 news:3BF91456.384BA778@student.tnw.tudelft.nl... some hacking. I was hoping that the single functions were callable something like XLIB 360 x, so I whipped out my Nosy and managed to crash the calculator within minutes :) contents of ~. Using emacs I managed to decompress the BZstring, and got the source. It's much more complicated than I was expecting it to be! Can you tell me what do ROMPTR 168 1b/1a do? I'm sure my questions denote my utter ignorance on the matter of SysRPL and ML, but I'm afraid I've too little time at the moment to delve into the deepness of the machine secrets... I'd be grateful if you can give me an enlightening insight. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143306 Indeed, some hacking is required. The simples way is to make a program with which will compress and remove the number 1., leaving the menu active. Another possibility with some hacking is 1. ~ DROP @ set the menu # 25845h SYSEVAL @ this is the SysRPL MenuDef@ {MENU} HEAD @ get the MENU command on the stack 2. 3tog 3tog @ make a program from these two items, using the @ OT49 command 3tog. The you have a program on the stack. You cannot edit this program (don't try!), but you can stor it into a variable. Executing it installs the menu. I have asked Wolfgang several times for a programmable command which allows to use the options in this very useful but difficult to access menu - he did not think it was important. Maybe this thread will convince him? - Carsten Xref: newsmst01.news.prodigy.com comp.sys.hp48:143309 news:qzuy9l2d1qy.fsf@sand.science.uva.nl... Trivial but slowest :) /me is very very grateful Why not? I've tried and still nothing has happNO CARRIER Xref: newsmst01.news.prodigy.com comp.sys.hp48:143310 Carsten Dominik wrote: The menu of the BZ-toggler x~ from OT49, more precisely, each option can fully be used in SysRPL-programs! For instance, the ACC-maker MakeACC from ACCeuro on my site produces from some input a self-decompressing and self-evaluating code object (the new ACC program), using the option BZCE from the BZ-toggler menu which has menu position 5. Hence, the end of the MakeACC program looks as follows: :: bla bla x~ (Compress the resulting program which) BINT5 (automatically sets the menu in question) GETPROC (Get the program of BZCE and eval it) EVAL %0 (Immediately restore the previous menu) Initmenu% ; Outside a program's run, the menu is IMHP useful only if a BZ-sting was produced with x~. So what? :-) Hope this helps Wolfgang Xref: newsmst01.news.prodigy.com comp.sys.hp48:143304 Hello! How do I make sybolic integration in the EQ. writer - I have no integration boarders to enter? Please help me! Regards Jacob Nielsen /Denmark Xref: newsmst01.news.prodigy.com comp.sys.hp48:143305 Forgot to say it is on a HP49! news:9tbl5s$7qt$1@eising.k-net.dk... integration Xref: newsmst01.news.prodigy.com comp.sys.hp48:143307 INTVX(f(x)) INTVX is accessable from [<-] [CALC] Jacob Sv276rke Nielsen wrote: Xref: newsmst01.news.prodigy.com comp.sys.hp48:143255 The file transfers ok (using hyperterminal) but I keep getting 'transfer failed' at the end. I have transfered programs/aplets before so I know that works. cheers /Victor news:9t9fsd$sl4$1@news.kolumbus.fi... across Xref: newsmst01.news.prodigy.com comp.sys.hp48:143268 Hi I think the problem you're having is that the converted file is in a format that is suited to the HP48 but not the HP39G. The HP39G doesn't regard images as separate entities that can be transferred to or from the PC in the same way that aplets and programs can. The HP39G only stores images as part of an aplet (the sketch view). Although they can be stored also in the graphics objects G1..G9,G0, these objects don't have a view of their own (like the List & Program views) and so don't have the usual SEND and RECV commands available. That limits their use. Do you, by any chance have access to an HP38G? If you do then there's a way to edit images stored as sketches in an HP38G aplet quite easily using a program called the ADK (Aplet Development Kit). You can then transfer the sketch to a program on the HP38G and then via IR to an HP39G. The problem is that the ADK only works with HP38G aplets. I've been playing around to see if I could find a way to fudge it but no success so far. -- Colin Croft Old mathematicians never die; they just lose some of their functions. ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ====================================== Victor wrote: Xref: newsmst01.news.prodigy.com comp.sys.hp48:143269 That explains a lot. I've just downloaded the ADK, I'll have a play with it as well. Thanks for your help Colin and VPN. cheers /Victor news:3BF8F6C7.A5B16D18@iinet.net.au... format the part way is see Xref: newsmst01.news.prodigy.com comp.sys.hp48:143279 If you have a 38G then ... 1. Store an image in the sketch view. If you don't already have one in there and it doesn't matter to you what it is (given you're going to replace it later anyway) then just create any old image using the sketch tools. 2. Transfer the aplet containing it to the PC 3. Use the ADK to open the aplet and then use the Views/Sketch view command to see the sketch 4. Click on the Edit button. This should launch Windows Paint to allow you to edit the image. When you exit from Paint the edited image is transfered back to the aplet sketch view. The point of the exercise is that while the image is in Paint you can do anything you want to it, including pasting another completely different image over the top (providing that it is the same size). Finally you exit from Paint to transfer the edited image back to the ADK, save it, transfer it back to the 38G and there you go. The final step is getting it from the 38G to the 39G. You can't just transfer the aplet as they are not compatible. If you transfer an aplet from one to the other it will appear to succeed but crash the calc when you run it. The trick is to transfer it via a program. Being purely text they are compatible. On the 38G, START the aplet so that it is active. Now go to the Program Catalog, create a new program and edit it. From within it, chose VARS, SK2(APLT), scroll down to Sketch and across to Page. Press SK4 (VALUE) and then OK. After a pause a massive collection of numbers will appear which is the image in numeric form. Page up to the top and you'll see that the image is stored as a MAKEGROB G0 command. If you were to execute this program it would store the sketch image into grob G0. Transfer the program to the 39G and execute it. The image is now in G0 on the 39G. Now START the aplet that you want to place the image into and go to the Sketch view (presumably blank). Press VARS, SK1(HOME), scroll to Graphic and across and down to G0. Press SK4(VALUE) and the image should be pasted into the sketch view ready for you to edit further or use in any other way. Hope this works for you. Victor wrote: -- Colin Croft Old mathematicians never die; they just lose some of their functions. ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ====================================== Xref: newsmst01.news.prodigy.com comp.sys.hp48:143300 also 32-level (32 shades) and 64-level. see http://ca-on.hpcalc.org/hp49/graphics/grayscale/ Xref: newsmst01.news.prodigy.com comp.sys.hp48:143256 wrote: It can either be binary or BCD. Which is used is determined by the current arithmetic mode which can be set via the instructions SETHEX and SETDEC. Note that the cycle counts are the same for both modes. I program in Saturn assembly and I use it. Like I said if for nothing else you can often use it to speed up an operation by operating on only the nibbles you need. The number of other uses is only limited by the amenability of the algorithm and the cleverness of the programmer. Well, the memory bandwidth of the Yorke Saturn just plain sucks. :D ( see http://groups.google.com/groups?selm=sok63tgmlr4l9sfnvsvum9et36hgd7gs50%404a x.com http://groups.google.com/groups?selm=44l73tksuemq8tcosas1piddfkdpd51glr%404a x.com ) This is due mainly to the antiquated 4-bit multiplexed address/data/command bus that is used internally to communicate with the rest of the on board peripherals and is compounded by the slow memory controllers. There is no fundamental reason why this ancient bus should still be in use. If the Saturn was reimplemented this could easily be remedied. For data the Saturn has two addressing modes : register direct and address register indirect :). M68K wins big time in that arena. The Saturn registers are *highly* asymmetrical. You have 2 address registers that can only be used with the GPR's A and C. A, B and C can work with each other but D can only be used with C. You also have 5 64-bit scratch registers that can only be used with A and C. And that's just the start of the idiosyncracies :D. Note that again there is no fundamental limitation in the Saturn architecture that demands this. Really it's due to the convoluted instruction encoding where they tried to pack all the possible register combinations into 1 nibble. ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143257 I think we have a misunderstanding here. I was never claiming that the Saturn *architecture* ( as it is now ) is superior to that of the 68K. If I were I'd have to be delusional. :D ( which might be the case anyhow ;). The 68K has a much better architecture ( for most things ) which is certainly a major influence on performance. What the debate is about is whether *Saturn instruction cycle counts* are much slower than those of the 68K. I think I've shown that in general the 68K isn't much faster in that area. On Wed, 14 Nov 2001 19:42:07 +1100, Jean-Yves Avenard Why get rid of it because of this when a real return stack could easily be added in a new Saturn implementation? The FPGA's needed to reimplement the Saturn cost even less. ( in quantity ) They aren't slow, at least compared to the 68K. Poor if you mean asymmetrical and hardly any addressing modes. I think the opposite would hold true in this case. :) At any rate, all these problems could easily be solved - and cheaply - by redesigning the Saturn on an FPGA. ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143258 wrote: You can just double many operations but for others it requires multiple instructions. At any rate 64-bit ops are slower on the 68K. For example, to do a 64-bit add you would need to do something like ADD.L D0, D1 ADDX.L D2, D3 which takes 16 cycles as compared to 11 on the Saturn. ( Note that if you want to test for zero you'll have to initialize the Z bit to 1 before the operation which would add two instructions.) A 64-bit single bit shift (multi-bit shifts are more complicated) might look like ROXR.L #1, D0 ROXR.L #1, D1 ( Note that above I again left out the initialization of the Z bit which would add another two instructions. ) which takes 20 cycles as compared to 11.5 cycles on the Saturn. ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143259 On Thu, 15 Nov 2001 10:30:48 +1100, Jean-Yves Avenard No, it wasn't. As others have pointed out there are problems associated with rounding and performing simple operations (such as addition, subtraction and multiplication ) on *decimal* numbers you enter from the keyboard. Not to mention the fact that for financial arithmetic you absolutely need the exact representation that BCD provides. What is that supposed to mean? Any arithmetic operation in binary mode is just as valid in bcd mode. It is just as possible to perform preemptive multitasking on the Saturn as many people have demonstrated ( eg. ShellOS ). Although it's a little more cumbersome because the hardware doesn't help you with saving the registers. ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143260 On Mon, 19 Nov 2001 01:50:37 +0100, Albert Graef Thanks for the link, I think it settles this debate. ( which has raged again and again in unsenet ) I vote for making this a FAQ entry :). ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143261 On Fri, 16 Nov 2001 22:45:36 -0200, /me250 In its current form, possibly. But as for the whole product line all I can say is Not if we can help it! :). All the HP calcs ( including the 48 ) are still readily available where I live ( near Houston, Texas USA ). And the 48 is still listed on HP's web site : http://www.hp.com/calculators/graphing/ . ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143263 news:jkfhvto5u7hdfkt6psot6l736vss5s22rm@4ax.com... X X ME, 2! VPN PS: Note that 2! gives 2, but the Millenium Edition may make your newsreader to give you a blue screen... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143264 On Mon, 19 Nov 2001 02:10:50 -0600, Jonathan Busby Oops. This should be the X bit of course. :) ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143266 Jonathan Busby wrote: I think that this thread in the MoHPC forum has been mentioned elsewhere, but anyway: http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/forum.cgi?read=12616 So the 48/49 calcs are *not* discontinued (at least not right away). Of course, the question remains whether HP will continue calc development. Albert -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag Date: Mon, 19 Nov 2001 17:19:40 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143291 Jonathan Busby wrote: I still wonder how DB can save *all* registers. I think you need at least one RSTK level to do something like this: RSTK=C CD0EX D0=(5) save_space DAT0=C.5 D0+5 C=RSTK DAT0=C.16 etc. I'm confused :-/ Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143316 Jonathan Busby wrote: I dont deny, that it is unusable, but simple it is rarely needed. As such it does not significantly increase program speed in general. Only in very special cases. M68K has for example command, that dumps (or retrieves) all registers onto the stack in one command. Altough it is a nice and fast feature, I don't consider it a big deal from the general speed point of view. I guess, that is an answer, why M68K so easily beats Saturn. Consider, that with those CPU's you don't have cache, so memory adressing is probably 70% of your program commands. On top of it manipulating highly assymetrical registers of Saturn where M68K doesn't need it and calculating memory adresses where M68K uses more advanced addressing modes is probably enoungh to outrun Saturn by tenfold if not more. Jack Xref: newsmst01.news.prodigy.com comp.sys.hp48:143317 Albert Graef wrote: TRADITION :-) BCD was used in the past, there is no reason not to have it standarized. Why neither Intel math coprocessor nor Motorola 68000 series math coprocessor are capable to use BCD math ? Why industry standard Visual C++ V6 does not have neither fundamental data type of BCD nor standard math library operating on BCD ? I'm afraid, BCD is only needed by exotic applications, where loops of bilions of calculations add up enough round off error to affect display. I will be happy to evaluate a program that will run several bilions of calculations on your HP49 within a lifetime of the single set of batteries :-) Again, I challenge anybody to present a reasonable set of calculations that uses 64 bit mantissa for BCD and 64 bit mantissa for binary and let's say 12 digits of LCD display to demonstrate less accurate binary outcome within reasonable caluclation time using either TI89/92 or HP48/49 typical calculation timing. Please, give the example wehen it is affecting 12 digits display. Check MS Windows calculator. It is working example of bianry math. Jack Xref: newsmst01.news.prodigy.com comp.sys.hp48:143318 Jonathan Busby wrote: That is not that much different. Altough M68K loses few cuycles here, on the other hand by having symmetric registers is capable to do way faster (A+B)*(C+D) for example, not even mentioning loading it into the registers where A,B,C,D are all 64 bit integers. Built in HW multiplication here is a bonus :-) You could easily do multiple bit shift using 16 bits chunks of data and 32 bit command. It still will be way faster than Saturn which as you said does not even support bit shift other than one bit. Jack Xref: newsmst01.news.prodigy.com comp.sys.hp48:143319 Jonathan Busby wrote: EXAMPLE PLEASE. There is no problem on MS calculator with binary. Why everobdy pretends, that there is a problem ? If there is a problem it is rather with BCD when 1 3. / 3 * clarly gives the wrong answer on both: HP48 and HP49 (altough it gives the right answer on TI89/92 which seems have a little better DISPLAY ROUND OFF functions than HP) :-) Jack Xref: newsmst01.news.prodigy.com comp.sys.hp48:143320 Julian Fondren wrote: If you use fractional notation, yes. But we are discussing here limited length BCD or binary notation, which are incapable to express many of the fractions exactly depending on the choosen base. For example base 10 will fail to represent correctly 1/3 because it is 0.333333333333(3).... 1/7 because it is 0.1428571(142857).... 1/11 because it is 0.0909090909(09).... etc. 60 and 12. Our time measurement is based on that system (12 hours, 60 minutes etc) This is not by any means theorem with proof. What I'm trying to say, is that for randomly choosen real number the delta between choosen value and represented value will be smaller with binary coding than with BCD coding. The RANDOMLY CHOOSEN here is the key. The assumption here is that you choose floating point number to represent as close as possible set of real numbers. Such set consist of course from rational and irrational numbers. The set of rational numbers is denumerable, where the set of irrational numbers is non-denumerable. In effect what it means, is that if you would randomly choose a point on the real number axis, the probability that it is rational number is belive or not ZERO and that it is a irrational number is ONE :-) Because irrational numbers cannot be represented with limited number of digits in any system based on integer base, therefore binary by having more digits to represent approxiamtion of such value is more accurate because the delta between represented value and true value is smaller. Exactly, you've made my point. The difference is so small, that binary by being more packed and on most CPU's (even Saturn) faster is simply a better coding :-) Jack Xref: newsmst01.news.prodigy.com comp.sys.hp48:143262 news:rjfhvts9u13nv9k9r9oj7mpi01lof3gdpt@4ax.com... X http://groups.google.com/groups?selm=44l73tksuemq8tcosas1piddfkdpd51glr%404a x.com This I have been telling lately! A temporary solution either from HP or from this NG members *must* design a Silver Edition HP 49G with _only_ the Yorke IC downscaled to, say 0.25 micron process and thus gain at least an 8x, propably a 10x speed increase without *any* re-designing of the rest of the hardware. I would call this an HP 49 XP I hate it when the Bitch calls of ACO & any new design !! Otherwise I still love Hewlett & Packard among the shareholders of the Company. Puuh! Veli-Pekka If you don't want to fiddle around with kermit then there's hptalx, with a GTK-based interface. It's available from www.hpcalc.org. Albert -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag Content-Type: text/plain; charset=Windows-1252 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143369 Too late ! But don't worry, there's a _lot_ left for you to document ! ;) Bye, Detlef -- `What a depressingly stupid machine' Detlef Mueller -- Marvin Detlef[DOT]M[AT]hamburg[DOT]de http://mein.hamburg.de/homepage/grendel Xref: newsmst01.news.prodigy.com comp.sys.hp48:143372 Detlef Mueller wrote: For instance, to complete Carsten's preliminary analysis of the System Choose Engine FPTR 2 72, applied, e.g., in LIBman (last version 4.2002 on my site). I took me over a week of hard work to realize the dance of choose boxes (Carsten Dominik), seen if pressing LBox in LIBman. Why? Because of lack of information on the involved message handling of FPTR 2 72, still more involved than $EXTPRG. LIBman is the first tool able to show HELP on library commands outside the catalog CAT in a general setting. - Wolfgang Xref: newsmst01.news.prodigy.com comp.sys.hp48:143376 No, this was SDIAG in Emacs :-) - Carsten Xref: newsmst01.news.prodigy.com comp.sys.hp48:143379 Carsten Dominik wrote: Pardon, Carsten is right. I myself discovered a bug in Emacs version 1.07 (meanwhile fixed) concerning HELP, and Carsten a deeply hidden bug in the first anouncement of LIBman a week ago. Sclerosis :-) - Wolfgang Xref: newsmst01.news.prodigy.com comp.sys.hp48:143388 Hello news:OZvK7.539$F46.140378@news.uswest.net... And to spare you all the fun ? Nah, I'm not that mean... It's all about fun and discoveries... Jean-Yves Hi Kolja! What an interesting thing to do! And it shows the nice capabilities of the 49G. Well, the 49 can't directly substitude dA, but it can substitude A in the integral, even if A appears only in dA. We can express variable 'A' as a function of variable 'a' by solving the differential equation: dA=sin(a) * r * b * da First enter your integral. Then enter 'd1A(a)=sin(a)*r*b' and 'A(a)' and then press DESOLVE. You get : { A(a)= -(r*b*COS(a))+cC0 } Press HEAD to get the equation out of the list. equation. constant. Enter 'A' and press SWAP and then = Now you have A = -(r*b*COS(a)) that is almost ready for substitution. The only thing that remains is the answer to the question: What will be the new variable of integration? Is it r, b or a? The HP49G can't tell that because at the righthand side we have three variables, none of which is somehow labeled to be the new integration variable. With such arguments for substitution the HP49G uses the first variable that it finds. (Which is also nowhere documented, I think) In this case it would use r, which is not what we want. To bring COS(a) to the first place, press down arrow once to get the equation in the EQW, press down arrow again and the press right arrow. Now the whole right handside should be selected. Press down arrow twice and right arrow twice. Now COS(a) should be selected. With COS(a) selected press blue-shift and the left arrow. This moves COS(a) past b to the left. Press blue-shift and then left arrow again. This brings COS(a) at the first place. Press ENTER to enter the edited equation. Now the integral should be on stack level 2 and the substitution equation on stack level 1. Press SUBST and you get what you wanted. (I hope ;-) ) All this could be of course done with a little program: << << SWAP REORDER = SUBST vx STOVX This takes as argument the integral and a substitution equation for da written in the form of a differential equation ( like 'd1A(a)=sin(a)*r*b' ) and returns the substituted integral. Hope it helped a little. Greetings, Nick. P.S. You may have noticed that the program recalls the current VX and restores it at the end. This is due to the fact that DESOLVE changes the current variable to the independend variable of the differential equation and I think you want your current variable to stay as it is until you want to change it. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143338 Hi and sorry for the delay, I was far awy... (What a poet ;-) ) Well, I think the HP49 has a chirality detector, that can tell right from wrong, errh I mean left of course. They made it that way to be conform with the chirality laws of the universe. Only this way it is possible for the HP49G to tell, if it is here or behind the mirror, you know. ;-) Strange enough, I am right handed, but nonetheless the first feet that went fishing, were both the upper right and the lower right. I always knew that I am incompatible with the rest of the universe, but such a discrepance I didn't expect. ;-) Perhaps you can further investigate the behavior of key wear out. If you find the same behavior, then something strange, something very unusual is going on with this calculator. Does it perhaps know that left and right is not in exact symmetry in nature? ;-) Greetings, keep on watching, Nick. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143339 Hmmm, dads tend to be overhanded ;-) Nick Xref: newsmst01.news.prodigy.com comp.sys.hp48:143365 Well, if you keep it in a case instead of in the backpack, it guaranteed that the low-right will fall earlier (if you aren't left-handed)Try to simulate it with ANSYS or your choice of FEM. This is because the keys are under the screen an the calc is at the right side of the table. Of course if you keep it on the backpack the missing feet will be random (and at least one). Julian, where did your poppa kept your calc???? Is this a calc-philia situation? It fitted!!!??? XDDDD Xref: newsmst01.news.prodigy.com comp.sys.hp48:143349 Exactly the same with my HP49G. Well, the calc gives the right answer even when it errors? ;-) Nick. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143354 On Mon, 19 Nov 2001 17:07:45 +0200, Veli-Pekka Nousiainen It's not new. I get the same error and the same result in my HP48SX, Rom D, (C)1989 cc Xref: newsmst01.news.prodigy.com comp.sys.hp48:143351 Just to provide a somehow more automated solution, let's program what Dan said: SWAP base LN / SWAP * base SWAP ^ * Feed this with the exponential expression and the new base (as real number) and it will return what you want. Greetings, Nick. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143360 Science and Engineering Mathematics with the HP 49 G By Gilberto Urroz Is it worth to buy or not? Xref: newsmst01.news.prodigy.com comp.sys.hp48:143362 Pek wrote: Yes, I have it (both volumes), and yes, IMHO, it's worth every penny. It's great to learn how to actually put the calc to work on real math problems, many examples throughout, including some UserRPL programming examples. (It's not about SysRPL hacking, though, but there are good docs on hpcalc.org on this.) AFAIK, this is the only such book available. A (very) short review is available on hpcalc.org. Contents can be found on the author's site, see http://www.engineering.usu.edu/cee/faculty/gurro/HP49G.html. -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag Xref: newsmst01.news.prodigy.com comp.sys.hp48:143371 I too, have both volumes and I'd agree they are certainly worth the $25 each I paid. What I really like is the two volumes really are a complete set. It starts with simple calculator stuff. In general The first 100 pages on basic calc operation, real numbers, complex numbers, lists, programs, functions etc. 35 pages on programing basics 30 pages on algebra basics 30 pages on vectors 80 pages on matricies and linear algebra 80 pages on graphics and char stringgs 40 pages on solving equations Volume II: 75 pages on calc II 90 pages on multivarite and vector calculus 100 pages on differential equations 30 pages on partial diff eq and fourier transformations 45 pages on statistics aside having just great general info on all those topics he also relates them all to the 49G wtih examples and such. I've already used it multiple times in linear algebra and physics, and sure wish I would have had it for calc I and calc II and stats. I dare say its' the best I guess 2 book set for the 49G I've found. Certainly for overall info. There are more complete programing references and the advanced users guide probably lists more functions but the huge problem with the advanced users guide IMO is two fold. First if your going to have a CAS and RPN there should be instructions on how to use each function in each mode. Second there should be a real example (not just syntax) for each mode and each function. On the interesting side when I got volume 1 every odd page was printed upside down, made for interesting reading. The company replaced the book for no charge. Todd On 20 Nov 2001 05:29:19 -0800, teerasarn_t@hotmail.com (Pek) wrote: Xref: newsmst01.news.prodigy.com comp.sys.hp48:143370 Hello, I have just updated the entry database for HP48 and HP49. Look at http://zon.astro.uva.nl/~dominik/hpcalc/entries Thanks to Thomas Rast for considerable input to the ML section, which now also contains compact references of object structures. - Carsten Xref: newsmst01.news.prodigy.com comp.sys.hp48:143377 Hi! I have little trouble with erable..It worked but now sometimes does weird things... 2: (x-1)^2 1: EXPA gives 1: x^2-2*x+1 but COLC this new one doesn t work it prints stuff from 2 to 10 lines (depending on the size of what i wanna factorize) with 1 's, x 's , External , {}, and a - error undefined name Any idea? Is it a flag i might have set or cleared ? Which one ?... Regards, Pied ps: You can answer in English or French...I was too tired to translate it... By all means start with MASD! Its super simple! MASD, Nosy, Emacs together will make you a sysrpl TIGER!! The bugs reported here refer to how MASD decompiles and then recompiles certain (ML?) code, IIRC. At any rate, I use it a lot and it doesn't bother me at all. If I have a lot of typing to do I use text files and copy them into the HP49 emulator on the PC, then compile with MASD. For small (even medium) sized files, Emacs is a dream to use. You can download even large compiled files and edit with Emacs. In bed, in the car, at the beach, the movies... Use HPTOOLS only for source code that exceeds the memory capabilities of the HP49, after you have mastered the language. And only if you are fluent in writing DOS batch files. Or can get through JYA's tight security at his website to download his freely available tools! Dennis Date: Tue, 20 Nov 2001 14:11:12 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143357 Dennis Straley wrote: You can't expect it to recompile *all* code correctly anyway. Consider this: GOSUB .data ... *.data C=RSTK D0=C Or even GOSUB rpl !RPL :: ( do something COLA !ASM CON(5) DOCODE GOIN5 end ... *rpl C=RSTK A=C.A PC=(A) *end Jazz tries guessing, and it works pretty good for the GOSUB...C=RSTK code (also, DB won't single-step such a GOSUB). However you can never know what tricks the programmers use. The movies?? I don't get that :-/ Why don't you just enjoy the *movie*? The Jazz sources contain a working script (well - needs to be modified to work with the 49) which builds the library. However it's csh, so you need to have a working *nix system or the cygwin tools. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143329 Can someone sort of give me the steps to how to write a program? Like I am at the command line when you turn it on. Where do I go from there? Also I want to make a program that would take an equation like x + y = a and it will ask for the variables. And whichever one I don't input, will solve for. Is this difficult to do? Any help as well as pointing me to any text files or websites on how to do this would be greatly appreciated. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143332 Strangly this is covered pretty well in the manaul. Use the NUM.SLV key (RS 7) select Solve Equation and that brings up an input form. Fill in the equation and variable values and initial guess, highlight the value to solve for then press F6. news:9tch6m$5kl$1@news.service.uci.edu... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143343 Regulars, bear with me, but I have to repeat myself: I *love* the built in SOLVR (30 MENU) to do this. You don't need to program one, there are some solvers built into the HP49. Search hpcalc.org for Cleansolver. That eases (IMO :-) the use of the SOLVR and also includes a description how to use it (which the HP49 manual doesn't have!) BTW, my sister asked me to write some programs for her stat classes (as I had done before) and when I saw what she asked me to do (calculating with sums), I just told her how to use the SOLVR and she never would want to miss the SOLVR now and even wonders how she could have lived without it ;-) Greetings from Cologne Peter -- Great HP48/49 links: http://www.hpcalc.org http://move.to/hpkb to find *old* postings search: http://groups.google.com/advanced_group_search Xref: newsmst01.news.prodigy.com comp.sys.hp48:143364 Hi Brandon! If I understood your first question, you want the command line to be active when you turn the calculator on (?). To do this, you just have to press a number key then backspace to delete the number and then turn the calculator off. When you turn it on again, you have the command line activated. Now about the program for inputing values and solving. Well, you don't need to program the input because the HP49G has it already built in. Just enter: << STEQ 30.01 MENU Store it in variable 'SOLVEQ' or something similar. Press VAR to go to the variables menu, enter your equation and press the soft menu button labeled SOLVE. This program takes the equation as argument and presents you a menu with all variables. Enter a number and press a soft menu key. The value will be stored in the variable. When you have entered all knowns, press blue-shift and then the soft menu key for the variable that you want to solve for. The HP49G will the use the known values to solve for that variable. Greetings, Nick. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143384 Nick, Hey that's a good one for RPN mode (knew it at one time but had since forgotten about it) but it's pretty scary in Algebraic mode. I could get it to get to the solvr in algebraic if I pressed solveq and added ('equation') but once in the solvr I could not figure out how to assign variables and when trying left shift variable it goes to the plot window :) In it's defence though it did plot the test equation rather nicely :) 2 good ways in algebraic, I'm sure there are others either left shift (blue) and S.slv (7) solve ('equation' , variable to solve for) or right shift (red) and num.slv (7) and solve equation, this will allow you to enter or use equation writer for the eq and keep track of your variables allowing you to enter whatever ones you like and solve for the remaining. On 20 Nov 2001 06:29:09 -0800, nk@imos-consulting.com (Nick Karagiaouroglou) wrote: Xref: newsmst01.news.prodigy.com comp.sys.hp48:143340 Hi, I need the SPANISH manual of the hewlwett packard 12c calculator Could you tell me where can I find it? Thanks SA Xref: newsmst01.news.prodigy.com comp.sys.hp48:143350 SA writes: I think it's not available on-line or as a pdf. And it's not in the MoHPC CD set (at least as of version 3.10). Keep in mind that the HP-12C is still on sale. You can post an ad to: http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/adforum.cgi If you have any specific questions about the HP-12C, I guess you can post them here or to these other forums: http://www.hpmuseum.org/cgi-sys/cgiwrap/hpmuseum/forum.cgi http://pub8.ezboard.com/bhpcccalculatorclub.html Hope this helps. Regards, Bye. Jordi Hidalgo johil@tv3mail.com - Xref: newsmst01.news.prodigy.com comp.sys.hp48:143378 Dear HP 48 fans: I need an old HP 48 SX to run Longfloat. I have an HP 42 to trade (no, I don't have the manual). Does anyone want to trade straight across? Both parties to pay their own shipping. Please reply to my e-mail address below. Thanks. First reply gets it! amccornack@home.com Xref: newsmst01.news.prodigy.com comp.sys.hp48:143334 All the new HP graphical calculators are now produced in China. Regards Jean-Yves news:9tbsm7$3vn$1@hecate.umd.edu... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143328 I'm also an engineer. Since you already prefer RPN the choice is HP32 or HP49. I find the greatest benefit of the HP49 is its ease of programing and file management. It's so easy and natural to store UserRPL and Equations that I use them all the time even for minor tasks. Other major benefits to me are its list operations, customization, libraries, solvers and matrix operation. If I was just after a calculator for adding and multiplying etc I'd get a HP32 for it's better keyboard. I'm thinking of getting one for a spare and for it's convenient size. news:1006205448.25614.1@eurus... than know Xref: newsmst01.news.prodigy.com comp.sys.hp48:143330 Come on, I don't flame people! :-) On a more serious note, if I am doing so, please let me know. Bhuvanesh. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143355 You might also want to check hpcalc.org's 49G engineering libraries in an emulator. Date: Tue, 20 Nov 2001 14:20:46 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143358 Bhuvanesh wrote: No, you don't. I apologize for that one. And I like your posts - it's always good to also have another view, especially in such religious questions as Which calculator do I need?. But what I wanted to say: My opinion I tend to have about the 89, and to a lesser extent about the 92+, is *worth* being flamed. Really. =) Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Date: Tue, 20 Nov 2001 22:03:50 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143380 I wrote: On a second thought, deserves would've been more appropriate. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143327 :-) It's not just the crashes... it's also possible (although I haven't seen my programs doing this) that assembly programs can cause built-in functions like solve() to return wrong answers. Since I use my calculator at work too, I don't want that happening. Besides, I have to write the documentation (I don't mean the Catalog help) on a computer anyway, and it's easier to capture screenshots on a computer (in my opinion). I also often need to refer to the online TIGCC documentation to make sure I'm using the right function and tag names. Oh, of course... I didn't mean it in that sense. I could just reset the RAM, but my confidence in answers would be lower until I cleared everything and restored it from a backup. Maybe I'm being overly cautious... Bhuvanesh. Date: Tue, 20 Nov 2001 14:02:43 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143356 Bhuvanesh wrote: Can it? Well it's a different thing then. I wouldn't want my calculator to do its job wrong either :-/ Thanks to SDIAG, I have some documentation with me and can hack away on the calculator (e.g. during dull chemistry classes...). Of course I write the documentation of my own programs on the computer - IF they're ever released. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Yes, but they realized that the people were plotting bode plots, nyquist, nichols, etc. And the program bode-plot (from apolonio) was a bit slow, so it couldn't be turned off before the teacher saw the graphic... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143381 In my circuit analysis exam (requiring a number of complex matrix solutions) I asked my teacher wether it was OK to use my 15C and he did not care. In the end I had enough time to work the answers by hand and use the complex matrix solve to check the numbers. Realy examiners should be astute enough to write examination material that makes a programmable calc irrelevant - however I think most examiners are just plain lazy. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143385 Is the num.slv only able to solve linear systems with one unique solution? Whenever I try a system with more than one solution, IE a solution in terms of it's other variables it gives me a no solution condition. Ex. x+y+z+w=0 x+w=0 x+2y+z = 0 Returns all zeros, when it should return something similar to x= -r y=r z = -r w = r or in terms of one of the others as linsolve does. It got even more frightning when I tried to give it something like x+y = a y= b x+z = c It would not even take a, b, c into the matrix either as a, b,c or 'a' 'b' 'c' Just curious if I'm missing a big part of it's functionality or if it's that limited Todd Xref: newsmst01.news.prodigy.com comp.sys.hp48:143342 Using the builtin Editor to edit programs which are not pure UserRPL may potentially hang you calculator. Furthermore, you will not be able to recompile the edited version correctly. or use RPLED. - Carsten Xref: newsmst01.news.prodigy.com comp.sys.hp48:143348 news:qzuelmthfjs.fsf@sand.science.uva.nl... That's ok, I use Emacs for my SysRPL needs :) Xref: newsmst01.news.prodigy.com comp.sys.hp48:143352 And yet it works. Are the menus cached, then? Xref: newsmst01.news.prodigy.com comp.sys.hp48:143363 Julian Fondren wrote: only the last one when you do RCLMENU in the ~ menu you get 0.01 0 because the menu has no number and .01 because it is the first page to MENU this is seen as the first page of menu 0 and menu 0 is the previous menu. In other words; it will only work if the last menu was ~ -- This message was written with 100% recycled electrons Pivo Xref: newsmst01.news.prodigy.com comp.sys.hp48:143367 The menu of the BZ-toggler x~ from OT49 is an example of an *application menu in a certain environement*. These have no numbers. There are accessible and inaccessible application menus. Accessible means here to be able to isolate it. For instance, the 3-page application menu of the Interactive stack (starting with ECHO and ending with LEVEL) is inaccessible because it is used in a rather involved POL (Parametrized Outer Loop), namely the Interactive stack. It can be isolated only by experienced hackers and its use outside the POL is very dangerous. In contrast to this, the application menu of x~ is accessible and can be used independently on x~. The only problem is to extract it somehow from the rather complex program of x~ :-) This is very easy and needs no hacker tools at all if the OT49 or Keyman tools are not counted as hacker tools (which I do not). You can just produce a copy of the menu, to set it as a CST menu or a TMENU. How to do this? Just bind the OT49-command S?UK provisorially to the TAN-key, say. Next ser the x~ appl-menu by doing 1 x~, say. Purge the result, press TAN and you are asked S?UKey Press! So press the fist menu key and the SysRPL program of UBZi is on the stack. Write the name 'UBZi' on the stack and make a 2-list from the name and its program. Repeat this for the other 5 menu option. Next, make a list from the six stack objects and store it in 'BZM', say. Finally, store the program Instead of S?UK you can also use SA from Keyman. The easiesr way is ASSIGN THE RIGHTSHIFT-HOLD SOFT KEYS WITH ASMs FROM KEYMAN! This is what I urgendly recommend anyway, independently on the current discussion. Hope this helps Wolfgang Hi Tom! I tried it exactly as you posted. But it works perfectly on my 49, and I think it should work on the 48GX exactly the same way. (With skipping line you mean that this line is not displayed, when the calc waits for input ?) Greetings, Nick. is see Isn't it possible to create an aplet using the ADK for the HP38, then convert it with the aplet converter for the HP39G and finally send it to the calculator without the use of an HP38G ?? I would find it surprising if it wasn't Nevertheless, it's possible (and quite easy) to create an aplet able to recognize the HP49G graphical (with greyscale) and display them on the HP39G... You can already use the greyscale engine of the HP39G when playing with my game aplet Jean-Yves Xref: newsmst01.news.prodigy.com comp.sys.hp48:143331 news:3BF917ED.F0996203@iinet.net.au... there later command to The EDIT button does nothing, it does not start PAINT but if I have PAINT running, it tells me to shut down PAINT. Weird. The CAPTURE does not capture the image from the 39 either, although I can screen capture with the Conn. kit and save the image as a .bmp. I am running XP and am wondering if that might be the problem. cheers /Victor Xref: newsmst01.news.prodigy.com comp.sys.hp48:143333 Ok I can create aplets, convert them to 39 format, and transfer them to the calc. I had to use Win ME instead. The images I want to use are larger than default (131x64) so I cannot import them anyway. Thanks once again. /Victor news:3bf9dc7a$0$10231$afc38c87@news.optusnet.com.au... it Xref: newsmst01.news.prodigy.com comp.sys.hp48:143337 Look at : http://www.hp-sources.com/cours/xnviewconv/ (in French) news:MUhK7.56$Lf2.532@newsfeeds.bigpond.com... the problem the playing Xref: newsmst01.news.prodigy.com comp.sys.hp48:143353 Jean-Yves Avenard wrote: Umm.... Yes, you're quite right that would be the simpler way. Create an aplet on for the 38G, edit the image with the ADK and then convert it to a 39G aplet with your utility. Forgot about that ;-( Well, that's true but us mere humans who don't program at sRPL level on the 38/39G can't do it that way! Anyway, the point's not worth following up since Victor apparently can't fit the image he want's to use into the available space. Well, it was interesting while it lasted..... -- Colin Croft Old mathematicians never die; they just lose some of their functions. ====================================== Applications in Mathematics ccroft@iinet.net.au http://members.iinet.net.au/~ccroft/ ====================================== Xref: newsmst01.news.prodigy.com comp.sys.hp48:143326 Hello news:ckfhvtkgf0a0gse7rkdvjjqvpsm7me8j84@4ax.com... There seems to be a misunderstanding here from the beginning. I disagree again on your point as I'm NOT talking about encoding floated point number, but encoding integers (that would be the ZINT or EXTENDED REALS on the HP49) where you do not give a damn about how precise they can be, the fractional part is always zero. Expressing the number 12345678 will take 32 bits in BCD, but will only take 24 bits in binary mode. Adding numbers in binary mode, or just working with them would have been much faster on the HP49 and would have required less memory (as I wrote previously, ALG48 and Erable 3, were both working with binary encoded integers). I asked Mika Heiskanen to rework all its ALG48 part to work in BCD, as I was more concern by the time needed to display such a long integers (probably the most common user object in the HP49 now), than by the time that was required to perform the calculation. So if we gained in speed for the display of such numbers, it is slower to perform calculations, and also it adds some limitation on how big a number can be. In particular, Alg48 was optimized to work on 64 bits at a time, and obviously, you fit more digits in a 64 bits register in binary than in BCD. fact Not at all. Only operation like register+1 or register-1 will work in BCD mode. everything else will work in hexa mode, no matter what mode have been set-up If you do: LC(5) 9 LA(5) 9 SETDEC C+2.A A+1.A SETHEX C is now equal to B and not 11 while A is equal to 10 (as you could expect) So as I wrote before, the BCD mode on the Saturn is not really great regarding instruction set But that's exactly what I mean, ShellOS was fun, but not serious enough for a commercial application. Having to scan which instruction is about to be run, save the registers, run the instruction and restore the context will make it less say... 50 times slower than a hardware solution ? that's just a guess of course. I wonder why you keep arguing on this matter, we've been through this over and over. And I've tried both solutions Jean-Yves Xref: newsmst01.news.prodigy.com comp.sys.hp48:143335 Hello Jean-Yves! You *must* be kidding... from SASM.DOC: 2.7 Arithmetic Mode The arithmetic mode is set by the SETHEX and SETDEC instructions. When SETHEX is executed, the arithmetic mode is set so that all register arithmetic is performed in hexadecimal mode. When SETDEC is executed, the arithmetic mode is set so that most register arithmetic is performed in decimal mode. The following instructions are *always* performed in hexadecimal mode, regardless of the arithmetic mode setting: P=P+1 P=P-1 C+P+1 D0=D0+ n D0=D0- n D1=D1+ n D1=D1- n A=A+CON fs,n A=A-CON fs,n B=B+CON fs,n B=B-CON fs,n C=C+CON fs,n C=C-CON fs,n D=D+CON fs,n D=D-CON fs,n It makes sense that addresses are always treated as HEX data, and the P register can take 16 values. (idem for the constant n that is added with the r=r+CON instructions). But all the other arithmetic instructions honour the mode. (not just r=r+1 and r=r-1, but r=r+s, r=r-s, r=s-r, r=r+r, r=-r, r=-r-1). How else would the coding of a BCD multiply have been possible?? (of course, part of the confusion arises from the difference between A+1.A and A+2.A which are two different instructions (A=A+1 A and A=A+CON A,2 in JAZZ syntax) - there does not seem to be a way to say A=A+CON A,1 in MASD?) Werner Huysegoms xwerner_huysegoms@freegates.bex (delete the leading and trailing x) Xref: newsmst01.news.prodigy.com comp.sys.hp48:143344 On Tue, 20 Nov 2001 02:28:44 +1100, Jean-Yves Avenard Oops, sorry. ;) It would have been slightly faster due to the fact that you'd have less data to process, but much faster? In BCD mode you can work on 64-bits at a time with the same arithmetic instructions as binary so there is no speed penalty there. As Werner pointed out from SASM.DOC this is because r=r+CON fs, n instructions *always* perform HEX arithmetic. Most of the other arithmetic instructions behave as expected in BCD mode. If you don't believe me try it for yourself : LC(5) 9 LA(5) 2 SETDEC C=C+A A SETHEX C equals #11 after the addition. What makes it not serious enough? Java, Jazz, ALG48, Erable etc. started out as fun and many either became commercial applications or ideas/code were incorporated from them into such applications. How about your own Meta Kernel ? For the life of me I have no idea what you are talking about. I don't know how you define it but I define preemptive multitasking as the repeated asynchronous interruption and switching of processes after a predetermined time slice. Assuming a (relatively) very short time slice of 1/64 second and a clock speed of 3.92 MHz then even if the context switch took 1000 clock cycles it would only incure a 1.6% performance hit. Most other OS's have larger time slices ( eg. 1/10 second of some Unix's ) which would make the performance penalty even smaller. ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143345 wrote: There is no problem with MS calculator because it doesn't use binary floating point :). To quote the MS calculator help page Calculator also stores rational numbers as fractions to retain accuracy. For example, 1/3 is stored as 1/3, rather than .333 . As has been stated many times BCD only has an advantage when you are working with numbers of the form 2^n*5^m (n and m can be negative of positive integers) and with the arithmetic operations of addition, subtraction, multiplication and division. In these cases BCD is an exact representation which is good because we (usually) all use decimal calculators. :) I'd rather my calculator tell me the truth instead of deceiving me about the value of a number :). ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Well, having the ability to load data into multiple registers in one swath and a hardware multiply are the real reasons for the 68K advantage ( which is huge ) in this case. The Saturn can perform the additions faster and easier though. The Saturn supports 1-bit and 4-bit shifts. If you do the calculations it turns out that the 68K is not much faster with regard to performing multiple-bit shifts on standard data sizes such as 32 bits. And doing multiple-bit shifts on a 64-bit quantity is not easy at all. One method of achieving this might be ( assuming the shift count is in D0 and the most and least significant long words are in D1 and D2 respectively ): LSR.L D0, D2 MOVEQ #0, D3 BSET.L D0, D3 SUBQ.L #1, D3 AND.L D1, D3 ROR.L D0, D3 OR.L D3, D2 LSR.L D0, D1 A little more cumbersome and time consuming than on the Saturn :). Xref: newsmst01.news.prodigy.com comp.sys.hp48:143347 wrote: So all these huge companies that use it for financial calculations and all the companies whose calculators use it are just following a tradition? If BCD has no advantages there exists a huge and readily available base of software that could take its place. This makes absolutely no sense whatsoever. Please go and read the link. I think the IBM UK Laboratories, the IEEE, the European Union, countless researchers and many others know what they're talking about :) . To quote http://www2.hursley.ibm.com/decimal/decifaq.html : binary floating-point cannot be used for financial calculations, or indeed for any calculations where the results achieved are required to match those which might be calculated by hand . Yes there is : because so many people use it :). Both are capable of 8-bit packed BCD arithmetic. I don't consider Micro$oft Visual C++ ( or anything Micro$oft by that matter ) standard by any measure of the word. :) There are widely available free BCD math libraries. If you consider financial calculations and simple mundane arithmetic exotic then you are right. :) Try [0.1 + 0.9] where [] is the greatest integer function. Or how about x_(n+1) = 4*x_n*(1-x_n) where x_0 = 0.1 . After a few repetitions the values begin to rapidly diverge. This is an example of how relatively few repeated multiplications and additions/subtractions and quickly amplify even the smallest error. Binary if you mean the numerators and denominators of the fractions it calculates with are binary. :) ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Sender: eric@ruckus.brouhaha.com Newsgroups: comp.sys.hp48 Subject: Re: Well, it's over then...Good Bye. Organization: Eric Conspiracy Secret Labs Xref: newsmst01.news.prodigy.com comp.sys.hp48:143374 That may be the only time when it has a technical advantage. However, most people consider it an advantage when their calculator gives the same results they would get by working the problem manually, i.e., in decimal rather than binary. To someone thinking that way, it doesn't matter if binary is better in some other regard. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143382 Eric Smith wrote: That's exactly the point I tried to make in my first post to this thread. If I use my calc to perform decimal calculations, as most ppl do, then it should correctly implement that representation. Binary floating point fails to do that, but it is easier and less costly to implement in hardware, that's why today's digital computers have it. Regarding efficiency of representation: Binary might facilitate numerical error analysis, but it's not the most efficient system, at least not in terms of radix times width cost. See, e.g.: http://www.sigmaxi.org/amsci/amsci/issues/comsci01/compsci2001-11.html -- I especially like the balanced ternary representation. :) Cheers, Albert -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag Xref: newsmst01.news.prodigy.com comp.sys.hp48:143383 Jacek Marchel wrote: Oops, sorry, I think I just formatted my Windows partition. ;-) -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag And how is that different to what I wrote ? I was referring to A+2 and such ... Jean-Yves Well, I'm talking about facts here, ALG48 on big numbers can be over 40% faster than the HP49 Thank you very much for teaching me the joy of the Saturn Xref: newsmst01.news.prodigy.com comp.sys.hp48:143359 Are the doublequotes here supposed to denote irony? (and ick -- why silver? The 49G is pretty enough.) Xref: newsmst01.news.prodigy.com comp.sys.hp48:143368 Maybe certain ideas were a bit ... facetious (except in my personal ideal world, through even there I had to stop short of suggesting a GPRS tranceiver), but certain things I was serious about. The picture I had in mind was of something physically small, but of significant processing power that wasn't just a souped-up PDA with a copy of some maths software on it. OGG, MP3 ... it's the idea that counts. Certain individuals may require an appliance capable of producing or recording sound data. I drew a picture of what I was thinking of the other day. I envisaged a screen with an aspect ratio of 3:1 as a means of allowing a simple, multitasking user interface: menu on the bottom; 'windows' occupy either the whole screen, or either the left or right half. This ratio would allow for a screen on the top half of the device, with a fold-away keyboard below, all set in a metalling sky-blue, much like the current HP 49 (and no, not with changeable covers). To say QWERTY doesn't necessarily imply anti-calculator; many people type on computers, and as such are familiar with this configuration. Keys could still be overloaded. And I don't see what's so bad about a touch-screen; I would find it very useful for manipulating graphics and equations in real-time. As for 3D acceleration, well, yes, that was silly; most portable CPUs these days are more than well-equipped when it comes to rendering a scene to an acceptable framerate on a small screen of the size I was talking about. I, for one, would like fast, portable, good-looking 3D graphics, as opposed to the difficult wireframe displays on offer at present. Of course, Li+ is better, but also more expensive. Many mobile phones still ship with NiMH batteries and provide adequate performance. Anyhow, that's a somewhat minor issue for me; users would change battery types to suit their needs. The suggestion of Linux was only as an example, and nor did I in any way imply that it was the only option or open-source operating system. Ideally it'd have an OS tailored to its application: something adept with dealing with [computer representations of mathematical] objects, and that allows them to plot a graph while performing a numerical integration (for example). Hmm. I really like RPL. Normal DRAM would be slower than SDRAM. I'm not an expert on memory technology; the idea was to have a low-power, lower-speed backing store that acted like a hard-drive (or flash ROM, but faster). I wouldn't want C++ as the only option for a mathematical system. I'd much prefer a high-level extension to RPL (e.g. some CLASS ... END syntax) that allows better object orientation; I don't think I'd require the sophisitication of C++ (although to include compilers either with the system itself or as part of a PC-based development kit would be wonderful). I think a PS/2 keyboard connector would be physically too big. A USB keyboard would be much more desirable. I did even consider mice and trackballs, but decided a pointer would just overcomplicate things. TCP/IP networking would be, I suppose, a function of system software. Here I was thinking in the broader terms of a portable mathematics system: you never know when you might need to download and install a new package. At the end of the day, it needs to be expandable and futureproof. J. H. Ettle. charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-Priority: 3 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143373 That is actually viable. With the relatively large display, GPRS would be nice. The harwdare for this do not use up much space, but battery power instead. power I agree. That will be an option - some form(s) or (an)other. the a That could be one configuration. The current design allow us to change things like display and keyboard without modifying anything but around 200 lines of C++ code. That will probably not even be necessary in a few months (we're still not even close to settled on the modularity concept). Touchscreen will probably be an option later on, but at first we won't do it - it implies alot more work in regard to character recognition and so on. It won't be a problem to incorporate a touchscreen at some point. these to Grayscale 3D? How - an OpenGL implementation of some sort? still a their I'm not sure it would be that easy for a user to change the batteries. It's not in the design now anyway. example). A specific multithreaded OO OS is on the drawing board at the moment. Me too :-) that I agree that DRAM of some sort (preferably SDRAM) will be a viable option. The only problem is the power consumption, but one can't have it all anyway. The amount of Flash, SRAM & SDRAM should be defined by the user when he or she buys the device. that system I agree. A USB port will probably be default. I the You would want TCP/IP for many reasons, and I surely believe (still) that it should be an option. Regards Steen == is there anyway to get the calculator to swap the first two in the stack without having to press enter. ie say i wanted to do 8 / 5 but i pressed 5 first. that additional keystroke is annoying, especially cos the enter is in the wrong place. thankyou. sv -- You should quit while I'm ahead. Date: Thu, 22 Nov 2001 19:30:29 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143511 der d212mon wrote: Right arrow. You might want to read Tim's 20 Essential Things document, it's got a long list of useful key shortcuts. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143516 Or use the built in one! Press right arrow on both the 48 and the 49 Dennis PS VPN: Forgot to drink his morning coffee before hitting the SEND key :) Xref: newsmst01.news.prodigy.com comp.sys.hp48:143519 I'm getting my HP 49G case engraved to stand SWAP above the appropiate key I also consider having XSERVER and SERVER on it but there is no more room AND why not add GRAPH or PICTURE above the left arrow... You're right on this one Dennis (as always) I'm out of Coke at the moment and therefore signing out... VPN, zzzzz.... news:20011122135502.10051.00000970@mb-fa.aol.com... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143526 yeah but i still have to press enter after the 8. i'll run threw it again: 5/9 32S 9 enter 5 SWAP Divide on a 49G 9 enter 5 enter (This is the keystroke i'm trying to eliminate) swap divide -- You should quit while I'm ahead. news:9tjfve$ogd$1@news.kolumbus.fi... the Xref: newsmst01.news.prodigy.com comp.sys.hp48:143528 You cant eliminate this keystroke on either the 48 or the 49. This is because you have editing capabilities not available on the wimpy 32S ;) Dennis Date: Thu, 22 Nov 2001 22:41:56 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143531 der d212mon wrote: If you're really after that one keystroke, you need to assign both ENTER and SWAP to the same key. SA from Keyman reveals that ENTER is :: PTR 2BF9E ( NoEdit?case ) :: (...) ; PTR 1B827 ( parse command line ) ; PTR 1B827 is the one we need. Then with another SA it turns out that Right-Arrow is :: PTR 2BF9E ( dito ) :: PTR 2BFCB xSWAP DoBadKey ; (...) ; Now lets combine the two. You need something like this: :: EditLExists? IT PTR 1B827 ( ENTER first if there's an EL ) PTR 2BFCB ( SWAP ) xSWAP DoBadKey ; Bind it to a key, it should work (I've tested it =). However there is something you should note: The addresses above are for ROM version 1.19-6. They're probably different in previous versions (though you can figure them out yourself with Keyman), and they'll most likely be different in newer yet-to-come (hopefully) versions. If you do not have a clue what the above code means, you should first get Emacs and extable from www.hpcalc.org and install them on your calculator. That's all you need to *compile* the assignment. Keyman, OT49 and Nosy are also very useful. If you want to know how it actually *works*, you can read Programming in System RPL, avaliable at hpcalc. HTH Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143496 yeah! There is nothin wgrong in the new ROM in this section BUT where are the new versions of the Graphics Commands working on the GREY's too???? Can you still do this JYA? This (native GRAY in the User level) make my HP 49G 1.19-7 to dance around the TI 89 :-D Veli-Pekka news:6hMK7.10597$Lf2.13173@newsfeeds.bigpond.com... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143520 Your Hp49 will dance around the TI89 when your Saturn gets to run three times it actual speed!. :-) BTW stop flaming Bhuvanesh. Unlike me, he hasn't made a direct comment to your calc... 9tjbvj$ihf$1@news.kolumbus.fi... but restore Xref: newsmst01.news.prodigy.com comp.sys.hp48:143472 OK, took it! Yep, done. Aha! Now we have the same results. :-) So it was only me misunderstanding your labor journal. ;-) Yes, I see. Perhaps that is why some people don't want keys but touch screen? Perhaps modelling keys behavoir could be done with statistical thermodynamics? ;-) Greetings, Nick. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143494 news:662e00ed.0111191943.2db05e3c@posting.google.com... GRRRR! When you're flaming my calculator you're flaming ME !!!! VPN PS: Buy a HP 49G, there's a lot of critic, but the build-in EQW beats the TI 89 add-on EQW Hah! Now there's a flame for you lalu_bhatt !! Xref: newsmst01.news.prodigy.com comp.sys.hp48:143499 I know it's not new! (AT least now I know ;) I'm just BUGging dear Monsieur Avenard to fix *ALL* the bugs in the upcoming 2002 release ROM 1.19-7 ;-) VPN, a v'ger fanatic !!! news:3bfa5747.243015@news.rdc1.az.home.com... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143508 news:3bfc9919.12447432@news.dcn.davis.ca.us... After reading this I had a terrible Anti-Deja-Vu from the future: Nokia is hiring Carly Fiorina as a new CEO after she got kicked out of from th HP in 2002 AAARRRGGGHHH! No more Nokia quality What a day-mare! Veli-Pekka Xref: newsmst01.news.prodigy.com comp.sys.hp48:143473 Hi Colin! Great proposal! For 2 reasons: 1) Much shorter/faster than mine. Greetings, Nick. Date: Thu, 22 Nov 2001 13:25:28 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143480 Timothy Ney wrote: Well if you wanted to create a bunch of 5-nibble variables in the SavMisc area (on the 49: 128 nibbles of memory at #8221D), you could do ABASE =SavMisc var1 ALLOC 5 var2 ALLOC 5 var3 ALLOC 5 etc. This way, var1 is at SavMisc, var2 at SavMisc+5, and var3 at SavMisc+10. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Thank you for that Eric. Now I see the light (light Coke on my table): BINARY *coded* decimals. A pure decimal machine ofcourse - becides 10 voltage levels - needs Decimal coded binaty mode ???! That's DCB Have fun VPN Xref: newsmst01.news.prodigy.com comp.sys.hp48:143513 In Eduardo Kalinowski's hp49 manual he states that the code string must end with a special character, which I cannot read or decifer from the acrobat reader file. What is the character and how do I get it using the built in editor. Date: Thu, 22 Nov 2001 19:56:00 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143517 Glen T wrote: The at character: @ (ALPHA RS ENTER). If you want to program on the calculator, you should really get Emacs. It does much more than just typing the @ for you =) Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143467 Hi, HP48 users! Several months ago I built a data logger for HP48. Since there is no availability of A/D converter to RS232 I had to build a UART (a parallel to serial converter), and to use a parallel A/D. Today some of my friends (a TI89 user) showed me a very simple project (more so then mine) that does the same thing. The site address is http://www.dvz.fh-koeln.de/~b012414/osci89/osci.html I wonder how to control the recieve bit in the HP48 as an input bit without using the RS232 protocol, (the transmit bit controlled by the OPENIO and CLOSEIO comands. I would be very thankful if you could help me and I hope its not too difficult of a task for you. Best regards, Tal. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143485 ----- Original Message ----- Newsgroups: comp.sys.hp48 Sent: Wednesday, November 21, 2001 10:38 PM Subject: Re: EMU49ASC question ??? I don't know what do you mean by secondary. I got the binary integers list from my HP49G and copied it to my Emu, then STOF it. This is my RCLF: { # 3494797178661045232d # 0d # 9230127453643964840d # 1048576d } As I said, I tried once and it worked. Then tried again, some time later, and the Emu processed the (ascii) program like it were a string: it just stripped away the CR, but didn't make a (binary) program. I'm using Emu48 with a 49G KML. I can't understand wht's going wrong. The Emu48Asc doc doesn't tell a thing about configuration, so I really don't know. Thanks, ============================================================ Paolo Cavallo I am a teacher at heart, and there are moments in the classroom when I can hardly hold the joy. P. J. Palmer, 1998 paolo.cavallo@iperbole.bologna.it http://www.alberghetti.it/paolo.cavallo/pc.htm ============================================================ EMERGENCY - UNO STRACCIO DI PACE - www.emergency.it Xref: newsmst01.news.prodigy.com comp.sys.hp48:143487 thing Know I understand, feeling rather silly. I assumed I had to type <<, like when I have to upload ascii programs to my HP49G. I was wrong. ASC is looking for « characters. Now it works... PC Xref: newsmst01.news.prodigy.com comp.sys.hp48:143515 news:9tj3ud$l7s$1@pinco.nettuno.it... to If you can use << or must use the « depends on the setting of the IOPAR variable, and if the string has a %%HP: T(x); header infront, from the header. If the header is missing the translation mode is fetched from the content of the IOPAR variable (when it exists, if not, one is generated with T(1)) else IOPAR is first set to parameters of the header. For using most of the cc sequences, especially for <<, you need translation mode 3. The sequence %%HP: T(3); will work fine, where %%HP: T(1); fails. But you're right, in the fail situation the binary number in stack level 1 should be purged. Will be fixed soon. Cheers Christoph Xref: newsmst01.news.prodigy.com comp.sys.hp48:143469 shrinking on daily basis. Yes, I like the idea. But maybe it would cause too much double work when I post here. Maybe we should make a chain: I will send the list of unknown entries it volunteers who would go through it and provide information about the entries which they either know or are willing to check out. When I get the list back from the first person, I will send the entries which are left to the next person etc. Who is willing to take part in this chain? Excellent. Many thanks. - Carsten Date: Thu, 22 Nov 2001 13:10:49 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143479 Carsten Dominik wrote: Why not use diffs? As you have seen (hehe), patches will usually succeed even if the original file was modified in the meantime. OTOH this requires the diffutils. Anyway, I'll keep working on the ML section =) Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143502 Carsten Dominik wrote: Some additions about undocumented pointers: 1. What I said about GETXMAX PUTXMAX holds similarily for GETRES, GETXPOS, GETINDEP (gets the independent variable in PPAR) and the corresponding PUT-commands. GETPTYPE recalls the current plot-type. If no PPAR exist in the current directory, the default PPAR is created (slow, needs about 0.5 sec) whose plot-type is FUNCTION. 2. An important undocumented command is DOERASE which - in contrast to KILLGDISP (UsrRPL is just PICT PURGE) - does *not* alter the actual GBUFF size but does simply erase. The others minimize GBUFF the to (10 bytes) NULLGROB. IMHO, Eduardo's comment Clears graphic display is misleading. That's what DOERASE actually does. All hithertoo mentioned commands run indepently on the actual HARDBUFFER. 3. GETIOPAR quotes IOPAR. The name IOPAR itself is just the unsupported but stable pointer PTR 271A9 (2.5 versus 8.5 bytes). However, the command GetIOPAR puts the IOPAR entries as a meta onto the stack (maybe without count, I don't remember). There should be a section Input/Output containing all those undocumented commands on which I only know that JHM knows their functionality. 4. TakeOver should perhaps be listed in the new created section Menu commands. What I said here should be verified by a second person to make the information more reliable :-) - Wolfgang Xref: newsmst01.news.prodigy.com comp.sys.hp48:143514 How about CHECKPVARS? Is it documented? At the risk of restating the obvious: Requires no arguments. Recalls contents of PPAR in current path to stack. Creates PPAR in current directory if non-existent. Errors (INVALID PPAR) if existing PPAR is invalid. Dennis Xref: newsmst01.news.prodigy.com comp.sys.hp48:143484 Dear all, now I made a quick and dirty kind of a printserver for all HP49 owners, who have a HP48, and a HP82240B infrared printer, too. Basically you have two programs, a remote line printer daemon one running on the HP48, and a remote line printer one on the HP49. The communication is handled with Kermit protocol (Attention! don't forget to -35 CF on both calcs for pure ASCII transfer.) with a server on HP48 and a client on HP49. All is only a first attempt, but at least it works for me. Every correction and improvement is highly appreciated. HP 48 part, RLPD (remote line printer daemon): %%HP: T(3)A(R)F(.); << DO SERVER 'TMPOBJ' RCL DUP 0 TMPOBJ } PURGE END == UNTIL END HP 49 part, RLP (remote line printer): %%HP: T(3)A(R)F(.); << { TMPOBJ } SEND FINISH The HP49 RLP eats up one object from stack, sends it to the HP48, and shuts down the remote Kermit server. The HP48 RLPD fires up Kermit server again and again, until he receives the string END (subject to further refinement). The HP48 converts the received object into a GROB, and prints it on the HP82240B. but then it is harder to send an understandable end of transmission message to the HP48. Now some things remain which I don't like: 1) A global variable TMPOBJ is created on both calcs, which has to be purged thereafter. 2) The end of transmission procedure is very inelegantly, especially since the string END is printed out. Examples: http://home.tiscalinet.de/RKF/printout1.gif http://home.tiscalinet.de/RKF/printout2.gif Greetings from Hamburg, ------------------------------------------------------------------------- Ralf Fritzsch Bundesanstalt fuer Wasserbau Federal Waterways Engineering and Research Dienststelle Kueste Institute - Department Hamburg ------------------------------------------------------------------------- Unix _IS_ user friendly - it's just selective about who its friends are. ------------------------------------------------------------------------- Xref: newsmst01.news.prodigy.com comp.sys.hp48:143495 news:JcldVHe8EppF-pn2-m8kYgdWMUvFV@localhost... but then discivered a take The [ON] key in both of my HP-48GX were failing I got rid of the one performing worse (ofcourse I told the guy that it has that problem, but he insisted on having it anyway and gave a slightly broken HP 75C in the trade. The 75C works fine as long as the tapes holding it together will last) but now the one that I have left starts also failing the shift keys and alpha like on the other GX My HP 48SX work as fine as my HP 28S does!! Of my four 41CX's only one is working... :-( VPN PS: Buy a HP 49G !!! Xref: newsmst01.news.prodigy.com comp.sys.hp48:143464 Thomas Sader wrote: Try to apply ObFx from OT49 on your string on the stack. That should fix any tranfer error and result in a binary. In the worst case, the binary itself is corrupt. Then there is no other way as to purge the file and contact the author. Hope this helps Wolfgang Xref: newsmst01.news.prodigy.com comp.sys.hp48:143465 Wolfgang Rautenberg wrote: Well, one should add that if you try the fixed binary and it is corrupt it might cause a TTRM, so you better back up your HOME first. :) -- Dr. Albert Graf Email: Dr.Graef@t-online.de, ag@muwiinfa.geschichte.uni-mainz.de WWW: http://www.musikwissenschaft.uni-mainz.de/~ag Xref: newsmst01.news.prodigy.com comp.sys.hp48:143478 Hi Thomas! When downloading libs or other binary files for the 49 results in such strings, then it could be that you have made the download in ASCII mode. Change this to binary transfer. Greetings, Nick. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143518 ************************************************* put in the stack the library put the port number and press sto or using the files maneger move the library to port and after warm start Xref: newsmst01.news.prodigy.com comp.sys.hp48:143486 Hi, Could you share your knowledge on 128K RAM extension cards for HP48SX? I am interested in what manufacturers there are, what prices can one expect, where they can be bought in web and if you know any problems in using some cards. All information is wellcomed. I am considering of byuing such an extension. Ari Xref: newsmst01.news.prodigy.com comp.sys.hp48:143492 news:etmK7.200$Lf2.1214@newsfeeds.bigpond.com... in Thank's JYA For the confused one's I'd like to add that some companies have a huge pile of... stocked Indonesian made HP 49G waiting for a poor... customer to buy one. The software may be great (no doupt! JYA is involved...) but the HW is not up to any HP Standard until Chinese models were introduced. BUT, BUT I do not own (yet) a Chinese made HP 49G AND I'm still happy to use it !!! AGAIN: If you, Glen, are a choosy guy, which is ok look carefully about the serial number in the back of the 49G you're about to buy! If it says CH.... your're ok (I think you're ok anyway, but...) I hope this clarified things more than confused. VPN Xref: newsmst01.news.prodigy.com comp.sys.hp48:143493 Not from HP, I'm afraid! I have some friends through the dealer channel and even I can't get a replacement for a perfectly workiing calc. If your dealer is friendly enough, you try to trade it... BUT if it ain't broken why fix it?! I have two ID... and they are both working wonderfully!! Veli-Pekka news:9td11v$clo$1@news.tpi.pl... keys Xref: newsmst01.news.prodigy.com comp.sys.hp48:143500 When I am using the built in hp 49 text editor how do I get a new line? VPN news:a7a5de04.0111220945.27347e04@posting.google.com... Date: Thu, 22 Nov 2001 19:31:01 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143512 Glen T wrote: RS-. Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman -- S.Cocquereaumont Xref: newsmst01.news.prodigy.com comp.sys.hp48:143481 Stephane Cocquereaumont writes: Thanks to Eric Smith, they are available in the MoHPC CD set: http://www.hpmuseum.org/software/swcd.htm#intern Regards, Bye. Jordi Hidalgo johil@tv3mail.com Xref: newsmst01.news.prodigy.com comp.sys.hp48:143463 I have been watching with some interest the lay-offs of people in the calculator division. I quit working for HP in Roseville last year. Since then they have come thru roseville twice with an independent company firing people. Using only numbers that needed to go. They then had to hire back about 40% of them, realizing they let the wrong people go. Seems like it is a corporate wide virus. When the numbers dont match shit-can people, or buy out other companies fire their employees and use their assets. Wished more companies were run by women? Don Xref: newsmst01.news.prodigy.com comp.sys.hp48:143482 you have to understand the false economics that companies are running under. its all book keeping entries. it is the result of banking fiat. www.anti-matrix.net Xref: newsmst01.news.prodigy.com comp.sys.hp48:143533 on MY new Hp49g i got the function Y=Root(9+sq(X) but when i want to draw it it doesn't fully work. I draws just half a circle in the positive Y-axes area. I already set it in TABLSET that it schould start with -20 but it still doesn't work. I m sure it is just some stupid setting i forgot, but anyway help me. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143534 Well, by definition, that value can only be positive. -- Thierry Morissette tmorissette@hotmail.com Xns9162DDCCA6n2@195.3.96.116... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143470 Hi Pied! Well, all 48 models have no real CAS. So for such more complicated cases you better use the 49. Greetings, Nick. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143476 Hi, well, my hp48gx took a while (aproximately 2 minutes) but it finally gave an answer 4.25 Check your flags and try it again CdU-ETSIG gregorio@rpn48.zzn.com http://www.rpn48.zzn.com http://www.etsiig.uniovi.es/asociaciones/clubusu Xref: newsmst01.news.prodigy.com comp.sys.hp48:143483 Rodriguez Rodriguez, Gregorio wrote: My HP49 with ROM 1.19-5 used some 2 minutes and 10 seconds to solve this in approx mode, but only 8 seconds in exact mode. Isn't that strange? I thought the HP49 worked faster in approx. mode than exact mode. Or doesn't it? Comments anyone? Regards Ray Xref: newsmst01.news.prodigy.com comp.sys.hp48:143506 news:3BFCF56B.2C9F4677@spam.no... Exact Mode: use CAS first to give a simplified solution Approx Mode: use 12/15 digits of presision to find a numerical answer You can limit the accuracy to get a faster responce Use SCI 6 Just a (quick) taught.... VPN charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Priority: 3 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143525 The HP49G tries to find a symbolic antiderivative first, if a number of conditions is met. this is usually much much faster than merely plugging in numbers. The overhead is usually small, because when the calc does not find an antiderivative, the expression is very complex, thus taking very long to compute manually too. A thing I don't understand right now: which should be the antiderivative of 'ABS(X-1)^3'. If I do which is equivalent to 'ABS(X-1)^3'. So far so good, but if I do The latter should be '1/2', not 0. I have been sitting for a couple of hours now doing this by hand, and I can't seem to grasp what is wrong. So far, my papers say that '1/2=0', but my intuition tells me that is wrong. What is wrong with the antiderivative the HP49G returns? Neither 'CASINFO' nor 'PRIMIT' tells me much. It must be late... :-) Regards Steen Xref: newsmst01.news.prodigy.com comp.sys.hp48:143535 I believe there's nothing wrong with the antiderivative your 49G returns. In fact, it took me a while to get to the point, and I should admit I should have reviewed my maths before attempting to solve this problem.However, now i think I have my knowledge in full state, so here it goes: Have you tried to plot that antiderivative?? What you propose is similar to saying that because int(sin(x),0,pi)!=0, then sin(pi)!=0, I mean, it should be clear that the antiderivative has several zeros in x=1, x=2, x=0, but these doesn't mean that the integration between 0 and 2 should be 0. Hope I've made this a bit clear. Regards Gregorio Rodr222guez CdU-ETSIG gregorio@rpn48.zzn.com http://www.rpn48.zzn.com http://www.etsiig.uniovi.es/asociaciones/clubusu Xref: newsmst01.news.prodigy.com comp.sys.hp48:143474 Hello everybody [Bonjour 210 tous] You can found a french version of this message at the end.. [vous pouvez trouver une version fran215aise de ce message 210 la fin] English version [version anglaise] : ==================================== I'm pleased to announce you that I have proposed the creation of the group fr.comp.sys.calculatrices.hp. It is of course a french group for the HP calcs ;-) The group is not created yet, it is just the first step... In fact I don't have the time to translate all the message in english, and I think that interested peoples can read the french version. Moreover, all the discussions about this new group must be in french, so sorry if there is a big message in french and a little in english... Bye ! French version [Version fran215aise] : ==================================== Je vous annonce que j'ai propos216 la cr216ation du forum fr.comp.sys.calculatrices.hp. Cela permettrait de remplacer les forums hpcalc (sur news.zoy.org), tout en 216tant plus simple d'acc217s (puisque le forum serait accessible depuis la majorit216 des serveurs de news des fournisseurs d'acc217s ou des 216coles/universit216s) et plus s236re (puisqu'il n'y a pas qu'un seul serveur de news contenant les messages : ces derniers sont recopi216s sur tout un ensemble de serveurs, ce qui offre plus de s216curit216 et de souplesse). Cela a aussi quelques inconv216nients : Dans un premier temps, il faudra se contenter d'un seul forum generaliste au lieu des forums dev/calcul/general. S'il y a un r216el besoin on pourra demander la division du forum par la suite en divers th217mes, mais mieux vaut commencer par un seul. Il est 216galement impossible de poster des binaires dans la hierarchie fr.*, ce qu'on pouvait faire ici dans hpcalc.binaries. Il faudra alors passer par des adresses web ou ftp (prennez un compte sur free par exemple ;-) ) IMPORTANT : La cr216ation de forums dans la hierarchie fr.* ne se fait pas simplement : elle doit commencer par un Appel A Discussion (AAD). Il s'agit d'une proposition d'un utilisateur qui doit 220tre d216batue. Au cours du d216bat ca peut donner lieu 210 de nouvelles versions de l'AAD. Enfin si le d216bat est constructif et dans l'espace d'un mois suivant le premier AAD, les utilisateurs sont appell216s 210 voter (210 l'aide d'un Appel A Voter : AAV), l210 encore selon une demarche bien d216finie. Du resultat de ce vote d216pendra la cr216ation du forum (notez au passage qu'il faut au moins 80 votes). J'ai donc poster un Appel A Discussion concernant la cr216ation de fr.comp.sys.calculatrices.hp (il n'est donc pas encore question de voter), et il est tr217s important que vous participiez 210 la discussion (ne serrait-ce qu'en disant je suis d'accord) pour que le forum ait une chance d'220tre cr216216. Les discussions doivent avoir lieu sur fr.usenet.forms.evolution (c'est un peu le bordel mais cherchez le message !) et uniquement l210 !!! Je ne vous poste donc pas une copie du message que j'ai envoy216 afin de vous eviter la tentation d'y repondre ici, ce qui ne servirait a rien, mais je vous invite 210 aller le lire sur fr.usenet.forums.evolution Cl216ment Pillias (HpFool). Xref: newsmst01.news.prodigy.com comp.sys.hp48:143507 Buy a HP 49G !!! VPN news:b1f5ba59.0111211223.370bdc5f@posting.google.com... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143536 Hi, I forgot to mention the fantastic Filer48 from Raul Del Cid Lopez, which seems to be the best file manager for the HP-48 G series. It works as fast as the filer in the 49G or the MK filer, it has nearly the same features, and it's free, so you can save your money for more important things;-) Of course available on www.hpcalc.org Raymond news:b1f5ba59.0111211223.370bdc5f@posting.google.com... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143471 I have been using a 2Mb card from Klotz Electronic in my HP48GX for a couple months now. It seems to work well for me but some have suggested that this card may somehow damage my machine. Can anyone shed some light on this? Has anyone else suffered damage from these cards? Thank you, Stephen Xref: newsmst01.news.prodigy.com comp.sys.hp48:143524 With the hp49 if I use ck&dispatch0 one I can only input data types real with a decimas eg 1. not 1 as valid input how can I get it to accept either 45 or 45. and give correct results? Date: Thu, 22 Nov 2001 22:20:08 +0100 Reply-To: t.rast@freesurf.ch X-Mailer: Mozilla 4.74 [de] (X11; U; Linux 2.2.16 i586) X-Accept-Language: de-CH, de, en MIME-Version: 1.0 Newsgroups: comp.sys.hp48 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143527 Glen T wrote: Depends. If you want #45h (2: array 1: list), you can simply write #45 and MASD will figure out the correct BINT address for you. If you want #45d (#2Dh or 2: complex 1: tagged) you can use BINT45 (or, if there is no corresponding built-in BINT, #2D). Thomas -- Thomas Rast If you cannot convince them, t.rast@iname.com confuse them. ICQ# 103670088 -- Harry S. Truman Xref: newsmst01.news.prodigy.com comp.sys.hp48:143530 Use CK&Dispatch1 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143537 Hello news:a7a5de04.0111221300.325a51d6@posting.google.com... If you are using CK&DISPATCH0, you ask the system to make sure that no conversion will occur. If you want your program to still accept real AND zint, then you simply have to put in the dispatchee list the zint code: #FF It will be something like this: :: CK&DISPATCH0 ONE action_for_real ; or something probably easier: :: CKREAL CK&DISPATCH0 ONE action_for_real ; CKREAL will check the argument, if it's a zint will convert it into a real, if it's then not a real, it will error out. If you want to conversion to be made automatically, you have to use CK&DISPATCH1 or CK&DISPATCH2, then you won't have to worry about if it's a real or a zint. The only issue you may have is that a tagged object may be detagged. But it's more into the HP4X calculators philosophy Cheers Jean-Yves Xref: newsmst01.news.prodigy.com comp.sys.hp48:143540 I don't know under what condition the rouding errors you mentioned occurred, but they are not likely to happen on both TI89 and HP49. Under very special circumstances there might be slight discrepancies between, but not anything that should be seriously considered. In an article published in Scientific American back in '84 or '85 (How to handle numbers with thousands of digits, and why one might want to), Professor Fred Gruenberger, of California State University at Northridge, presented a way to check the accuracy of various computers and calculators then available. He would enter the number 1.0000001 and square it 27 times consecutively and compare the result with the exact answer, equivalent to 1.0000001^(2^27), that is, 674530.470741 . Doing this on the HP28, 48 and 49, we get 674514.86877 (percent error: 0.002313%). The equivalent operation on the TI-92 yields 674529.413051 (percent error: 0.0001568%). But this only means the TI-92 (or 89) has a more accurate floating point multiplication algorithm. One should test all other real functions in order to determine which calculator is more accurate. Just a curiosity: Percent errors for some (old) calculators and computer languages: TI SR-52: 0.00146% HP 33, 67, 41, 15C: 0,0050% Turbo Pascal 7.0 (real): 0.0091% Turbo Palcal 7.0 (extended): 0.00000071% Apple II Basic 96.631% !!! G. W. Barbosa Hi, Julian! I was talking about NATIVELY BUT thank you for this additional information! Xref: newsmst01.news.prodigy.com comp.sys.hp48:143489 news:SUhK7.58$Lf2.532@newsfeeds.bigpond.com... X for run But this fun! Keep onm counter-arguig Jean-Yves! There are always newcomers which will keep wondering about multitasking and other strange things while we old-timers have fun re-remembering these debats... Tervehdys Jussi-Iivarille ! Fr217re-Pierre Xref: newsmst01.news.prodigy.com comp.sys.hp48:143490 news:3BFC4B99.F294FC7@simul8.demon.co.uk... at binary ? What are the (Donals) Ducks covering? ;-) Is there a ultra low power version of Vax?? Do you have a direct URL into Vax-CPU references? Veli-Pekka Xref: newsmst01.news.prodigy.com comp.sys.hp48:143497 Jonathan Busby wrote: Sure, but here is my immediate counterexample: double x = 1.0; x /= 262144.0; x *= 262144.0; print(floor(x)); The system should produce 1.0 right ? Wrong if you use HP48 with BCD :-) It produces 0.0; What you do is to use recursion used by chaotic systems to magnify intial errors. The same ideology can be used to prove, that BCD by being thousand times less exact will produce garbage much faster than same memory size binary. Consider the following recursion: x = 1.0/3.0; x[n+1] = 10.0*x[n]-3.0; x should be always equal 1/3 right ? Here is the outcome: Step HP48 Binary double 1 0.33333333333 0.333333333333333 2 0.3333333333 0.333333333333331 3 0.333333333 0.333333333333315 4 0.33333333 0.333333333333148 5 0.3333333 0.333333333331483 6 0.333333 0.33333333331483 7 0.33333 0.333333333148296 8 0.3333 0.333333331482962 9 0.333 0.333333314829616 10 0.33 0.333333148296163 11 0.3 0.333331482961626 12 0.0 0.333314829616256 13 -3.0 (NONSENSE) 0.333148296162562 ... Binary holds answer at least correct at the first digit for the few more steps. I'm sure, one can easily find out other recursions that reinforce quickly initial error of BCD being incapable to propely show inversions or relatively small powers of 2 where binary is capable to show it correctly. Jack Xref: newsmst01.news.prodigy.com comp.sys.hp48:143503 And I have a blue screen and have to reformat all the hard-disk partitions. The NT40 was bit by the BIOS: A) BIOS was set-up for a different condiguartion of hard disks by an automated detection. B) the cable select option was used on both hard disks C) The NT incorrectly claimed to have cross-linked files and immediately continued to resolve this by automatically changing them in the hard disk D) everything seems to be screwed up BUT the old hard disk shows everyting right after a _new_ automatic hard disk detection of the drives fater the cable select was jumpered again to Master and Slave E) The moral of the story: Too much automation can destry your NT40 hard disk use Linux, it stated in Finland, the land of great programmers like the Nokia GSM guys (& girls), Mika Heiskanen, Kari Pasanen, Teppo Vuori, Tapani Tarvainen and: Veli-Pekka Nousiainen Veli-Pekka Nousiainen PS: a great programmer doesn't always imply a hardware guru :-( news:3BFAE8E3.7000500@t-online.de... Xref: newsmst01.news.prodigy.com comp.sys.hp48:143505 LOL !! Please, both of you, stop, you're killing me!! NO, NO, NO ! Don't stop! This is the most enjoyable way to go.... VPN news:d8dnvtcdnvb40fe9js5ijgr55vk43qm7vq@4ax.com... it's ----- Xref: newsmst01.news.prodigy.com comp.sys.hp48:143521 Here is a simple recursion, that binary carries without error (even with float size 24 bit mantissa of data, not double), and where BCD implemented on HP48 miserably fails: x[n+1] = ((x[n]/262144)*10)*262144-9 where x[1]=1.0 The outcome is as follows: Iteration BCD (HP48) Double binary Borland C++ builder 1 1.0 2 0.99999999999 1.0 3 0.99999999991 1.0 4 0.9999999991 1.0 5 0.999999991 1.0 6 0.99999990999 1.0 7 0.99999909989 1.0 8 0.99999099891 1.0 9 0.9999099891 1.0 10 0.999099891 1.0 11 0.99099890999 1.0 12 0.90998909991 1.0 13 0.09989099911 1.0 14 -8.0010900089 1.0 :-) Jack charset=iso-8859-1 Content-Transfer-Encoding: 7bit X-Priority: 3 Xref: newsmst01.news.prodigy.com comp.sys.hp48:143523 Hi Jonathan. I have watched this conversation in amusement, but I have to say something now - it's getting a bit out of hand. You should definetely be able to see clearly what I say now; OT, but Java is not known for absolute matematical precission either ;-) In ANY case when using more than 1 bit, FLOOR & CEIL should be used with uttermost care. There is an indefinite number of cases where any coding will fail (in the following, *packed* BCD is used of course). Let me explain: Let's say you want to have a discrete representation of the real numbers between 0 and ~1. If you decide to use 8 bits, you'll be able to represent these numbers: Binary: 0.00000000 0.00390625 0.00781250 0.01171875 ... 0.99609375 BCD: 0.01 0.02 0.03 0.04 ... 0.99 If you want to represent the number 0.625452452842, the closest value in binary representation would be 0.625, and in BCD it would be 0.63. The BCD representation is around 10 times less accurate. This is just one simple example, but it is obvious; In any given range of real numbers, represented by an evenly distributed miss rate. this means that for any given encoding, you'll have the same chance of a miss, but whereever you hit per random, the even distribution will make certain that your average miss is only half of the distance between two adjecant numbers in the set. The smaller that distance, the smaller any accumulated error will be. That is fact. Binary representation will always be much more accurate than BCD. At 64 bits, the binary distribution distance is 1/2^64, while the BCD distribution is 1/1E16. In this case binary representation is 1845 times more accurate than BCD. It'll always be that way of course. In any random calculation, you could end up with any rational number, and the more accurately you could represent that number, the smaller your final error will be. As to accumulated errors - you cannot determine the roundoff direction in any particular iteration, but you can guarentee, to a high degree, that it won't be the same direction all the time (by choosing the proper algorithm). Whenever roundoff will be the same direction, you'd minimize the accumulated error by making certain you had the smallest possible distance between adjecant numbers. This is obvious too. Binary representation will be more accurate than BCD in far the most cases in any application. What's a correct answer? The resault is more accurate when you're closer to the real result - right? You cannot expect to get '8' as reault all the time - not even on a financial calculator... BCD was invented at that point in time, where it took much more time to display a binary number in decimal representation, than there was overhead in BCD calculations. That time is by far over, and we now take advantage of the true precission of our registers. No reason to do otherwise. BCD is very rarely used today, and was almost solely used in for example cash registers, since they did not rely on complex arithmetic. The overhead when dealing with BCD multiplication and division is simply enormeous, which is also why most BCD capable processors internally convert to and from binary whenever doing BCD mulitiplication and division. Think about it, instead of citing age old books (no pun intended). Regards Steen Xref: newsmst01.news.prodigy.com comp.sys.hp48:143532 On Thu, 22 Nov 2001 21:22:09 +0100, Steen Schmidt Try telling that to a calculator user who only cares if the result is what they expect. Thank you for telling me why binary floating point is more accurate than BCD. Now I know how to do 2^64/10^16 :). Sigh. I never implied ANYWHERE that BCD was more precise and/or accurate than binary. It's a mystery to me why you keep reading that in. Let me restate the assertion : When using numeric data containing decimal fractions and performing the operations of addition, subtraction and multiplication *without regard to the delicate issues of round-off* binary floating point WILL NOT MATCH the results achieved via hand decimal arithmetic. What this means for the calculator user is that unintuitive and unexpected behavior can result. Why do you think almost every calculator to date uses BCD arithmetic when they could have more easily used binary? Yes you can. As long as additive and multiplicative arithmetic is used exclusively on decimal data and any calculations don't exhaust the precision provided by the BCD representation. Do a search on the web or visit the various links that have been posted and then tell me that again. Why don't you e-mail Mike Cowlishaw (http://www2.hursley.ibm.com/mfcsumm.htm mfc@uk.ibm.com ) the maintainer of the page on decimal arithmetic at the IBM UK Laboratories, at Hursley Park, Winchester, England. I'm sure you'll get a wholehearted response of agreement from him. Please, continue to live in your delusion while the rest of us move on. Also, keep me informed on the sales of your new calculator which will be using binary floating point in an interactive arithmetic environment :). ---------------------------------------------------------------------------- --- Remove the random permutation of NOSPAM from my e-mail address before replying. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143529 I tried to calculate the exact value of PI using both BCD and binary. No matter when I stopped, I got a rounding error :) DOH! Dennis Xref: newsmst01.news.prodigy.com comp.sys.hp48:143475 Yeah, you name it! Seems to me that appearance gets more and more important. Greetings, Nick. Haa! You found the answer! To be more TI-copy you don't only need to misplace the [ENTER] and give an ALG-Mode, but you need to copy the Model Names, too Irony ofcourse !!!! BUT There's some truth in it. If you don't change the keys then thecase should have, say more metallic color in it. Suggestions on case colouring ?? @--------------------------------------------------- The biggest problems (on my top ten list) of the TI-83 is the lack of more than 1-letter variables. Another problem is the lack of CAS. The top ten best features include a huge Flash !!! And this is missing from the great HP 39/40 calculators (and the CAS from the 39G :-P ) Veli-Pekka Xref: newsmst01.news.prodigy.com comp.sys.hp48:143468 3D Acceleration: Surely OpenGL would be an overkill. Some stripped-down GL customised to a scientist's/engineer's needs would suffice. Something small and object-oriented that fits well within the OS paradigm. Xref: newsmst01.news.prodigy.com comp.sys.hp48:143538 Hello news:DBxK7.692$fj1.31641@news000.worldonline.dk... I doubt it will be viable if you want your hardware to stay in a reasonable price. The cost of a GPRS hardware is prohibitive (minimum of $75 in big quantity with the intellectual property (IP) paid), and you will need to integrate a GPRS software stack which will cost you around 5 millions US dollars for an unlimited use license. A GPRS module would also use your RS232 port if you have only one on your CPU Jean-Yves