A166 Subject: Further Observations on HP49+ Strength and Limitations Here are two observations regarding the quity of the HP49+ from an engineering point of view. One is positive and one negative. While working on the chemistry module of EqaNexMan, an equation-connection-manager, I wanted to test its Insert function which is based on ISOL and MATCH. The following 4 equations from water chemistry served as test object: n = m / Mw eqw = Mw / delox meq = m * 1000 / eqw mgCaCO3eq = meq * 0.05 where: n = moles m = mass Mw = molecular weight eqw = equivent weight meq = mg equivent delox = change in charge mgCaCO3eq = mg CaCO3 equivent After consecutive insertion of the top equation into the next lower one (only Mw had to be isolated) and simplification the HP49+ returned mgCaCO3eq = 50 * n * m * delox / m Try as I wanted, I never could get the ultimately simplified form of mgCaCO3eq = n* delox * 50 (easy to verify with pencil and paper). Is anybody more successful? This might not be new (negative) information since criticism of the COLLECT, COLCT, EXPAND and SIMPLIFY functions were posted earlier, though I believe not in a context like the above insertion sequence. The good news is that with an intelligent program this set and other sets of equations can be solved elegantly with the multiple equation solver, this greatest but so most neglected tool of the HP49+. So great and superfast is this tool, that from an engineering point of view I 't mind other shortcomings of the HP49+, like keyboard and other issues that have been described as beta testing camity. BTW, the above equations (and l the others too) were written into MS Word with a PC keyboard, together with units, pictures, explanations and other embellishments. Before downloading, a VB program strips l superfluous additions and returns a text file that is sent to the HP via Con4x. Some modules contain up to 100 equations and can be converted in a matter of seconds to text files. Wt. === Subject: Re: Further Observations on HP49+ Strength and Limitations X > mgCaCO3eq = 50 * n * m * delox / m > Try as I wanted, I never could get the ultimately simplified form of > mgCaCO3eq = n* delox * 50 (easy to verify with pencil and paper). > Is anybody more successful? > This might not be new (negative) information since criticism of the COLLECT, > COLCT, EXPAND and SIMPLIFY functions were posted earlier, though I believe > not in a context like the above insertion sequence. X I just pressed EV === Subject: Re: Further Observations on HP49+ Strength and Limitations > I just pressed EV > [VPN] Which does not do anything! Still an m both in numerator and denominator, that m is what I want to get rid of. Nor does pressing of SIMPLIFY + EV seem to help. Not on my machine! === Subject: Slash codes () in strings Is there an easy way to extract the slash codes used for formating (bold, itic) text from the string? I'll like to replace those code with the closest PCL codes, in order to send them to a printer. === Subject: Re: Slash codes () in strings > Is there an easy way to extract the slash codes used for formating > (bold, itic) text from the string? I'll like to replace those code > with the closest PCL codes, in order to send them to a printer. I take at that you're looking at a character strings with styles in it, as transferred or edited as text by Conn4x. Actuly, on the cculator, those are ASCII control codes, fairly well conceed by the cculator, but you can find them with the POS command and examine them with the SUB command, and the SIZE command certainly shows that some hidden codes are in the string. ASCII control codes are somewhat problematic to view or edit in a text editor, and might be a re problem if you tried printing the file to a printer in text mode, so Conn4x gives you the option of translating the control codes, except Line Feed (ASCII 10 decim), to the nnn translation codes (where nnn is the 3-digit decim vue of the code), similar to what you can use for Kermit ASCII downloads. See the Help file for Conn4x for more information on the translation options. The following should give you an idea of how the styles and font changes work. Note that styles and font changes can be nested. I'm using the nnn codes here, as using the control codes in a newsgroup post might just a bit of a be a problem, but of course on the cculator, or if transferred with the Conn4x option to translate them turned off, they'd be actu ASCII control codes. %%HP: T(3)A(R)F(.); 019001019BOLD019001019 019002019ITIC019002019 019003019UNDERLINED019003019 019004019INVERSE019004019 019001019019002019019003019019004019LSTYLES019 001019019002019019003019019004019 STARTING WITH SYSTEM FONT 019011019019013019THIS LINE MYFONT 019013019019018019NOW A MINI FONT 019018019019012019NOW A CARRE FONT 019012019019011019 Note that the font changes start with the font that you're leaving, followed by the font that you're going to. so, the vues used for the font change codes are actuly the font ID number plus 11, to avoid duplicating a style code or New Line code. so note that if you end the string without turning off the style or switching back to the system font, the cculator automaticly adds the appropriate code to do it. I suppose that you'd do a search and replace to substitute the printer codes that would be appropriate, but note, for example, that the first occurrence of 019001019 turns bold on, and the next occurrence of that substring turns bold off, so a simple SREPL on a cculator wouldn't be good. With a little work it could be e with the POS and SUB commands in loops. But it may be easier to write some macros for the text editor on your PC. I 't know that the documentation for the cculators have anything about style and font change codes, but a search of the Google archive of the newsgroup should find the information for you. -- === Subject: Re: Slash codes () in strings Is there an easy way to extract the slash codes used for formating (bold, itic) text from the string? I'll like to replace those code with the closest PCL codes, in order to send them to a printer. > I take at that you're looking at a character strings with styles in it, > as transferred or edited as text by Conn4x. Actuly, on the cculator, > those are ASCII control codes, fairly well conceed by the cculator, > but you can find them with the POS command and examine them with the SUB > command, and the SIZE command certainly shows that some hidden codes > are in the string. ASCII control codes are somewhat problematic to view > or edit in a text editor, and might be a re problem if you tried > printing the file to a printer in text mode, so Conn4x gives you the > option of translating the control codes, except Line Feed (ASCII 10 > decim), to the nnn translation codes (where nnn is the 3-digit > decim vue of the code), similar to what you can use for Kermit ASCII > downloads. See the Help file for Conn4x for more information on the > translation options. > The following should give you an idea of how the styles and font changes > work. Note that styles and font changes can be nested. I'm using the > nnn codes here, as using the control codes in a newsgroup post might > just a bit of a be a problem, but of course on the cculator, or if > transferred with the Conn4x option to translate them turned off, they'd > be actu ASCII control codes. > %%HP: T(3)A(R)F(.); > > 019001019BOLD019001019 > 019002019ITIC019002019 > 019003019UNDERLINED019003019 > 019004019INVERSE019004019 019001019019002019019003019019004019LSTYLES019 00101901900 2019019003019019004019 > STARTING WITH SYSTEM FONT > 019011019019013019THIS LINE MYFONT > 019013019019018019NOW A MINI FONT > 019018019019012019NOW A CARRE FONT 019012019019011019 > Note that the font changes start with the font that you're leaving, > followed by the font that you're going to. so, the vues used for the > font change codes are actuly the font ID number plus 11, to avoid > duplicating a style code or New Line code. > so note that if you end the string without turning off the style or > switching back to the system font, the cculator automaticly adds the > appropriate code to do it. > I suppose that you'd do a search and replace to substitute the printer > codes that would be appropriate, but note, for example, that the first > occurrence of 019001019 turns bold on, and the next occurrence of > that substring turns bold off, so a simple SREPL on a cculator > wouldn't be good. With a little work it could be e with the POS and > SUB commands in loops. But it may be easier to write some macros for the > text editor on your PC. > I 't know that the documentation for the cculators have anything > about style and font change codes, but a search of the Google archive of > the newsgroup should find the information for you. > -- > === Subject: Re: Slash codes () in strings You're welcome! I've been thinking of doing something similar, and I've written various Not optimized, but I think that it does what we're looking for. Of course, replace , , and so on with the appropiate codes for your printer. It should be easy enough to add printer codes for font changes if you want to. Tested on the 49G only, but I think that it should be the same on a 49g+. Sorry, I haven't taken the time to comment it, but if you insert a HT at the beginning of the sub-program that I store as a loc variable, and another HT at the beginning of its REPEAT clause, it should be easy enough to single-step through it and see what's going on. Cut and paste the program below to a text file and download it to your cculator. %%HP: T(3)A(R)F(.) @ Example program to replace Style codes. @ Tested on 49G ROM 1.19-6 @ Output from BYTES command: @ # 6EA2h @ 451. << DUP RCLF UNROT STD 64. STWS TYPE DUP UNROT 2. IF =/ THEN ->STR END << 4. ROLL WHILE DUP 5. PICK POS DUP REPEAT OVER 1. PICK3 1. - SUB 5. PICK + UNROT 3. + OVER SIZE SUB + DUP 5. PICK POS IF DUP THEN OVER 1. PICK3 1. - SUB 4. PICK + UNROT 3. + OVER SIZE SUB + ELSE DROP END END DROP 4. ROLLD 3. DROPN >> -> p << 019001019 p EV 019002019 p EV 019003019 p EV 019004019 p EV >> SWAP 2. IF =/ THEN STR-> END SWAP STOF >> -- === Subject: Re: Slash codes () in strings PS: > so note that if you end the string without turning off the style or > switching back to the system font, the cculator automaticly adds the > appropriate code to do it. This seems to be true for styles turned on by using the keys in the Style menu, but not if you add the style code manuly. For example, you can write a string with the style code to turn bold on but leave out the matching style code to turn it back off. so, if such a string is embedded in a composite (program or list), then the style applies to objects (even non-strings) that follow; view my program in the command odd number of occurrences of the same style code. -- === Subject: Re: Slash codes () in strings Thank you, thats exactly the task. I want to send text to an Ir-capable hp printer, but I want to keep the formating. PCL, the printer's language does follow the pre-postfix model for font underlining, style and weight, with minor variations. The plan is to process the file, replacing l '' codes with PCL equivents, and then beam it to the printer. > You're welcome! > I've been thinking of doing something similar, and I've written various > Not optimized, but I think that it does what we're looking for. Of > course, replace , , and so on with the appropiate > codes for your printer. It should be easy enough to add printer codes > for font changes if you want to. Tested on the 49G only, but I think > that it should be the same on a 49g+. Sorry, I haven't taken the time to > comment it, but if you insert a HT at the beginning of the sub-program > that I store as a loc variable, and another HT at the beginning of > its REPEAT clause, it should be easy enough to single-step through it > and see what's going on. Cut and paste the program below to a text file > and download it to your cculator. > %%HP: T(3)A(R)F(.) > @ Example program to replace Style codes. > @ Tested on 49G ROM 1.19-6 > @ Output from BYTES command: > @ # 6EA2h > @ 451. > << > DUP > RCLF > UNROT > STD 64. STWS > TYPE > DUP > UNROT > 2. > IF > =/ > THEN > ->STR > END > << > 4. ROLL > WHILE > DUP > 5. PICK > POS DUP > REPEAT > OVER > 1. > PICK3 > 1. - > SUB > 5. PICK > + > UNROT > 3. + > OVER SIZE > SUB > + > DUP > 5. PICK > POS > IF > DUP > THEN > OVER > 1. > PICK3 > 1. - > SUB > 4. PICK > + > UNROT > 3. + > OVER SIZE > SUB > + > ELSE > DROP > END > END > DROP > 4. ROLLD > 3. DROPN > > - p > << > 019001019 p EV > 019002019 p EV > 019003019 p EV > 019004019 p EV > > SWAP > 2. > IF > =/ > THEN > STR- END > SWAP > STOF > === Subject: Re: Slash codes () in strings Is there an easy way to extract the slash codes used for formating (bold, itic) text from the string? I'll like to replace those code with the closest PCL codes, in order to send them to a printer. === Subject: 49g+ screen capture Is there any way to make an screen capture of the 49g+ without the connectivity kit and even without a pc?. I'd like to take an image (a grob) from the filer or a FAST3D graphic, etc. and save it in HOME or in a port or in the SD in the cc. I guess it's possible to write an ML code (I 't know anything about ML) to modify a little [ON]+UP to make it store the grob in the SD instead of sendind it trough the USB. . jorgecevlos_at_hotmail_dot_com === Subject: Re: 49g+ screen capture draw your 3D curve as you would usuly do, then press on the EXIT button, and enter the graphic environment. then, press STO to save the graphic on the STACK. exit the graphic environment, your graphic is there, ready for you to be stored anyplace you want and reused anytime you want. cyrille > Is there any way to make an screen capture of the 49g+ without the > connectivity kit and even without a pc?. > I'd like to take an image (a grob) from the filer or a FAST3D graphic, > etc. and save it in HOME or in a port or in the SD in the cc. > I guess it's possible to write an ML code (I 't know anything about > ML) to modify a little [ON]+UP to make it store the grob in the SD > instead of sendind it trough the USB. > . > jorgecevlos_at_hotmail_dot_com === Subject: Re: 49g+ screen capture > > Is there any way to make an screen capture of the 49g+ without the > connectivity kit and even without a pc?. > > I'd like to take an image (a grob) from the filer or a FAST3D graphic, > etc. and save it in HOME or in a port or in the SD in the cc. > > I guess it's possible to write an ML code (I 't know anything about > ML) to modify a little [ON]+UP to make it store the grob in the SD > instead of sendind it trough the USB. What you need is the LCD-> command, which puts the current display on level 1 as a grob. But I 't have my 49g+ right now, so I'm not sure that it works as expected with the larger display. I thought that there was a shortcut key to it, but if so, I 't remember it. But you could ways assign it to a user key. But then, user keys are disabled while using the filer, and I'm not sure how you'd access LCD-> from there. Ok, not I complete solution, but I hope that this at least helps. -- === Subject: Re: 49g+ screen capture Is there any way to make an screen capture of the 49g+ without the connectivity kit and even without a pc?. I'd like to take an image (a grob) from the filer or a FAST3D graphic, etc. and save it in HOME or in a port or in the SD in the cc. > What you need is the LCD-> command, which puts the current display on > level 1 as a grob. A quick way to capture the Filer as a grob: << FILER LCD->> Run the program, exit filer, and the grob appears on the stack. Yet another way: store << LCD->> in your dir of choice, load the filer, select it, and press EV. Bill === Subject: Re: 49g+ screen capture try these commands: lcd-> and ->lcd. The first captures the stack and menu and returns that capture as a GROB on to stack level 1. Next you could store this GROB to a variable like you would any other type of variable (i.e. 'a' sto). Recl the GROB you just stored ('t use the ' delimiters). Go into the PRG menu and hit NXT and into the PICT menu. Press PICT and STO. PICT is the variable that holds the current GROB. You can view your capture by just hitting the left arrow key! Hope this helps! -Coiln > Is there any way to make an screen capture of the 49g+ without the > connectivity kit and even without a pc?. > I'd like to take an image (a grob) from the filer or a FAST3D graphic, > etc. and save it in HOME or in a port or in the SD in the cc. > I guess it's possible to write an ML code (I 't know anything about > ML) to modify a little [ON]+UP to make it store the grob in the SD > instead of sendind it trough the USB. > . > jorgecevlos_at_hotmail_dot_com === Subject: Re: 49g+ screen capture >try these commands: lcd-> and ->lcd. The first captures the stack and menu >and returns that capture as a GROB on to stack level 1. You'll need to use Headman. LCD-> only retturns the smler size. Headman will get l of it. Get it at HPcc.org TW === Subject: Re: 49g+ screen capture >try these commands: lcd-> and ->lcd. The first captures the stack and menu >and returns that capture as a GROB on to stack level 1. > You'll need to use Headman. I know I can use it to take a grob from the screen while in the stack, I'm a coauthor of Headman, but LCD2-> cannot be used in the filer or while anything else is running (like a choose box or an input form). I'd like to have something like [ON]+UP to override anything else. > LCD-> only retturns the smler size. > Headman will get l of it. Get it at HPcc.org === Subject: Re: 49g+ screen capture > > jorgecevlos_at_hotmail_dot_com I understood the question wrong I guess. ;-) TW === Subject: Who is Helen? Who is Helen? In comp.sys.hp48, she says she is leading a research group and in the physics department of an Ivy-League university. Yet nowhere in the Ivy League does the physics department claim a faculty member named Helen. Not Brown: http://www.physics.brown.edu/pages/faculty.htm Not Columbia: http://columbia-physics.net/faculty/faculty.htm Not Cornell: http://www.physics.cornell.edu/physics/people/faculty.abc.htm Not Dartmouth: http://www.dartmouth.edu/~physics/people/faculty.html Not Harvard: http://physics.harvard.edu/faculty.htm Not Penn: http://dept.physics.upenn.edu/directory.html Not Princeton: http://pupgg.princeton.edu/www/jh/contact.html http://pupgg.princeton.edu/www/jh/faculty.pdf Not Ye: http://www.ye.edu/physics/faculty/index.html Is Helen a fraud? Is she a re professor who simply chooses to hide behind a pseuym? Helen is only traceable by the email GHMohiam02@yahoo.com, indicating that she is a fan of the Dune series. This, and her hatred of both objectivism and the New HP, are l we truly know about her. Oh yeah, and she likes to troll the newsgroups of the people she hates. comp.sys.hp48 and humanities.philosophy.objectivism See some h.p.o trolls at Since Helen has chosen to hide her identity while trolling publicly, I have e likewise. === Subject: Re: Who is Helen? > It would appear the question asked in the origin post > is still unanswered. Is this because Helen is not who > she claims to be? Or did she claim to be anything in the > last months of posting? The mystery still prevails. The question was answered before the question was asked, in fact. (the origin post even refers to it...) You are free to believe it or not. That's usenet. Camille === Subject: Re: Who is Helen? Who is Helen? She is a lady who entitled to her own opinion. Perhaps you should not get so worked up over her and just ignore her postings if they bother you. I read the postings on this site because I love HP cculators, mathematics, and programming these wonderful little machines. When I read Helen's posts I am sometimes annoyed with her, but maybe that is because at least some of the things she says are painful - but true. But that does'nt make me less of an HP fan at l. Instead of wondering about who Helen is maybe you should focus on the postings that you like and l the cool things that this place offers to HP fans. I mean come on man, HP is back, and the 49+ is awesome. Enjoy! === Subject: Re: Who is Helen? >Who is Helen? She is a lady who entitled to her own opinion. Perhaps >you should not get so worked up over her and just ignore her postings >if they bother you. >I read the postings on this site because I love HP cculators, >mathematics, and programming these wonderful little machines. When I >read Helen's posts I am sometimes annoyed with her, but maybe that is >because at least some of the things she says are painful - but true. >But that does'nt make me less of an HP fan at l. >Instead of wondering about who Helen is maybe you should focus on the >postings that you like and l the cool things that this place offers >to HP fans. >I mean come on man, HP is back, and the 49+ is awesome. Enjoy! Let's have her go after Microsoft. They need lots of work too. === Subject: Re: Who is Helen? I will admit that about a week ago I took a modest snipe at Helen myself, and then wondered what in the heck prompted me to do that. Clearly I'm an adult, just ask my children! Helen is active in this newsgroup, and her name is easy to remember. She does tend to highlight the negatives, though she speaks positively as well. Her comments are intelligent, which is why many of us respect Helen. Actuly, on occasion there are other posters (none of whom can I remember) who make far more negative comments, and seemingly without the consideration and/or anysis that l of us might hope. But, those are their opinions, right? Here's where I'm rely at. I do have a deep-seated fondness for HP cculators. I have derived considerable profession vue and person entertainment with these little machines. Okay, I've frittered away a lot of time with them; but, I've frittered away just as much time on the hockey rink and on the golf course! The bottom line is this: I was ecstatic when I heard that HP was getting back into the business. I'll bet many of the folks on this newsgroup were too. Just look at the posting activity! My big concern is that HP may exit this business again. Surely they need to hear from their customers and fix any and l identified problems. Surely they need to restart their tradition of continuous improvement, just like any good company would. I'll bet many folks here have been involved with product development, and therefore know how difficult it is to create the dreamed-of product. Generly, product development pressures come from l directions: move quickly, reduce costs, add functionity, etc, etc. Many of the same product engineers and business managers who face these pressures read this newsgroup, at least that's my person opinion and experience. These folks listen, they consider the ternatives, they wonder how accurate the newsgroup is in representing the marketplace reity, and they compare newsgroup observations with other sources of customer feedback (distributors, user groups, etc, etc). Anyway, I just want to be strong voice on this newsgroup supporting and encouraging HP. I think the 49G+ is a wonderful new cculator. Perfect? Of course not. Better than the competition? In my opinion and for my purposes, yes. Better than prior models? Well, in virtuly every respect (except perhaps the keyboard), yes. Should HP stay in the cculator business? ABSOLUTELY!!!!! Many of us (especily me) are THRILLED with your newest product. Helen, I apologize for my uncharacteristic snipe. It will not happen again. === Subject: Re: Who is Helen? > Helen, I apologize for my uncharacteristic snipe. 't worry about it. As you say, your swipe was a very mild one in any case. -- Helen. === Subject: Re: Who is Helen? > Who is Helen? and where is Wdo? === .=== > Yet nowhere in the Ivy League does the physics department claim a > faculty member named Helen. > Not Brown: > http://www.physics.brown.edu/pages/faculty.htm > Not Columbia: > http://columbia-physics.net/faculty/faculty.htm > Not Cornell: > http://www.physics.cornell.edu/physics/people/faculty.abc.htm > Not Dartmouth: > http://www.dartmouth.edu/~physics/people/faculty.html > Not Harvard: > http://physics.harvard.edu/faculty.htm > Not Penn: > http://dept.physics.upenn.edu/directory.html > Not Princeton: > http://pupgg.princeton.edu/www/jh/contact.html > http://pupgg.princeton.edu/www/jh/faculty.pdf > Not Ye: > http://www.ye.edu/physics/faculty/index.html === Subject: [ANN] HP4x::Codec released I have released the HP4x::Codec module that Carsten Dominik used to build SDIAG2.00 and mentioned in his announcement. Until it shows up on hpcc.org, you can download it from http://n.ethz.ch/student/trast/download/HP4x-Codec-1.00.zip It's a Perl module that can encode and decode most object types used on the 48 and 49 series cculators. Most notable among the supported types are libraries, directories and BZ (de)compression; the only unsupported ones I can think of right now are linked arrays and ACPTRs. (Which, by the way, could both be integrated if someone actuly needs them. Just write a mail...) and for Emacs and SDIAG! === Subject: Re: Mac and 49G+ connectivity?? > There is NO gen USB seri driver available for Mac OS X. oops Joshua, I stand correct re my above comment and OSX. Cameron Downunder. === Subject: Re: Mac and 49G+ connectivity?? > Once again seeking stories of successful or problematic connectivity > with a Mac OS 9.x or OSX via USB port or other means. > There should be a way to force the Mac to recognize the 49G+ as a > gen USB seri device ... 't ask me, I have no Macs and no > experience with them, but that's the way it works on Linux (according > to Yoann IIRC), and OS X is Un*x too! > As soon as that works you can use any Kermit or XModem capable program > on the seri device. , I agree, though I do not know how. OS 9.2.2 and OS X has gen drivers for reading USB Ram disks. However the re story maybe more complicated than this. Other HP 49+ and Mac users have not found such a direct way yet. Cameron Downunder. === Subject: Re: Mac and 49G+ connectivity?? > Cameron, which model USB->seri cable do you have, there are a few > drivers > available for Mac OS X from non-obvious places you might want to try. > Do a search on http://www.macosxhints.com so. === Subject: Re: conect 49g and 49g+? > Can someone explain me how to make a cable to connect a HP 49 g+ and a 49g. Since the hp49g+ uses USB and the HP49G uses RS232, this is not possible by any simple mean (like would say). BUT you may take advantage of the IrDA -> RS232 adaptor that someone used and mentionned here, connected to the IR port of the 49g+, in order to then use the HP49/HP49 link cable. Yoann. === Subject: Bug report in Warm reset with Try to recover memory ? and NO answer running ROM 1.22 (actuly there were two bugs in my post, but I was explained later that my second bug was a feature :-) I have now upgraded to ROM 1.23 (the offici build downloaded from HP support site) and that bug is still there, so I post it again, hoping to see it fixed sometine - it's not a *big* bug, but still it can be ennoying if you 't know the workaround. Bug : Warm reset with Try to recover memory ? and NO answer When using warm reset : hold down F1 and F6 and press ON, I get the usu message Try to recover memory ? When I hit F1 for YES, everything runs well. When I hit F6 for NO, nothing happens. The workaround is to hit any other key (but F1). The cc beeps, and after that you can hit F6 for NO, which performs the desired operation. === Subject: Re: Bug report in Warm reset with Try to recover memory ? and NO answer > running ROM 1.22 (actuly there were two bugs in my post, but I was > explained later that my second bug was a feature :-) I have now upgraded > to ROM 1.23 (the offici build downloaded from HP support site) and that > bug is still there, so I post it again, hoping to see it fixed sometine - > it's not a *big* bug, but still it can be ennoying if you 't know the > workaround. > Bug : > Warm reset with Try to recover memory ? and NO answer > When using warm reset : hold down F1 and F6 and press ON, I get the usu > message Try to recover memory ? > When I hit F1 for YES, everything runs well. > When I hit F6 for NO, nothing happens. The workaround is to hit any other > key (but F1). The cc beeps, and after that you can hit F6 for NO, which > performs the desired operation. I had no problem doing ON-A-F before my cculator competition. === Subject: Re: Bug report in Warm reset with Try to recover memory ? and NO answer Guess what ? I checked again, AND NOW IT WORKS ! Some time I think of moving back to my old Aristo slide rule. May be computation was not that accurate, at least behaviour was predictable... Grrr... 104mf8sdv0gbj4d@corp.supernews.com... HP49G+ > running ROM 1.22 (actuly there were two bugs in my post, but I was > explained later that my second bug was a feature :-) I have now upgraded > to ROM 1.23 (the offici build downloaded from HP support site) and that > bug is still there, so I post it again, hoping to see it fixed sometine - > it's not a *big* bug, but still it can be ennoying if you 't know the > workaround. > Bug : > Warm reset with Try to recover memory ? and NO answer > When using warm reset : hold down F1 and F6 and press ON, I get the usu > message Try to recover memory ? > When I hit F1 for YES, everything runs well. > When I hit F6 for NO, nothing happens. The workaround is to hit any other > key (but F1). The cc beeps, and after that you can hit F6 for NO, which > performs the desired operation. > I had no problem doing ON-A-F before my cculator competition. > === Subject: Re: Bug report in Warm reset with Try to recover memory ? and NO answer The HP 49G+ is protected against an undesired No key stroke. To get the No you must at first stroke previously rigth shift or left shift (I 't remember with one but please try) Matt M. a .8ecrit dans le message de HP49G+ > running ROM 1.22 (actuly there were two bugs in my post, but I was > explained later that my second bug was a feature :-) I have now upgraded > to ROM 1.23 (the offici build downloaded from HP support site) and that > bug is still there, so I post it again, hoping to see it fixed sometine - > it's not a *big* bug, but still it can be ennoying if you 't know the > workaround. > Bug : > Warm reset with Try to recover memory ? and NO answer > When using warm reset : hold down F1 and F6 and press ON, I get the usu > message Try to recover memory ? > When I hit F1 for YES, everything runs well. > When I hit F6 for NO, nothing happens. The workaround is to hit any other > key (but F1). The cc beeps, and after that you can hit F6 for NO, which > performs the desired operation. > I had no problem doing ON-A-F before my cculator competition. > === Subject: Re: Bug report in Warm reset with Try to recover memory ? and NO answer Would it then be linked to a hardware version ? My S/N is CN33403366, made in China (I guess thay are l made in China...) Herve 104mf8sdv0gbj4d@corp.supernews.com... HP49G+ > running ROM 1.22 (actuly there were two bugs in my post, but I was > explained later that my second bug was a feature :-) I have now upgraded > to ROM 1.23 (the offici build downloaded from HP support site) and that > bug is still there, so I post it again, hoping to see it fixed sometine - > it's not a *big* bug, but still it can be ennoying if you 't know the > workaround. > Bug : > Warm reset with Try to recover memory ? and NO answer > When using warm reset : hold down F1 and F6 and press ON, I get the usu > message Try to recover memory ? > When I hit F1 for YES, everything runs well. > When I hit F6 for NO, nothing happens. The workaround is to hit any other > key (but F1). The cc beeps, and after that you can hit F6 for NO, which > performs the desired operation. > I had no problem doing ON-A-F before my cculator competition. > === Subject: How to locate scratchpad memory? What I would like to do is: 1) Use SysRPL / Emulated Saturn assembly to reserve a chunk of memory to play with. 2) Jump to ARM land, and do something that requires a decent amount of will be used as an object to be placed on the stack, so I need to tell the OS not to override it. This could be e in SysRPL or assembly. Are there any pitfls involving the Garbage Collector? If someone could just give me a link or an explanation I'd be grateful. === Subject: Re: How to locate scratchpad memory? The greysce program has been put on hold since HP support still havn't given me a replacement cculator :-( === Subject: Re: How to locate scratchpad memory? > 1) Use SysRPL / Emulated Saturn assembly to reserve a chunk of memory > to play with. > 2) Jump to ARM land, and do something that requires a decent amount of > will be used as an object to be placed on the stack, so I need to tell > the OS not to override it. This could be e in SysRPL or assembly. The solution below is a variation on the following theme: :: ERRSET CODE GOTO start ENDCODE ERRTRAP :: GARBAGE CODE *start % ML code goes here... % if you run out of memory cl GOVLNG GPMEMERR % ... GOSBVL Shrink$ GOVLNG GPPushR0Lp % or whatever ENCODE ; ; job without garbage collecting first, and simply restarts from the beginning if it runs out of memory. This works very well as long as the average execution time is short, and the ML code part does not change the stack. Pivo's ML tutori file has this and more examples that show how you can locate memory for use in ML programs: http://www.hpcc.org/details.php?id=5007 It sets up some registers before entering ARM mode, and expects you to return something too: * Saturn register contents when entering ARM code: R0[A]: Start address of your memory zone, pointing at prologue R1[A]: Address of the object on stack level 1 D[A]: Number of free nibbles remaining * Saturn register contents when leaving ARM code: D0: End address of the object you want to push CRY: ON if you ran out of memory, the code will cl the garbage collector and then run your ARM code again from the start. Note that the carry flag is ways clear after MOVEDOWN on the 49G, and since that's deliberate (it returns with RTNCC) I assume it's the same on the 49G+. Note that even though the arguments passed in for the ARM code to use are only A fields (5 nibbles), the code clears the rest of the registers. So unless I'm missing something, you can simply load the lower hf of the register into an ARM register, which will then be l zeros in the upper 12 bits and the A field in the lower 20. Here's what (Saturn) memory looks like when entering ARM mode: R0 +5 +10 +10+D[A] ?? | | | | | v v v v v C2A20sssssxxx..xxxcccc...cccc0..0 * C2A20 is the string object prologue (DOCSTR=#02A2C), which you likely want to change to 41620 (DOINT=#02614) for a ZINT either from ARM code or by uncommenting the line near the end * sssss is the current size of the string object, which as I said you must leave untouched * xxx..xxx are D[A] nibbles of scratch space that you can use for anything you want, and write your result to before returning * ccc..ccc is the ARM code igned to a 32-bit boundary * 0..0 are 0-7 nibbles of padding required to achieve the 32-bit ignment Remember that R0 does not necessarily point at a byte boundary. Here's the source code: :: CK1NoBlame DUPTYPEZINT? NcaseTYPEERR ERRSET CODE GOTO start ENDCODE ERRTRAP :: GARBAGE CODE *start GOSBVL SAVPTR A=0.W A=DAT1.A R1=A.W A=0.W A=R0.A R0=A.W AD0EX A+C.A B=A.A LC(5) (end)-(begin) RSTK=C B-C.A LC FFFF8 C&B.A D1=C C=A-C.A D-C.A SKNC { *memerr GOVLNG GPMEMERR } GOSUB end *begin % ARM code goes here *end C=RSTK B=C.A D0=C C=RSTK GOSBVL MOVEDOWN C=B.A INTOFF $80BFF INTON % If the following line fails (jump too long) try to substitute the % next one GOC memerr % SKNC { GOVLNG GPMEMERR } GOSBVL Shrink$ % Uncomment the following line if you want a ZINT as a result % C=R0.W D0=C LC(5) DOINT DAT0=C.A GOVLNG GPPushR0Lp ENDCODE ; SWAPDROP ; Note: This is only a template. Do *not* try running it without adding ARM code that correctly sets up registers before returning, as explained above. It won't work if you just remove the $80BFF or leave it unchanged. I have again only tested this on the 49G, as far as I could. Use at -- If you want to reply by mail, substitute my first and last name for 'foo' and 'bar', respectively. === Subject: Re: How to locate scratchpad memory? There's no need for condition garbage collections in this case. Just do :: NULLHXS #8000 EXPAND % reserve 32k CODE GOSBVL =SAVPTR ... code ... GOSBVL =D0=DSKTOP A=DAT0.A R0=A.W % setup string start LC 0400A C+A.A D0=C % .. and end % Shrink$ will need R0.A to point at the string obj in TEMPOB % D0 to point just past the new end point % and the string will need to be the last object in TEMPOB % (EXPAND took care of that) GOSBVL =Shrink$ GOVLNG =GPPushR0Lp ENDCODE ; @ Werner === Subject: Re: How to locate scratchpad memory? >What I would like to do is: >1) Use SysRPL / Emulated Saturn assembly to reserve a chunk of memory to >play with. >2) Jump to ARM land, and do something that requires a decent amount of >will be used as an object to be placed on the stack, so I need to tell >the OS not to override it. This could be e in SysRPL or assembly. >Are there any pitfls involving the Garbage Collector? >If someone could just give me a link or an explanation I'd be grateful. > :: GARBAGE ( Make sure the maximum amount of memory is free. ) CODE GOSBVL =SAVPTR ... * nibbles ) to a string . String * prologue address is returned in R0. * ... * Here, D0 points one nibble past the end of the data that you * want the fin string to contain. R0 points to the * prologue. GOSBVL =Shrink$ * Truncate l the string data * starting at D0 and resize * accordingly. Unused memory is * reclaimed by the OS. GOSBVL =GPPushR0Lp * Push string to stack. ENDCODE ; ------- --- === Subject: Re: How to locate scratchpad memory? I just saw Page 18 of the pdf in http://www.hpcc.org/details.php?id=5007 My code needs to have some free room for working, and so return a string or integer. I'd like to use MAKE$ to create a 32kb string, use the last 16kb as scratch space (so it gets filled with junk) - Then I'd use the first 16kb or so to store the result. Finly I'd ter the type/size fields so only the first part of the old string (that which contains the result) are returned. To put it graphicly... [ABCDEFGHI] <-- The string as it was created, size 9, type string [2d4f5tftg] <-- The 'string' being used as working room during the ARM code. Size is still set to 9, type is set to string. Contents is junk as far as the HP is concerned. [1234hdfgy] <-- The 'string' after my code completes. The type is now set to Integer, Length 4. I'd want 1234 to be on the stack in this example, and not corrupted by the garbage on the end. I so 't want the junk data wasting memory. Will this work? Or am I totly lost? === Subject: Re: How to locate scratchpad memory? > I'd want 1234 to be on the stack in this example, and not corrupted by > the garbage on the end. I so 't want the junk data wasting memory. I think, in this case, that it will work, but the nibbles unused will not be freed. Your 32kb will ways be reserved, though there is nothing interesting in most of them. -- === Subject: Re: How to locate scratchpad memory? > What I would like to do is: > 1) Use SysRPL / Emulated Saturn assembly to reserve a chunk of memory to > play with. > 2) Jump to ARM land, and do something that requires a decent amount of > will be used as an object to be placed on the stack, so I need to tell > the OS not to override it. This could be e in SysRPL or assembly. > Are there any pitfls involving the Garbage Collector? > If someone could just give me a link or an explanation I'd be grateful. While I Know most nothing about SysRPL a string comes into mind, either in Stack or in a (hidden variable) but what if memory moves? Maybe you could force GC first. === Subject: Re: ARM code extraction utility symbols. 't know what you mean exactly here, but I 't believe the linker is needed if you're just using loc and/or absolute references. But, it wouldn't hurt. >I *do*, however, have a fix for this: use the arm-aout target. Why use the ancient a.out format? ELF is a superset of this and is the most sophisticated and widely used Unix object module format. There's no extra effort involved in its use. But, it rely doesn't matter for the purposes here what format you use. I do agree though that it would be a better idea to use ld just to be on the safe side. Plus, you'll have to use it anyway if you want to link multiple modules or use some type of entry points list. So, there's rely no advantage in using objcopy, since both are just one-liners . ------- --- === Subject: Re: ARM code extraction utility >>but I doubt objcopy works if you've got other >>symbols. > 't know what you mean exactly here, but I 't believe the linker > is needed if you're just using loc and/or absolute references. But, > it wouldn't hurt. Symbols are places where it can jump to. If you use objcopy, I doubt these would be correct, as l it does is copy. However, ld understands about symbols and linking, so it'll be l right. >>I *do*, however, have a fix for this: use the arm-aout target. > Why use the ancient a.out format? ELF is a superset of this and is the > most sophisticated and widely used Unix object module format. There's > no extra effort involved in its use. But, it rely doesn't matter for > the purposes here what format you use. Well, if the armlf-ld supports --oformat binary, then it rely doesn't matter. However, I use a.out format because I tried wince-pe (big mistake), and it didn't support --oformat binary (Error: PE operations on a non-PE file). Since Matthew G. was using the a.out format (it was precompiled for MSDOS), I gave it a try, and it worked well. I see no reason to recompile if it's working fine over here. :-) And you're exactly right: it doesn't matter, if you use objcopy. > I do agree though that it would be a better idea to use ld just to be > on the safe side. Plus, you'll have to use it anyway if you want to > link multiple modules or use some type of entry points list. So, > there's rely no advantage in using objcopy, since both are just > one-liners . Actuly, using objcopy requires three lines -- as -> ld -> objcopy. Using ld, however, only requires two commands -- as -> ld. > -------- -- === Subject: Re: ARM code extraction utility >> 't know what you mean exactly here, but I 't believe the linker >> is needed if you're just using loc and/or absolute references. But, >> it wouldn't hurt. >Symbols are places where it can jump to. If you use objcopy, I doubt >these would be correct, as l it does is copy. However, ld understands >about symbols and linking, so it'll be l right. Well, I'd say it's a little more complicated than that, since a symbol can be any arbitrary (with some restrictions) name-vue pair or section-offset reference, and it has various associated attributes such as type, size etc. But, this is l mute as relative loc symbols and absolute symbols are resolved during the assembly process ie. their referent's are ready patched/fixed-up in the output module and their vues are vid relative to their associated section. >I *do*, however, have a fix for this: use the arm-aout target. >> >> Why use the ancient a.out format? ELF is a superset of this and is the >> most sophisticated and widely used Unix object module format. There's >> no extra effort involved in its use. But, it rely doesn't matter for >> the purposes here what format you use. >Well, if the armlf-ld supports --oformat binary, then it rely >doesn't matter. Well, heh, seems you win here, at least under windoze. ;) The documentation doesn't indicate any problem but it craps out under Cygwin. >> I do agree though that it would be a better idea to use ld just to be >> on the safe side. Plus, you'll have to use it anyway if you want to >> link multiple modules or use some type of entry points list. So, >> there's rely no advantage in using objcopy, since both are just >> one-liners . >Actuly, using objcopy requires three lines -- as -> ld -> objcopy. Only if your module contains extern or inter-section references. (and possibly section relative references as well) But, as I stated in my post, I was tking about when this wasn't the case. ------- --- === Subject: Re: ARM code extraction utility >But, this is l mute Wow, even sentences are starting to tk now. What a crazy world we live in. (Well, unless you're a schizophrenic, in which case it's nothing speci. ;) On another note, I hope these guys http://www.cyc.com/ hurry up and finish before my posts turn into incomprehensible babble. Ooops! Too late. ;) ------- --- === Subject: Re: Samson Cables told the truth as they knew it Funny, I checked the Samson HP-49G+ page and could see NO mention of it being the newer model with l the fixes. On their eBay listing it's mentioned rather prominently. === Subject: Who am I ? boundary=----=_NextPart_000_001B_01C40433.29461B00 charset=iso-8859-1 Yeah, try to decipher me. Let's see how good you are! Here's a taste: yboT === Subject: earthquake problem hp49g+ Hello l ! I have the hp49g+ rom 1.22 Yesterday my cc indicated the low battery sign. It has been only a month since the last replacement with no ROM upgrade between. I didn't replace the battery immediately cause I waited for the LOW BATTERY text box to jump. I did the autotest just for fun and when it got to the ROM TEST the earthquake started... the screen filled with dots in a stripes pattern. Then, I had to get out the batteries and put them back, but then when it turned on the earthquake came back. Of course after I replaced batteries everything was OK. t you think its a bit annoying way to notify me for a low battery ? Did it happen to you too ? There has been too few time from the low battery sign till the batteries died in my opinion. Not enough time even to go to the next store to buy new ones or finish a test ! What do you think ? Idan === Subject: Re: earthquake problem hp49g+ > I have the hp49g+ rom 1.22 < t you think its a bit annoying way to notify me for a low battery ? > Did it happen to you too ? There has been too few time from the low > battery sign till the batteries died in my opinion. Not enough time > even to go to the next store to buy new ones or finish a test ! > What do you think ? I think you should upgrade to 1.23 ASAP. That version gives you a good Low Battery message in plenty of time to buy new ones. === Subject: Re: I'm a beginner...why SysRPL ? > [...] > 2. Normy sysRPL programs are shorter (e.g. DROP SWAP > becomes DROPSWAP) and about 10 % faster than a good > usrRPL-Program. usrRPL is highly redundat in argument checking, > which costs running time. > Is 10% gain i na speed worth fiddling with SysRPL? Well, it's not lways only 10% gain in speed. I know the next program is useless but I'll take it as an example, it takes no arguments and return a list of string representing numbers 1 to 100: UserRpl: << 1 100 FOR I I ->STR NEXT 100 ->LIST >> SysRpl: :: CK0NOLASTWD BINT100 #1+_ONE_DO INDEX@ #>$ LOOP BINT100 {}N ; The first program is 45.5 bytes and takes 0.6598 secs to run (TEV in a 49g+) while the SysRpl program is 25 bytes and takes 0.0707 secs. It's 45% smler and is most 10 times faster. So, it depends of what you want to do to consider learning SysRpl. . jorgecevlos_at_hotmail_dot_com === Subject: Re: I'm a beginner...why SysRPL ? charset=iso-8859-1 Jorge schrieb ... > SysRpl: > :: > CK0NOLASTWD BINT100 #1+_ONE_DO INDEX@ #>$ LOOP BINT100 {}N > ; > The first program is 45.5 bytes and takes 0.6598 secs to run (TEV in > a 49g+) while the SysRpl program is 25 bytes and takes 0.0707 secs. > It's 45% smler and is most 10 times faster. Hi Jorge, this is a good example to sysRPL-beginners, because it use binary-arithmethics, which could give drasticy speed and memory improvements against re-arithmethics. This implies to use binary-arithmetics where ever it is possible in sysRPL! In norm situations you can not gain that much in speed or memory improvement against well written usrRPL. The only benefit o sysRPL is *functionity (things you can not reize with usrRPL). What is your average forcast for speed improvements when using sysRPL against usrRPL and what could be spared in terms of memory space? I would appreciate comments. === Subject: Re: I'm a beginner...why SysRPL ? X > this is a good example to sysRPL-beginners, because it use > binary-arithmethics, > which could give drasticy speed and memory improvements against re-arithmethics. > This implies to use binary-arithmetics where ever it is possible in sysRPL! X A Side Note: Considering that the ARM9 does not have to do BCD->binary->BCD conversions anything with binaries goes faster in the 49g+, even UserRPL loops, etc... === Subject: Re: I'm a beginner...why SysRPL ? [...] > The first program is 45.5 bytes and takes 0.6598 secs to run (TEV in > a 49g+) while the SysRpl program is 25 bytes and takes 0.0707 secs. > It's 45% smler and is most 10 times faster. > So, it depends of what you want to do to consider learning SysRpl. in a need of extended functionity, since I can wait few seconds more :-) I skimmed through 2nd ed. of progSysRPL, but I'd need some more light introduction (if something like that exists), i.e. more tutori-like than reference. Sincerely, Gour === Subject: Re: I'm a beginner...why SysRPL ? >> Do you recommend to first write program in UserRPL and then rewrite in >> SysRPL? > For learning sysRPL it is maybe helpfull, but in gener I would say no. > The main pro for sysRPL is functionity. UserRPL. Sincerely, Gour === Subject: Re: I'm a beginner...why SysRPL ? charset=iso-8859-1 Gour schrieb > The main pro for sysRPL is functionity. > UserRPL. Yes Gour, and knowing usrRPL is a condition/quification for starting with sysRPL. Optimized usrRPL is very close to sysRPL speed. Size of usrRPL programs can be handled by applying BZ-toggling see OT49(+) from Wolfgang: http://page.mi.fu-berlin.de/~raut/WR49/index.htm In case you rey need big speed improvements you should start with assembly code (mashine language: ML) or especily ARM-code for the g+ :-), but this would be a big step for a norm user. === Subject: Re: I'm a beginner...why SysRPL ? > Yes Gour, > and knowing usrRPL is a condition/quification for starting with sysRPL. Is 48 AUR (still) the best reference for UserRPL programming, or 49G+ AUG is sufficient? > Optimized usrRPL is very close to sysRPL speed. Any further info about optimizing UserRPL i.e. what is the 'cost' of different function cls and/or some profiling tool? > Size of usrRPL programs can be handled by applying > BZ-toggling see OT49(+) from Wolfgang: > http://page.mi.fu-berlin.de/~raut/WR49/index.htm Thank you for that. > In case you rey need big speed improvements you should start with > assembly code (mashine language: ML) or especily ARM-code > for the g+ :-), but this would be a big step for a norm user. I'm not so thrilled on the thought of assembler programming :-) Sincerely, Gour === Subject: Re: I'm a beginner...why SysRPL ? charset=iso-8859-1 Gour schrieb > Is 48 AUR (still) the best reference for UserRPL programming, or 49G+ AUG > is sufficient? Yes (HP, forgive me ;-) > Optimized usrRPL is very close to sysRPL speed. > Any further info about optimizing UserRPL i.e. what is the 'cost' of > different function cls and/or some profiling tool? For high speed: - Avoid gebraic expressions. - The STO command takes some time. - Use the stack manipulation, wherever you can (especiy within loops). - make a documentation about your program. ASM > I'm not so thrilled on the thought of assembler programming :-) Me too, but it may change, when thinking about the g+ :-) === Subject: Re: I'm a beginner...why SysRPL ? >> Is 48 AUR (still) the best reference for UserRPL programming, or 49G+ AUG >> is sufficient? > Yes (HP, forgive me ;-) fees - shipping, customs, forwarding ...) I have to pay to import something into Croatia. > For high speed: > - Avoid gebraic expressions. > - The STO command takes some time. > - Use the stack manipulation, wherever you can (especiy within loops). > - make a documentation about your program. > ASM >> I'm not so thrilled on the thought of assembler programming :-) > Me too, but it may change, when thinking about the g+ :-) I'd prefer to have some language compiler available to produce nice ARM code. Sincerely, Gour === Subject: Multiple Equation Solver ? I would like to learn about the Multiple Equation Solver, hp49g+ style, please. Where can I find information on how to use it? === Subject: Re: Multiple Equation Solver ? charset=iso-8859-1 Surely schrieb > I would like to learn about the Multiple Equation Solver, hp49g+ > style, please. Where can I find information on how to use it? Hi Surely, what about the users guide (you should have get it via CD) The multiple Equation Solver is well described. === Subject: Spare parts for 48gx? Where can I get hold of a replacement LCD screen for a HP-48gx? I intend to repair it myself. -- =o Morten Holm () N69¡39,0' E018¡56,9' PGP: 0x898DDDA5(rsa) 0x7EBA2D5D(dh) ( )~( ) The more I de with people, the more I love my dog. === Subject: Re: Spare parts for 48gx? > Where can I get hold of a replacement LCD screen for a HP-48gx? > I intend to repair it myself. Never mind, I just found the FAQ. -- =o Morten Holm () N69¡39,0' E018¡56,9' PGP: 0x898DDDA5(rsa) 0x7EBA2D5D(dh) ( )~( ) The more I de with people, the more I love my dog. === Subject: Re: Spare parts for 48gx? Sorry but buy a new one! Morten Holm a .8ecrit dans le message de > Where can I get hold of a replacement LCD screen for a HP-48gx? > I intend to repair it myself. Never mind, I just found the FAQ. -- =o__ Morten Holm _() _ N69¡39,0' E018¡56,9' PGP: 0x898DDDA5(rsa) 0x7EBA2D5D(dh) (_)~(_) The more I de with people, the more I love my dog. === Subject: Re: gener question: HP49G+ > i've been out of HP-situation developments for a longer time, and > yesterday > i see HP49G+ in a shop. WHAT? i thought the whole project was dead and > those were nothing more than rumours? > anyhow, i was just wondering, what were the practic differences between > 49G and 49G+? i read info on HPs site, but that didn't tell me much. > Faaaaster (can now INVert num matrices 10*TI speed) > Better Garbage Collector, pauses are very short now > SD card reader for 1GB extern storage > +25% screen > screen protector closer to LCD, better view > possibilities for ARM9 assembler programming > continuing ROM upgrades > backup battery instead of a capacitor > better keyboard > better color scheme > [ ' ] and [EV] as primary keys > IR support up to 115,200 > USB support up to 115,200 > no RS-232C seri > no overhead display device connectivity > most l games need a rewrite to slow them down > some other programs need rewriting > gray sce pictures flicker (a rewrite needed) > Forget the trolls and go buy it (year 4000 model eg. seri CH4... ) > mine was fixed by wiggling the few stiff keys > [VPN] You summary is right on the mark. I've been an HP cculator user for over 15 years, and with the newer version of the 49g+, HP finly got it right. Not only was the keyboard fixed, but it seems that they have added some weight in order to make the unit feel more substanti. I am very happy with the cculator and look forward to using it extensively in my engineering work. Unfortunately, those with negative experiences are the most voc in expressing their opinions and those opinions seem to have dominated this newsgroup; therefore, it is important for those of us that are happy with their units to speak out. I look forward to reading these posts everyday; however, it is starting to get old listening to the same old bitching about the 49g+, HP's quity (or lack thereof), Carly Fiorina, etc. I understand that a device will never be l things to l people (particularly the old-timers); however, HP went beyond what I ever expected from them since they decided to resurrect the se of RPN cculators. Remember, for us RPN users, any HP cculator is better than a TI or Casio any day. Later... === Subject: Re: gener question: HP49G+ > It's a free country, > Which country are you tking about ? > [I just wonder which country regulates this forum ;-)] > The USA, where I live and where this is a common phrase. I found it funny to read this phrase on a usenet forum. > Being European > myself, I 't know if there is any specific truth to the statement It's a > free country (look up, for instance, McCarthy era on Google: > The United States has a proud history of politic dissent, but doesn't it > so have an equly notable history of government repression of unpopular > politic speech? Each country has had its dark eras, maybe Cou.8e helps some people to forget it... > About the re power of the memory structure of the HP49+, what do you > mean ? > Until now, I assumed it was pretty much the same memory structure as > the 48 (I 't own a HP49G+). > The power (it's a re power) to create subdirectories and there is more to > it than meets the eye. I would say it is useful. But I assumed that you could do it with any cculator (maybe using some memory manager replacement). Camille === Subject: IRDA link between HP 48G and HP 49G+ ? How can I proceed to transfer files between HP 48 G and HP 49G+ via IrDA link? I 't understand it does not work (It is no pb between HP 48S to HP 48G) === Subject: Re: IRDA link between HP 48G and HP 49G+ ? By the way IrDA works on different frequencies. Maybe try the seri->IR module > How can I proceed to transfer files between HP 48 G and HP 49G+ via IrDA > link? > I 't understand it does not work (It is no pb between HP 48S to HP 48G) > === Subject: Re: IRDA link between HP 48G and HP 49G+ ? Nope. === Subject: Re: IRDA link between HP 48G and HP 49G+ ? By Nope I think you mean IMPOSSIBLE??? <@coqui.net> a .8ecrit dans le message de > Nope. > === Subject: Re: IRDA link between HP 48G and HP 49G+ ? IrDA's minimum transfer rate is standard at 9600. The 48 transmits in the 4k range (can't remember the exact number). There are some assembly tricks to speed it up, but it does not reach 9600. I want to say 'impossible', but there's ways a little hope... > By Nope I think you mean IMPOSSIBLE??? > > <@coqui.net> a .8ecrit dans le message de > Nope. > === Subject: Re: IRDA link between HP 48G and HP 49G+ ? > IrDA's minimum transfer rate is standard at 9600. The 48 transmits in the > 4k range (can't remember the exact number). There are some assembly tricks > to speed it up, but it does not reach 9600. I got the 49g+ to transfer files with a pocket pc through IrDA using the xmodem protocol and DOS. So you're saying, as-is the 48g isn't fast enough for transfer with 49g+? Is the IrDA transfer speed of the 49g+ at 115200? The program on my pocket pc says I'm transfering at 44000, but it feels slower than my 28K modem connection. === Subject: Re: IRDA link between HP 48G and HP 49G+ ? > IrDA's minimum transfer rate is standard at 9600. The 48 transmits in the > 4k range (can't remember the exact number). There are some assembly tricks > to speed it up, but it does not reach 9600. > I got the 49g+ to transfer files with a pocket pc through IrDA using > the xmodem protocol and DOS. So you're saying, as-is the 48g isn't 48G is 4800 IIRC, the 49g+ can be set to this speed.... > fast enough for transfer with 49g+? Is the IrDA transfer speed of the > 49g+ at 115200? The program on my pocket pc says I'm transfering at > 44000, but it feels slower than my 28K modem connection. Are they both set to 115,200? === Subject: Re: IRDA link between HP 48G and HP 49G+ ? >> IrDA's minimum transfer rate is standard at 9600. The 48 transmits in the >> 4k range (can't remember the exact number). There are some assembly tricks >> to speed it up, but it does not reach 9600. >I got the 49g+ to transfer files with a pocket pc through IrDA using >the xmodem protocol and DOS. So you're saying, as-is the 48g isn't >fast enough for transfer with 49g+? Is the IrDA transfer speed of the >49g+ at 115200? The program on my pocket pc says I'm transfering at >44000, but it feels slower than my 28K modem connection. > The IrDA to seri link cable that I have is software setable from 9600 to 115200. === Subject: Re: IRDA link between HP 48G and HP 49G+ ? Could you please discuss, in greater detail, the IrDA to seri link cable you're using for the hp49g+? 1. Does it actuly work well enough to function as a seri port? 2. Can I purchase one like it somewhere? 3. Are there any adaptations necessary to make it work, like drivers or hardware fittings for the cculator? 4. Any other details you've discovered that make it more useful. Thank you, Surely > IrDA's minimum transfer rate is standard at 9600. The 48 transmits in the > 4k range (can't remember the exact number). There are some assembly tricks > to speed it up, but it does not reach 9600. >>I got the 49g+ to transfer files with a pocket pc through IrDA using >>the xmodem protocol and DOS. So you're saying, as-is the 48g isn't >>fast enough for transfer with 49g+? Is the IrDA transfer speed of the >>49g+ at 115200? The program on my pocket pc says I'm transfering at >>44000, but it feels slower than my 28K modem connection. > The IrDA to seri link cable that I have is software setable from >9600 to 115200. === Subject: Re: IRDA link between HP 48G and HP 49G+ ? >Could you please discuss, in greater detail, the IrDA to seri link >cable you're using for the hp49g+? >1. Does it actuly work well enough to function as a seri port? >2. Can I purchase one like it somewhere? >3. Are there any adaptations necessary to make it work, like >drivers or hardware fittings for the cculator? >4. Any other details you've discovered that make it more useful. >Thank you, >Surely >> IrDA's minimum transfer rate is standard at 9600. The 48 transmits in the >> 4k range (can't remember the exact number). There are some assembly tricks >> to speed it up, but it does not reach 9600. >I got the 49g+ to transfer files with a pocket pc through IrDA using >the xmodem protocol and DOS. So you're saying, as-is the 48g isn't >fast enough for transfer with 49g+? Is the IrDA transfer speed of the >49g+ at 115200? The program on my pocket pc says I'm transfering at >44000, but it feels slower than my 28K modem connection. > The IrDA to seri link cable that I have is software setable from >>9600 to 115200. You can find out l about it here. http://www.actisys.com/instantir.html It is so sold as an OEM board( No wires or shell). for about $65.00 U.S. in quantity. It is $100.00 U.S. as show on the web page. I have only had time to use the software to set the speed and I got it to tk to a Vernier Univers Lab Interface Box and their new LabPro. I haven't e anything useful yet. Hopefully this summer when my teaching load is not so heavy I can rely use it to create a data collector using the 49G+. === Subject: Re: IRDA link between HP 48G and HP 49G+ ? What's that little cable extending from the DB-9 end? >Could you please discuss, in greater detail, the IrDA to seri link >cable you're using for the hp49g+? >1. Does it actuly work well enough to function as a seri port? >2. Can I purchase one like it somewhere? >3. Are there any adaptations necessary to make it work, like >drivers or hardware fittings for the cculator? >4. Any other details you've discovered that make it more useful. >> IrDA's minimum transfer rate is standard at 9600. The 48 transmits in the >> 4k range (can't remember the exact number). There are some assembly tricks >> to speed it up, but it does not reach 9600. >I got the 49g+ to transfer files with a pocket pc through IrDA using >the xmodem protocol and DOS. So you're saying, as-is the 48g isn't >fast enough for transfer with 49g+? Is the IrDA transfer speed of the >49g+ at 115200? The program on my pocket pc says I'm transfering at >44000, but it feels slower than my 28K modem connection. >> The IrDA to seri link cable that I have is software setable from >>9600 to 115200. > You can find out l about it here. > http://www.actisys.com/instantir.html > It is so sold as an OEM board( No wires or shell). for about $65.00 > U.S. in quantity. It is $100.00 U.S. as show on the web page. I have > only had time to use the software to set the speed and I got it to > tk to a Vernier Univers Lab Interface Box and their new LabPro. > I haven't e anything useful yet. Hopefully this summer when my > teaching load is not so heavy I can rely use it to create a data > collector using the 49G+. === Subject: Re: IRDA link between HP 48G and HP 49G+ ? >>Could you please discuss, in greater detail, the IrDA to seri link >>cable you're using for the hp49g+? >>1. Does it actuly work well enough to function as a seri port? >>2. Can I purchase one like it somewhere? >>3. Are there any adaptations necessary to make it work, like >>drivers or hardware fittings for the cculator? >>4. Any other details you've discovered that make it more useful. >>Thank you, >>Surely > IrDA's minimum transfer rate is standard at 9600. The 48 transmits in the > 4k range (can't remember the exact number). There are some assembly tricks > to speed it up, but it does not reach 9600. >>I got the 49g+ to transfer files with a pocket pc through IrDA using >>the xmodem protocol and DOS. So you're saying, as-is the 48g isn't >>fast enough for transfer with 49g+? Is the IrDA transfer speed of the >>49g+ at 115200? The program on my pocket pc says I'm transfering at >>44000, but it feels slower than my 28K modem connection. > The IrDA to seri link cable that I have is software setable from >9600 to 115200. > You can find out l about it here. >http://www.actisys.com/instantir.html >It is so sold as an OEM board( No wires or shell). for about $65.00 >U.S. in quantity. It is $100.00 U.S. as show on the web page. I have >only had time to use the software to set the speed and I got it to >tk to a Vernier Univers Lab Interface Box and their new LabPro. > I haven't e anything useful yet. Hopefully this summer when my >teaching load is not so heavy I can rely use it to create a data >collector using the 49G+. I forgot to mention that a gentleman surveyor in Ohio has ready used the IrDA to seri cable (OEM version) to use his G+ to collect data from his tot station survey instrument. He has built a prototype of a cradle for the cculator, the OEM board, a sml mirror to shorten the light path, and a seri plug. You can read more in the P.O.B. Point of Beginning web site and look at discussions from October to December of last year. I have not heard anything more about it since just before tmas on this group. === Subject: HP 49G+ cable to transfer to HP 49G Do you know if any cable exist to trnasfer files between HP 49G & HP 49G+? === Subject: Re: HP 49G+ cable to transfer to HP 49G If youre high on electronic try this: There exist usb<->seri modules I 't exactly remember where I saw them. Its pretty easy to assemble, you maybe need an extern im and probably to write some bits in the EEPROM of the module (this is only for seri->usb) Just search over the internet... Dimitri > Do you know if any cable exist to trnasfer files between HP 49G & HP 49G+? > Thank you in advance > === Subject: Re: HP 49G+ cable to transfer to HP 49G Nope. === Subject: HP 49G+ ROM 1.23 With ROM 1.23 I can not transfert files between cculator and my computer?? But I have no problem with the 1.22 ROM. Why? Do you have the same issue? How can I solve it? Thank you for your help === Subject: Re: HP 49G+ ROM 1.23 I have no problem with ROM 1.23. - First be sure that you are using ROM 1.23 coming from HP web site - Use the latest release of conn4x from HP site. If you cannot transfer from PC to HP use SD card to instl the ROM. Sebastien <@yahoo.fr> a .8ecrit dans le message de > With ROM 1.23 I can not transfert files between cculator and my computer?? > But I have no problem with the 1.22 ROM. > Why? Do you have the same issue? How can I solve it? > Thank you for your help > === Subject: Re: HP 49G+ ROM 1.23 ROM 1.23 is HP and I Use the latest release of conn4x from HP site. Then It is no Pb to instl it on HP 49G+ with conn4x then again no pb to come back on the 1.22 ROM version with conn4x. But no transfer between PC and HP 49G+ with ROM 1.23 while it is 100% working with ROM 1.22 (PC conn4x can not find the cculator???). Sebastien Devroe a .8ecrit dans le message de > I have no problem with ROM 1.23. > - First be sure that you are using ROM 1.23 coming from HP web site > - Use the latest release of conn4x from HP site. > If you cannot transfer from PC to HP use SD card to instl the ROM. > Sebastien > <@yahoo.fr> a .8ecrit dans le message de > With ROM 1.23 I can not transfert files between cculator and my > computer?? > But I have no problem with the 1.22 ROM. > Why? Do you have the same issue? How can I solve it? === Subject: Re: HP 49G+ ROM 1.23 > But no transfer between PC and HP 49G+ with ROM 1.23 while it is 100% > working with ROM 1.22 (PC conn4x can not find the cculator???). > On the cc: flag -33 must be cleared, flag -34 must be set. Be sure you press and RELEASE right-shift then press right-arrow. Go to Windows Device Manager and check USB driver version for 1.2.0.0 If not 1.2.0.0 click Update Driver, etc Have Disk and browse to Program FilesHewlett-PackardConn4xUSBDriver This will make sure you have the latest driver. There rely are no reported problems using Windows XP and the connections. It should work fine if you do the above. -- - - - - - - - - - - - - - - - - Bill Graves RKBA! bgraves@ix.netcom.com === Subject: Re: HP 49G+ ROM 1.23 ... or a digit camera, if it can mount as a removable hard drive, as mine does. > I have no problem with ROM 1.23. > - First be sure that you are using ROM 1.23 coming from HP web site > - Use the latest release of conn4x from HP site. > If you cannot transfer from PC to HP use SD card to instl the ROM. > Sebastien > <@yahoo.fr> a .8ecrit dans le message de > With ROM 1.23 I can not transfert files between cculator and my > computer?? > But I have no problem with the 1.22 ROM. > Why? Do you have the same issue? How can I solve it? === Subject: Externs on HP 48GX 1- How can I define the RPL routines (Extern) with in a machine language source code with the HP 48GX. For instance RPL Routines are very usefull at the beginning of a LM program to verify number of arguments, to enter a number in the stack.... 2- With Assembly software is teh best for HP 48GX? Thank you for your help === Subject: Re: Externs on HP 48GX >1- How can I define the RPL routines (Extern) with in a machine language >source code with the HP 48GX. >For instance RPL Routines are very usefull at the beginning of a LM program >to verify number of arguments, to enter a number in the stack.... I assume you're asking about how to cl RPL routines from within a code object. This is a FAQ, but unfortunately there isn't a system programming FAQ (that I know of) and even the gener one is quite out of date, so I guess I'll post the umpteenth answer to this question. See : Note that the extra code needed for embedded RPL usuly outweighs any speed, size, or convenience benefits unless you have a significant amount of it. It's usuly smler and easier to just embed your code object in a secondary. >2- With Assembly software is teh best for HP 48GX? === Subject: Re: Externs on HP 48GX > 1- How can I define the RPL routines (Extern) with in a machine language > source code with the HP 48GX. > For instance RPL Routines are very usefull at the beginning of a LM program > to verify number of arguments, to enter a number in the stack.... > 2- With Assembly software is teh best for HP 48GX? > Thank you for your help > Try out Jazz. With Jazz you can assemble/disassemble everything on your HP48. There is even a sys-rpl and ml debugger. Get it from here: http://www.engr.uvic.ca/~aschoorl/ http://www.fi.tartu.ee/danli97/hp48/hp48_jazz.html There is even a port to the HP49G. -- === Subject: HP 49G+ Transfer via Hyper Termins How to proceed?? Thank you for your help === Subject: Last cl for HTML v.s. plain-jane text boundary=----=_NextPart_000_0023_01C4046A.0ACF51A0 charset=iso-8859-1 I've received a few requests to quit using html in this ng. I rely 't think it is a big de, but I do prefer html. I would like to get the killer reason (please, keep me in your lists), the unnegotiable one. Am I the only one using it? If not, please voice your opinion so. Of course, No responses means that the issue is just not that important here. I promise this is my very last thread on this issue. === Subject: Re: Last cl for HTML v.s. plain-jane text Well, sever posts against HTML, not a single one for it. like the deodorant commerci: 'I switched'. (I think I switched. hopefully this one comes out right) === Subject: Re: Last cl for HTML v.s. plain-jane text > I've received a few requests to quit using html in this ng. I rely > 't think it is a big de, but I do prefer html. I would like to get > the killer reason (please, keep me in your lists), the unnegotiable > one. Am I the only one using it? If not, please voice your opinion so. I've ready exposed my view against html in USENET. I'd like it better if you 't use it. My reasons against it are byte-size and the fact that your posts use a too sml font which is most unreadable at my norm font size. -- === Subject: Re: Last cl for HTML v.s. plain-jane text The simplest and most fundament reason for plain text to a newsgroup is that it is universly compatible and readable. That is why it IS a big de--plain text is ways plain! Of course, if you, like me, you use M$ Outlook Express, you might accidently send html-specily when replying to an html message---you have to remember to go up to format and choose plain text and then acknowledge to the annoying warning box that yes, I rely do want to use plain text {why do they have those annoying messages!}. Bill I've received a few requests to quit using html in this ng. I rely 't think it is a big de, but I do prefer html. I would like to get the killer reason (please, keep me in your lists), the unnegotiable one. Am I the only one using it? If not, please voice your opinion so. Of course, No responses means that the issue is just not that important here. I promise this is my very last thread on this issue. === Subject: Re: Last cl for HTML v.s. plain-jane text [...] [...] > > I do di up, by the way. > I've received a few requests to quit using html in this > ng.  I rely 't think it is a big de, but I do prefer html.  I > would like to get the killer reason (please, keep me in your lists), the > unnegotiable one.  Am I the only one using it?  If not, please voice > your opinion so. Of course, No responses means that the issue is just not that > important here.  I promise this is my very last thread on this > issue.  I do di up, by the way.   I suggest you get a newsreader that works as it should. I.e. one where you can set the word-wrap at 80 or so, see the posts as monospaced so you can follow any code samples that are posted, and insert/remove signatures the correct way (a signature comes below dash-dash-space). About the only thing I miss from using windows is using Diog. The best newsreader I've ever used: http://www.40tude.com/diog/ I recommend it to anyone still using Outlook Express. === Subject: Re: Last cl for HTML v.s. plain-jane text > I've received a few requests to quit using html in this ng. I rely > 't think it is a big de, but I do prefer html. I would like to get > the killer reason (please, keep me in your lists), the unnegotiable > one. Am I the only one using it? If not, please voice your opinion so. this category: The third link has 7 biggies that you should pay attention to. Second point is 'NO HTML'. You should trust the advice of other more experienced Usenet users, as they know what is good/bad/annoying/stupid/smart/asinine/etc. Using HTML in your posts has very little advantage to it, compared to the disadvantages of the other users who may be forced to use newsreaders that are rather incompatible with HTML. It would so do good for even the older users who may not have posted to Usenet in a while, such as myself, to re-read these FAQs if we have forgotten simple things that annoy others. === Subject: Re: Last cl for HTML v.s. plain-jane text Aside from the bandwidth issue (the text part of your post has more than 1800 characters instead of about 450) your mailer tries to use HTML to force its idea of what the text should look like on other users, which I do find somewhat obtrusive behaviour. FYI, I use Gnus/Emacs and have set it to never render any HTML. It will simply display the text/plain MIME part. -- If you want to reply by mail, substitute my first and last name for 'foo' and 'bar', respectively. === Subject: Re: Last cl for HTML v.s. plain-jane text >I've received a few requests to quit using html in this ng. I rely 't think it is a big de, but I do prefer html. I would like to get the killer reason (please, keep me in your lists), the unnegotiable one. Am I the only one using it? If not, please voice your opinion so. >Of course, No responses means that the issue is just not that important here. I promise this is my very last thread on this issue. I do di up, by the way. One good reason not to use it is that micious code cam be lirking in HTML code. Even if it does not originate with you, it could be hijacked and added. === Subject: Problems with SysRPL/sasm code I've written (actuly, partily copied) the following code for person use to manipulate contrast on my HP49g+. It's written with HP's debug4x. Works great in the emulator. On my cculator: 1) GETCT ways returns 14, regardless of the actu contrast. 2) SAVECT ways stores 14, regardless of the actu contrast. 3) RESETCT sometimes does nothing, and sometimes acts like UP or DOWN. I'm at the end of my rope. Can someone else take a look and hunt for problems? RPL INCLUDE HPContrast.h ASSEMBLE CON(1) 8 RPL :: CODE GOSBVL =POP# GOSBVL =SAVPTR C=0 A LC(2) #1F ?C ITE :: DROP % 440 % 0.25 DOBEEP ; :: !!SETCTST ; ; :: !!GETCTST #1- DUP # 1 #< ITE :: DROP % 440 % 0.25 DOBEEP ; :: !!SETCTST ; ; :: !!GETCTST UNCOERCE ; :: ' ID DEFCONTRAST @ ITE :: CK&DISPATCH1 ONE :: COERCE !!SETCTST ; ; :: # 0D !!SETCTST ; ; :: !!GETCTST UNCOERCE ' ID DEFCONTRAST STO ; -- === Subject: Re: Problems with SysRPL/sasm code >I've written (actuly, partily copied) the following code for >person use to manipulate contrast on my HP49g+. It's written with >HP's debug4x. Works great in the emulator. On my cculator: > 1) GETCT ways returns 14, regardless of the actu contrast. > 2) SAVECT ways stores 14, regardless of the actu contrast. > 3) RESETCT sometimes does nothing, and sometimes acts like > UP or DOWN. >I'm at the end of my rope. Can someone else take a look and hunt for >problems? Unfortunately, the 49G+'s emulation of many of the Yorke hardware registers is bug ridden. Here's is some information from Yoann Desir about the broken contrast emulation and how to get around it : Unfortunately, while you can *change* the contrast, there doesn't appear to be any immediately obvious way to retrieve the current contrast setting in Saturn assembly. Further, after consulting the S3C2410 User's manu, I couldn't find any mention of the ability to change the LCD contrast via one of the LCD speci function registers. So, this means either: (a) The contrast is controlled via some GPIO lines connected to the display driver, or (b) There is no contrast control capability and the apparent contrast change in Saturn mode is just a different graysce level generated by the hardware various usenet posts asking about how to change the iPaq contrast, I think it's probably possibility (b). ------- --- === Subject: Re: Problems with SysRPL/sasm code > I've written (actuly, partily copied) the following code for > person use to manipulate contrast on my HP49g+. It's written with > HP's debug4x. Works great in the emulator. On my cculator: > 1) GETCT ways returns 14, regardless of the actu contrast. > 2) SAVECT ways stores 14, regardless of the actu contrast. > 3) RESETCT sometimes does nothing, and sometimes acts like > UP or DOWN. > I'm at the end of my rope. Can someone else take a look and hunt for > problems? [...] Take a look at how I do it on the HP48. I'm not sure about differences with HP49G+, but it might be worth a look: http://www.fi.tartu.ee/danli97/hp48/hp48_contrast.html -- === Subject: Re: HP49G LGPL CAS release Manel. >Hopefully, both problems will be corrected one day. Anybody can explain me the meanning or ... one day. Could be 1, 2, 3 years or... never. I'd like to think JYA or Parisse have no responsability in this subject. HP would have to pay to us (people how write in this group) for offering a free technic support. Of course the way to pay us is publishing buggy roms. Absolut emetic. Or you think the HP49G is what it is due to the extensive information provided by HP (I tk about the wonderful manus of the HP, where you can solve any doubt. (of course it is a irony). The treatment of HP towards us would have to be more correct. They must publish a new USEFUL ROM for the OLD HP49G with sources, of course!!!!! === Subject: Re: HP49G+ Return/New Keyboard > I sent my cculator to New Age after getting a case number from an HP > rep and was promised that a new cculator would be sent out > immediately upon reception of my old unit and would ARRIVE within > 2-business days. > Well, New Age had my cculator on Friday morning, 02/27 and shipped > out a replacement on Tuesday, March 2nd (2 business days later). It > is due to arrive on Tuesday, MARCH 9TH (a full 7 BUSINESS DAYS LATER). > So, the mor of the story is to push HP into sending a replacement > unit BEFORE you part with your cculator. This was not given to me > as an option when I cled, even with a cc #. > BTW, Just in case this happened, I purchased a NEW 49G+ from > outpost.com (seri CN402...) BEFORE parting with my OLD unit (seri > CN33...). The new keyboard is not as noisy and doesn't miss as many > keystrokes as the origin keyboard. In addition, the new cculator > is slightly heavier. > I am disappointed in the HP Cculator Division since the new keyboard > shows only increment improvement over the old unit and the tech > support rep was less than honest about the replacement shipping > policy. I was in the process of starting an HP Cculator User Group > at RIT but have decided against it for the time being. My students > deserve a fully function cculator. > I will continue to use my 49G+ cculators and hope for a better > keyboard/unit downstream, though the Voyage 200 by TI looks like a > nice unit. > - Former owner of the HP15C, 28S, 48GX, 49G > - Current owner of (2) HP49G+ cculators and an HP IPAQ 5450 (which > l have their own problems) It seems that HP is getting super paranoid about these keyboard related returns because some urchins had been taking advantage of HP's initi good spirit. (Free 49G+ anyone?) When I cled their support line to initiate the return process, they wanted l sorts of information, including the cculator's seri number, where I bought it, the order number (from HPshopping) and a DETAILED explanation of the problems with the keyboard. I have yet to receive my replacement, but I have high hopes that this new one will have better key recognization (the old one required extremely hard keypresses to register anything at l). --Jason Pratt jspREMOVETHIS@asu.edu === Subject: Matrix Operations (SysRPL) I'm trying to perform some matrix operations in a SysRPL program (^MAT+, ^MATSQUARE, etc), but cannot figure out why they won't work. I'm creating an array with ^XEQ>ARRY. I've then converted it to a matrix with ^ARRAY2MATRIX and I've so tried ^ArrayToMatrix. When I then try to use ^MAT+ or any other matrix operation it bombs out. Can anyone tell me what I'm omitting or doing wrong? === Subject: Re: Matrix Operations (SysRPL) >I'm trying to perform some matrix operations in a SysRPL program (^MAT+, >^MATSQUARE, etc), but cannot figure out why they won't work. I'm creating >an array with ^XEQ>ARRY. I've then converted it to a matrix with >^ARRAY2MATRIX and I've so tried ^ArrayToMatrix. When I then try to use >^MAT+ or any other matrix operation it bombs out. Can anyone tell me what >I'm omitting or doing wrong? > Are4 you using debug4x or on cc? In debug4x you need to decre ROMPTR and FPTRS as EXTERN in the beginning. I know that threw me off in the beginning. === Subject: Re: Matrix Operations (SysRPL) I should probably clarify that I'm not getting compiler errors. I'm getting errors on the cculator when I try to run it. Debug4x doesn't have a problem with the program at l... > Are4 you using debug4x or on cc? In debug4x you need to decre > ROMPTR and FPTRS as EXTERN in the beginning. I know that threw me > off in the beginning. === Subject: Re: Matrix Operations (SysRPL) > I should probably clarify that I'm not getting compiler errors. I'm getting > errors on the cculator when I try to run it. Debug4x doesn't have a > problem with the program at l... I haven't used Debug4x in a long time, but I remember sometimes it gave me no problems during compilation, but it didn't compile what I thought. It was a problem with the declaration. I suggest you to use Nosy in the emulator and see the code compiled by Debug4x. -- === Subject: Re: Matrix Operations (SysRPL) > Are4 you using debug4x or on cc? In debug4x you need to decre > ROMPTR and FPTRS as EXTERN in the beginning. I know that threw me > off in the beginning. Tim, Yes, I'm using Debug4x, but I'm using FLASHPTR where appropriate. Any command with the '^' would be written in Debug4x as: FLASHPTR XEQ>ARRY FLASHPTR ARRAY2MATRIX FLASHPTR ArrayToMatrix FLASHPTR MAT+ FLASHPTR MATSQUARE etc... Is that what you're referring to? === Subject: Transposing a formula with the 49g+ Hi Folks, If I have a simply formula such as ... 2x + 3y = 10 is there a way to make the hp49g+ turn it around and express it in terms of a variable of my choice, eg ... x= (10-3y) /2 CC PS: It's the actu transposition of the formula I'm after, not the num result === Subject: Re: Transposing a formula with the 49g+ > Hi Folks, > If I have a simply formula such as ... > 2x + 3y = 10 > is there a way to make the hp49g+ turn it around and express it in terms of > a variable of my choice, eg ... > x= (10-3y) /2 > CC > PS: It's the actu transposition of the formula I'm after, not the > num result '2*X+3*Y=10' 'X' ISOL If you have a quadratic, switch ISOL for QUAD. -- === Subject: Re: Transposing a formula with the 49g+ > If I have a simply formula such as ... > 2x + 3y = 10 > is there a way to make the hp49g+ turn it around and express it in terms of > a variable of my choice, eg ... > x= (10-3y) /2 Do this: '2*x+3*y=10' Enter x Enter Solve The cculator gives 3y-10 x = - ------- 2 Note the leading minus sign. -- Capaci Occasio === Subject: Re: Transposing a formula with the 49g+ I keep ending up with ... X = -(2.*Y-5.) where if you take x and y as being 2 in my origin equasion, to me (and keep in mind that math isn't my strong suit), I end up with a wrong answer. Where am I going wrong here? CC > If I have a simply formula such as ... > 2x + 3y = 10 > is there a way to make the hp49g+ turn it around and express it in terms > of > a variable of my choice, eg ... > x= (10-3y) /2 > Do this: > '2*x+3*y=10' > Enter > x > Enter > Solve > The cculator gives > 3y-10 > x = - ------- > 2 > Note the leading minus sign. > > -- > Capaci Occasio === Subject: Re: Transposing a formula with the 49g+ You probably have your cculator set to display numbers in FIX 0 format; that is, your numbers will be rounded to their integer part. If you change this mode (MODE => Number Format => Standard => OK => OK), you'll see the fraction part too. > I keep ending up with ... > X = -(2.*Y-5.) > where if you take x and y as being 2 in my origin equasion, to me (and > keep in mind that math isn't my strong suit), I end up with a wrong answer. > Where am I going wrong here? > CC >If I have a simply formula such as ... >2x + 3y = 10 >is there a way to make the hp49g+ turn it around and express it in terms >>of >a variable of my choice, eg ... >x= (10-3y) /2 >>Do this: >>'2*x+3*y=10' >>Enter >>x >>Enter >>Solve >>The cculator gives >> 3y-10 >>x = - ------- >> 2 >>Note the leading minus sign. >-- >>Capaci Occasio === Subject: Re: Transposing a formula with the 49g+ Next issue though ... When I put in something like X=2 PI SQRT (LC) and try to solve Isolate for L it says it can't do it - any ideas anyone? CC > You probably have your cculator set to display numbers in FIX 0 > format; that is, your numbers will be rounded to their integer part. If > you change this mode (MODE => Number Format => Standard => OK => OK), > you'll see the fraction part too. > I keep ending up with ... > X = -(2.*Y-5.) > where if you take x and y as being 2 in my origin equasion, to me (and > keep in mind that math isn't my strong suit), I end up with a wrong answer. > Where am I going wrong here? > CC >If I have a simply formula such as ... 2x + 3y = 10 is there a way to make the hp49g+ turn it around and express it in terms >>of >a variable of my choice, eg ... x= (10-3y) /2 >>Do this: >>'2*x+3*y=10' >>Enter >>x >>Enter >>Solve >>The cculator gives >> 3y-10 >>x = - ------- >> 2 >>Note the leading minus sign. === Subject: Re: HP49g+ in Oslo? > does anyone know if/where the HP49g+ is available in Oslo centrum? Months ago, I posted to this group that I had seen it demonstrated at Tanum's on Karl Johan. Presumably it is still available. Another possibility is Norli in Universitetsgaten (I think I have seen it there, too). === Subject: Re: HP 49G+ Questions > use this link: > http://www.hpcc.org/hp49/math/misc/eqlibhp49.zip > this is a pretty god library > the info is on this link: http://www.hpcc.org/details.php?id=4577 > i'm not sure if you need bz49 but you can get it l at: > www.hpcc.org/hp49 === Subject: Re: HP 49G+ Questions > GX did/does, but for the 48SX you need a ROM card for that. > You have been ready answered about the extern library > for the 49g+ in another post, but the www.hpcc.org > has so many - many other good programs. > [VPN] VPN, unsure which is the best for gener use. I guess I should be more specific, and ask for opinions on which library would best serve my purpose. I am mainly using the cculator in a trigonometry class, and will later be using it in multiple levels of Cculus classes as === Subject: Re: HP 49G+ Questions > GX did/does, but for the 48SX you need a ROM card for that. > You have been ready answered about the extern library > for the 49g+ in another post, but the www.hpcc.org > has so many - many other good programs. > [VPN] > VPN, > unsure which is the best for gener use. I guess I should be more > specific, and ask for opinions on which library would best serve my > purpose. I am mainly using the cculator in a trigonometry class, Ahaa! Try this one: http://www.hpcc.org/details.php?id=5103 > and will later be using it in multiple levels of Cculus classes as Cculus? Search Marathon > === Subject: Re: HP 49G+ Questions > Ahaa! > Try this one: > http://www.hpcc.org/details.php?id=5103 > and will later be using it in multiple levels of Cculus classes as > Cculus? > Search Marathon > > [VPN] VPN, Thank you so much for l the help. I am going to try the Trigonometry link above. I have downloaded it, and will be loading it on my 49G+. Where you said: > Cculus? > Search Marathon I am not sure exactly what you mean. Do you mean run a search on www.hpcc.org for Marathon or do you mean to run a search on any === Subject: Re: HP 49G+ Questions Search on hpcc. > Try this one: > http://www.hpcc.org/details.php?id=5103 > and will later be using it in multiple levels of Cculus classes as > > [VPN] > VPN, > Thank you so much for l the help. I am going to try the > Trigonometry link above. I have downloaded it, and will be loading it > on my 49G+. Where you said: > Cculus? > Search Marathon > I am not sure exactly what you mean. Do you mean run a search on > www.hpcc.org for Marathon or do you mean to run a search on any > === Subject: Re: HP 49G+ Questions > Ahaa! > Try this one: > http://www.hpcc.org/details.php?id=5103 > and will later be using it in multiple levels of Cculus classes as > Cculus? > Search Marathon Where the heck is Nick?!?! I've completely lost contact with him :( -- === Subject: Open Source for HP49G+ I've got an idea. Besides the obvious hardware problems I think everybody is forgetting one (very) important thing. We have a 75Mhz processor and we've found that it is possible to run native ARM asm programs! Even more, we have the possibility of adding an sdcard which enable us to have more memory then on any other cculator. So, we have something like a sml pda on our hands.... Now imagine that HP gives us the necessary documentation to write native arm programs for the hp49g+.... programs that run, not in emulated Saturn processor but,at 75 Mhz.... See the possibilities... The only problem, so far, is that not everybody knows how to write in ARM asm (me, for example) so if HP provide us with a tool (something like HPPasc) to simplify the process then they would have something that makes the competition truly vanish... And since HP doesn«t provide any programs for their ccs (at least no good ones I'm aware of) and they ways depended on users to make the cculator part (at least to compensate for their poor quity materi, but that's another story) Power to the people :) === Subject: Re: Open Source for HP49G+ Actuly, the ARM processor is quite a standard platform used in many embedded devices. There are ready a lot of tools out for it, including GCC for ARM. This means you could code in C and get ARM output in binary format. We currently know how to get into ARM mode and run short code snippets in it. The problem now is making this process more efficient/automated and discovering the memory locations and such for accessing the cculator's functions directly -- writing to the screen and such, instead of processing data in ARM mode then returning the result to the Saturn emulator for it to display -- which is what Cyrille's ARM example does. I am not sure just how efficient/inefficient GCC compiled code would be for the cculator, but I imagine it would be quite good for medium programs. > Now imagine that HP gives us the necessary documentation to write native > arm programs for the hp49g+.... programs that run, not in emulated Saturn > processor but,at 75 Mhz.... See the possibilities... The only problem, so > far, is that not everybody knows how to write in ARM asm (me, for example) > so if HP provide us with a tool (something like HPPasc) to simplify the > process then they would have something that makes the competition truly > vanish... === Subject: Symbolic substitution (hp49g+)? Is there a built in method to join multiple equations (including more than two) connected by a common variable? For example, given two equations: A = B * C; and D = B / E I'm seeking a result that would take the form: D = (A / C) / E (Parenthesis for emphasis only, not needed in actu result). In rpn: [ A=B*C D=B/E] [D] doesn't seem to do it for me with ISOL, SOLV, or EV Thank you for your help. === Subject: Re: Symbolic substitution (hp49g+)? > Is there a built in method to join multiple equations (including more > than two) connected by a common variable? > For example, given two equations: A = B * C; and D = B / E > I'm seeking a result that would take the form: D = (A / C) / E > (Parenthesis for emphasis only, not needed in actu result). > In rpn: > [ A=B*C D=B/E] [D] > doesn't seem to do it for me with ISOL, SOLV, or EV > Thank you for your help. Starting with Stack: 'A=B*C' 'D=B/E' 'B' ISOL [ENTER] 'A=B*C' 'B=E*D' SUBST [ENTER] 'A=E*D*C' 'D' ISOL [ENTER] 'D=A/(E*C)' === Subject: new 49G+ direct from HP has SN CN333... Argh! My cat yacked on my 49G, ruining it. (Stomach acid does nasty things to circuit board traces.) So I ordered a 49G+ direct from HP Shopping, thinking that would minimize my chances of getting an old seri number with a crap keyboard. WRONG! Ordered last Wed., arrived 20 minutes ago. SN is CN3330..., and the keyboard is indeed crap. Loud, hollow-sounding (can you spell CHEAP?), and missing keystrokes, just as many here have encountered. I cled the INVENT number, and when I finly got a human they had me cl instead the HP Shopping number: 1-888-999-4747. Save yourself the trouble of deing with that stupid voice menu system and cl the 888 number first! Got a human right away, he asked for my order number and what was wrong with the cculator, and he set me up with an exchange -- no problem, very courteous. They're charging my credit card for a new purchase and will refund the old one when the broken cculator gets back to them. They pay the return and new shipping. Seems reasonable. I only hope the new cculator's keyboard is likewise reasonable... --Marc Murison === Subject: Re: new 49G+ direct from HP has SN CN333... I only hope the new cculator's > keyboard is likewise reasonable... I got a CN352... seri number as a replacement and it's just fine and dandy - no missed keystrokes and solider. - Ed === Subject: Can the new HP33S isolate a variable? Anyone have the new cculator? I looked on the data sheet, I didn't see where it said it could isolate a variable. Can anyone confirm this? === Subject: Re: Can the new HP33S isolate a variable? It appears to behave just like the 32sii: it will solve for any variable in an equation . However, when you say isolate it sound like you want it to do symbolic gebra? (It won't do that). > Anyone have the new cculator? > I looked on the data sheet, I didn't see where it said it could isolate a variable. > Can anyone confirm this? === Subject: Re: HP51GX sold out > Hp has a smash hit on its hands with the new HP51GX scientific > cculator. The slimer cousin of the HP49G+ is virtuly sold out > after less than two weeks in stores, with nearly 100,000 snapped up. April 1st is a little early this year, eh? -- at: home: www.smith-house.org | http://www.smith-house.org/books/list.html === === Subject: Re: Relevance of RPN (please, no holy war...) > Is RPN still relevant (outside of c.s.hp48 and other core HP fanatics)? RPN is simple, regular, predictable, straightforward, and extensible - l things that gebraic is not. Even today, different ccs have different interpretations for gebraic notation, and it is still very rare to find a cc that can rely display gebraic notation - rather than some bowdlerized, computerized, ascii-ized representation hack. The HP-48 series is one of the ones that _can_ display gebraic correctly. gebraic notation is just an extreme example of baby duck syndrome. The system learned first is ways easier because people 't care to invest effort in learning another. If just one generation learned RPN first, gebraic would be quickly relegated to the dustbin of history it so richly deserves. -- at: home: www.smith-house.org | http://www.smith-house.org/books/list.html === Subject: Re: Relevance of RPN (please, no holy war...) > Finly, 'RPN has many advantages over AOS' needs some heavy quification. > Undoubtely, 4-level stack RPN was a marvel when compared to early TI > cculators, such as the SR-52, TI-58 or TI-59, where you felt cramped and > absolutely lost as to just where you are in the middle of a many-parenthesis > expression. I think most RPN fans ways have in mind those kind of cculators > when they think 'gebraic'. > But next to more advanced pocket computers and cculators of the time, such as > the many worthwhile SHARPs, classic 4-level RPN just doesn't hold a candle. > Not when you can type and see your full gebraic equation l at once on the > display, written on the display as it is in the book, with named variables if > necessary, and can go forth and back inserting, deleting or changing anything > before you even press ENTER. Then after the result has been computed you can > recl the whole expression to the display again, to edit or recompute with > different vues. And if you think it merits being considered a program, just > stick a line number in front of it, press ENTER, and there you are. > Only sheer stubbornness can claim than 4-level RPN is better than such a > typic, modern gebraic system. And of course, if you take on someone having > such a SHARP and try to proselitize with your 4-level RPN cculator, he or she > will merely think you're pathetic at best, and nuts at worse. Of course, SHARP > is just an example, the same can be argued for the CC mode of the 71B, which > has the added advantage of showing intermediate results and is capable of > backstepping a cculation one step at a time (but has the serious disadvantages > of a 21-character display and no multicharacter names for variables). I think there is a difference in outlook going on here. When I hear people speaking of entering a 'full gebraic equation' generly they are not 'cculating' (as in seeking a num answer) so much as programming or performing computer gebra. For students that is fine so far as it goes. However as a cculating device I think it is arguable that the 4-level stack is just about as good as it gets. If a problem gets much more complicated than that then I either use a spreadsheet, or I write a program (function) to evuate it. In practice such complexity happens rarely and when it does either a custom program or commerci software is the answer. Anyway that's my 2-bits ... and if, after the above, I've earned the 'stubborn' moniker then I'll wear the label proudly. === Subject: Re: Relevance of RPN (please, no holy war...) X-Mimeole: Produced By Microsoft MimeOLE V5.50.4910.0300 X-Msmail-Priority: Norm As a working engineer, when I am using a cculator I usuly have a lot of numbers in front of me generated from other sources that I want to do some multistep cculation with (short of importing the whole lot into a spreadsheet). The postfix approach of RPN lets me enter sever numbers and then decide what to do with them. I 't have a written equation in front of me or maybe even a firm idea what operations I need to perform. This is more compatible with the way I think ( emphasis on I, everyone is different ). With gebraic entry, there is a certain amount of pre-planning invovled that I have never mastered. Now in a learning situation, perhaps it makes much more sense to use gebraic or an Equation Writer type of entry format to better follow textbook examples. > Is RPN still relevant (outside of c.s.hp48 and other core HP fanatics)? > This isn't meant to incite a flame war. I've been thinking about this, > and have grown away from RPN and the HP48 I've used since my > engineering undergrad days. > 1) It's obvious to any programmer why RPN was conceived, and how > postfix notation makes stack-based processing simpler and more > efficient - for the *computer*. But as humans, we are concerned with > making things simpler for *humans. In any case, modern computers care > little about how data is entered. > 2) In my graduate programming courses, we've studied the advantages of > taking advantage of the user's context and existing knowledge, rather > than requiring him/her to learn a new UI. Buttons, windows, icons > behave (somewhat) similarly across platforms and applications these > days. Developers who stick to the prevailing standards find their > programs are learned, mastered, and _used_ much more than arcane > applications. This is not (as we hardcore geeks like to imagine) a > concession that users are soooo dumb. It's the way people learn > things. Gas pe on the right, QWERTY keyboard, etc. It's not just > marketing - transfer of training is a very re and important effect. > 3) In my graduate human factors courses, we've seen how verb > (language processing) skills are by far the most ingrained in the human > mind. When people say one plus one, they literly think one plus > one. You are fighting a very re and measurable part of your brain > power to think one plus one and write one one plus. The fact that > one one plus is natur to us is because we have *learned* that > language. But there is nothing inherent about it (unless some human > language I 't know of uses postfix). The idea that RPN better > reflects the way we write simple arithmetic (one over one, plus sign, > line for equs) is likely fse. Ask anyone to *read* that arithmetic > to you, and you're back in the strong verb mode. (It might be > different if the display actuly showed the numbers one over the > other, like arithmetic on a page, but that would take some study). > 4) The last argument I know of for RPN is that it is faster to enter > (quite true for experienced users) and that you can see intermediate > results (so true). Faster entry does not mean higher accuracy or ease > of use. For example, we used to program using variable names like SC, > but now, anyone would require you to spell out screenColor. (We can > similarly debate UNIX regular expressions vs. a GUI Find box) The > lesson over time was that cutting out typing makes for logic errors > and vastly increased training requirements. > 5) Finly, we have to admit that RPN as a barrier to entry to > beginning users has a certain macho appe. Maybe not to _you_ ;-), but > I think you've probably seen this. (Re men code in FORTRAN, we used > to say). > So, I have next to me a PDA that cost $120, with a 320x320 full-color > could easily have similar specs (and, hopefully, a decent keyboard!). > The CPU could handle whatever wacky entry method we dream up. Why still > use RPN? > This is meant as a reasoned discussion, not a geeked-out flame. My HP49 > is firmly stuck in RPN mode, in case you're wondering. The question is, > is there any vue added to making new users learn RPN, when they could > arguably be more productive writing equations just as they do on paper > (my arguments above)? I'm especily curious if anyone knows of actu > studies, and not just anectdot evidence (i.e. RPN is faster for me, > so shut up). === Subject: Re: Relevance of RPN (please, no holy war...) ># that I believe it may be relevant in other areas than cculators. For >example, ># learning the concepts of RPN on a cculator helped me greatly in learning >how ># to make estimates and to improve my ment cculating speed overl. >How? As one example, learning RPN in junior high school emphasized grouping and subtots. Especily with the 3 stack models, which only displayed the current tot on screen and you had to remember the other 2 stack tots (or at least how many of the stacks were in use so that they wouldn't roll out of the stack!) Once it becomes natur to reorganize what you see visuly into something that cculates more simply, it becomes much simpler to do estimates and ment arithmetic. I am sure this comes naturly to those that are adept at this sort of thing, but for me RPN was a fun way to learn about the commutative and associative properties :) As a young student first learning math, it was helpful for me to see on a cculator how many ways the following equation could be entered : 2+3*6+8 One possible way is the same order it is written: 2 ENTER 3 ENTER 6 * 8 + + Or regrouping: 3 ENTER 6 * 8 + 2 + which saves a keystroke and one precious level of stack space! Correct understanding of operation hierarchy is required to come up with the answer 28, rather than 38 when e left to right. I won't go into the argument over which answer is correct :) Of course this is a trivi example, but doing these types of manipulations mently are of great importance if you want to estimate and do ment of numbers, checking for easy to cculate sequences and groups, etc. === Subject: Re: URPL code need help for speed charset=iso-8859-1 Giancarlo schrieb ... > It works, the only thing is that it is very slow. Any propos for > improving speed? I 't need the code, maybe some suggestion only: > %%HP: T(2)A(R)F(.); > Ç 0. 'A' STO C->PX 1. ->LIST > WHILE OBJ-> DUP > REPEAT { # 0d # 0d } PVIEW 1. - ->LIST 'LST' > STO B->R { 1. 0. } { 0. 1. } { -1. 0. } { 0. -1. } 5. ROLL 1. 4. > START SWAP OVER ADD R->B DUP > IF PIX? > THEN DROP > ELSE DUP PIXON 'A' INCR DROP LST SWAP 1. ->LIST + 'LST' STO > END > NEXT DROP LST > END DROP > È I think you should not use STO within the loop. The list objekt should be handled within the stack. === Subject: Re: URPL code need help for speed I could think that the code is efficient and nobody has nothing to add but I am too reist... ;-) Maybe we l had a busy code weekend... > gorithm is very simple. You input user coordinates of a pixel inside > the box, circle or anything you want to paint and then the program > recursively look for 'white' pixels. > It works, the only thing is that it is very slow. Any propos for > improving speed? I 't need the code, maybe some suggestion only: > %%HP: T(2)A(R)F(.); > Ç 0. 'A' STO C->PX 1. ->LIST > WHILE OBJ-> DUP > REPEAT { # 0d # 0d } PVIEW 1. - ->LIST 'LST' > STO B->R { 1. 0. } { 0. 1. } { -1. 0. } { 0. -1. } 5. ROLL 1. 4. > START SWAP OVER ADD R->B DUP > IF PIX? > THEN DROP > ELSE DUP PIXON 'A' INCR DROP LST SWAP 1. ->LIST + 'LST' STO > END > NEXT DROP LST > END DROP > È > Any help appreciated, > === Subject: Re: Questions about 48GX operation Your equations seem so simple that I wouldn't even worry about storing them. Just as long as you understand the materi, you should be able to cculate manuly without any trouble. Other than that, to store a program oject, for instance <> or an equation, `E=m*c^2` you put the object into level 2, and a new variable name in level 1, and push STO. That will put your object into a variable. To see the equation, just push its softkey. If what I am making makes no sense, then you should read the manu to understand the basic operations.... Good Luck. > I am taking a class about wireless communication this semester and want > to put an equation into the HP48GX I got last year for another > networking class. In that class the teacher beamed the equations > through IR and they got stored on it. I can enter the equations into > the cculator but I have no idea how to store it so I can cl it up > during an exam. > The equation is pretty basic for sign energy per noise bit something > like this: > Eb/No = S dBW - 10 log R -10 log k - 10 log T > that is in decibel notation. > There is so Eb/No = S/R / No = S / kTR > There are a few other equations which would be nice to have stored but > I 't know how to store them into memory once they are entered. so, > should I just give the ratio Eb/No a name or leave it as is? > Its hard to say since I haven't seen the exam yet and 't know which > parts of the equations will be known. > Joshua === Subject: Re: Questions about 48GX operation > variable. To see the equation, just push its softkey. > Sorry, I ment to sat, to run the program, push its soft key. To see the equation of program, push righshift and the softkey... === Subject: Replacement 49g+ from Samson Cables working flawlessly Recieved the replacement 49g+ today and everything seems to be working great. Ser. # CN4020 with the latest ROM software. Keyboard is rely improved. Much more solid feel to the keys.They 't feel hollow like on the other cculator. Have tried, but been unable to to get the keys to miss. If everything continues to hold up I am going to be a very happy camper with this one. Much, much better than the 49g. === Subject: writing user RPL on my computer how do I write a user RPL program on my computer so I can transfer it to my cculator? using windows notepad doesn't seem to work. I couldn't figure out how to put Cyrille's mandelbrot program on my 49g+ === Subject: Re: writing user RPL on my computer > how do I write a user RPL program on my computer so I can transfer it > to my cculator? using windows notepad doesn't seem to work. I > couldn't figure out how to put Cyrille's mandelbrot program on my 49g+ %%HP: T(3)A(R)F(.); << D9D20430F2CCD20212008FB9760808F70E00F14D29E4192195E4293195E20070A1E30080A1 E 10C60A3E2920400E4 4640A1E3930500E54650A1E500E480E1090E53EA00000AC5004440E2930300E4002780E3C530 A1E3003880E100665 2E1FFFFFA186961D5E10066C3E86961C5E0F18DB8E86961D5E1006683E86961C5E0F18DB8E07 1341F00108340E000 80B60AE280B331FE9228319AAF0A7C151717FA6E55F1B0210034E922814413437FFFEFFFFAF7 31F481A60BAF23300 81BFAAF53128AC0B4406340010880BFF860D01564B421544A44580160B44AF23104A7107A6E5 6C161AF23166A7311 BA6E10B54931FF8018F2120096A6F8F2120096E6F8080310180B338FC77621361B021001448D 34150B2130 13 CHR SREPL DROP 10 CHR SREPL DROP H-> EV >> Notice the slashes in front of the <<, >> and /-> ? These tell conn4x to replace these characters with the speci HP equivlent. Transfer that program using con4x in ASCII mode. > === Subject: Re: Howto assign angle symbol to unshifted key > I had the same problem as an electric engineer. Not only would I need > polar notation, I would so need to quickly convert to rectangular (to > cculate resistance and reactance). I have created a custom menu as > follows: > {< DEG RAD CYLIN SPHERE RECT}, > where < is the angle symbol enclosed by speci parentheses. > I assigned this list to the 'CST variable. Once that is complete, enter > left-shift custom where the soft menus are available. This way l items > that would be needed to enter complex numbers are available with one press > of a function key. This has worked out rely well and is very efficient > as I can visuly see what key I need to press. This is where I started, but I quickly found out that there were a lot of menus I wanted to access, and having my angle key in one of them had me shifting back and forth an awful lot (which is precisely what I wanted to avoid). polar/rectangular degrees/radians thing. I think my favorite solution so far is Bill's little program. By assigning that program to a single user key, I can not only replace the angle key, but skip the parenthesis key so. Ruthlessly eliminating keystrokes, -Bear === Subject: 49g+ KEYBOARD question Ok, I know l people are tired about this subject.. but I want to say what I have to say... After playing with my NEW 49g+ I just see something that could explain why the new 49g+ are missing keypress... As l says it misses rely rarly keypress but when it does I think this is due to the Software. My tough on this subect is based on the fact that if you press two keys simultanouly, they simply 't register.. So when the cc miss keypress, I think, that simply because we type too fast and we presse two key without the needed interv to make the two register... Anyway, we'll see that in some new rom release since this is probably due to the Saturn Emulation Part.. Fell free to argument on it.. but Please 't start with like the keybord is shit.. hp is.... We should need constructive comments =)) === Subject: Conn4x Xsrvr library port on 48GX (Warning) Because Conn4x has some nice features that aren't available with other communication packages, I've decided to use it with the 48GX. Of course, the 48GX doesn't have an Xmodem server built-in, so instling the XSrvr library adds a lot of functionity. The Conn4x help says to instl the library in port 0 or 1, and indeed that seems to work just fine. But I usuly keep port 1 merged, so it's not available as a port, and of vulnerable to memory loss than an independent port. So anyway, I decided to try the library in port 2. Bad move; trying Conn4x with the XSrvr library in port 2 causes the 48GX to hang, and the ON-C warmstart doesn't work. Forcing a warmstart with a paperclip in the hardware reset hole does work, sort of. It comes back to life with an unacknowledged arm warning, the flags, both system and user, changed (RCLF returns { # 2o # 3o }), and last stack and command line saves disabled. I've never had a warmstart do those things, so I surmise that Conn4x made the changes before hanging, and wasn't able to restore various settings. I 't get a TTRM, and the home directory seems to be intact, and I'm able to, as far as I've noticed, get everything back to norm by doing an doing an ACKL, {#0 #0} STOF, renabling the saves, and using the flag browser to set the system flags to my preferences. It would be nice if the library could be used in port 2 or higher, but if that's not feasible, perhaps it could handle a mis-instlation a bit more gracefully, and I think that at least the instlation instructions should have a strong warning against trying anything except port 0 or 1. -- === Subject: integrability of a mathematic function Is there any program available for hp48 to check the integrability of an indefinite integr? If i enter sin(x)/x the program should tell whether it can be integrated or not. I have read somewhere that there is an gorithm for finding the integrability of any mathematic function. I want to know more aboutit. Please help. === Subject: Factoring using HP 48g? I know on the TI89's there is a built-in function for factoring polynomis. I was wondering if there was a similar built-in function for the hp48g or if there was one i could download. === Subject: HP-33S could have been nice if... ...you could see the fricking decim point! This is such a disappointment. It's VERY difficult to see, even when set to display a 'comma'. My 32SII's decim point is very easy to see. Same with my 48SX and GX. l my other HP's have had easy-to-see decim points. What in the HELL is with HP, these days? Do they do Beta testing -- and if so, listen to the feedback? Bob === === > 13 CHR SREPL DROP 10 CHR SREPL DROP H-> EV >> === Subject: Re: test >can i get this on the HP48G+ ? Yes. But you'll run out of memory rely fast. . . each is about 13K. . . (so remember to snip off unnececary stuff. . . i deleted about 150 lines of the code off the end =) TW === === Subject: German User Guides for HP 49G can someone tell me where I can get GERMAN user guides (basic, advanced, pocket) for the HP 49G in digit version to download, or can someone send it to my Email-address.. On www.hpcc.org I didn't find one. Thank you in advance! === Subject: Re: Directions for Using GNU Binutils for ARM9 Programming for 49g+ in Linux > MOV R2, #0x00001200 @ make R2 = 0x00001200 > ADD R2, R2, #0x00000034 @ add 0x00000034 to R2, making it 0x00001234 > Hey Matt, we were tking before on IRC about how to load large vues > into a register. You were wondering why MOV R2, #0x01234567 wouldn't > work. If you look up the ARM reference sheets for the instructions, you > will see that the 'immediate vue' field is only 8 bits wide. You can > instead specify a shifted register -- the assembly instruction has other > placeholders for shifting so it can hold more info that way. exactly, the constants in MOV can only have 8 bits of consecutive data (knowing that bit 0 is consecutive to bit 31...) if you want to load complex constantes, you need to do something like the add you are doing above, or a LDR as in: label: DD #12345678 code.... LDR R0, [label] ; this is in fact encoded as: LRD R0, [PC, -displacement to label] > Here's an example: > ADD R2, R3, R4,LSL#4 @ R2=R3+(R4 << 4) > The machine code format for this type of add instruction (32 bits, big > endian) is: 1110 000 0101 0 0011 0010 00100 00 0 0100 or 0xE0832204 > 1110 - condition (ways execute) > 000 - no re definition, basicly a way for the processor to > - recognize what type of instruction this is: > 0101 - opcode, 'ADD' > 0 - 'S' flag, if equ to '1', sets the flags after this instruction > - would be set to '1' if instruction were 'ADDS' > 0011 - Rn -- the 2nd parameter > 0010 - Rd -- the 1st parameter > 00100 - shift amount (shifts by 4) -- so you can shift up to 31 > 01 - type of shift, 01 is LSR (logic shift right), 00 is LSL, 10 is > - ASR (arithmetric SR - preserves sign), 11 is ASL (same as LSL) > 0 - same thing as the 2nd field -- IDs the type of instruction > 0100 - Rm -- the 3rd parameter > An add instruction like this: > ADD R2, R2, #0x34 has a format of: [cond(4)],[001],[opcode(4)],[S(1)],[Rn(4)],[Rd(4)],[rotate(4)],[immediate(8) ] > As you can see, there's only room for 8 bits for the immediate address. > MOV R2, #0xAB > ADD R2, R2, R2,LSL#8 @ R2=R2+(R2<<8) or R2= 0xAB00 + 0x00AB = 0xABAB > Hope this helps. I like reading machine code. :D > On a side note, I modified the machine code from Cyrille's Mandelbrot > program so it runs 32 iterations instead of 256, and it runs faster, > though the pic is a bit rougher. Pretty cool. though I'm still > trying to figure out the gorithm he used. Seems he used fixed point > registers?