B219 No, I was actually emailed the link to the survey directly. When I registered my 49g+ on hp's website, I apparently indicated that I such. After your question, I actually went on the survey.com website to see if they had a list of surveys people could participate in but could not see that anywhere. I just checked my email and I haven't I don't have the original email here at work anymore but if I recall, the email came from: hp@survey.com so you could maybe try sending a message to that address indicating your desire to participate... Mike Mander posting-account=NOgskgwAAADQRHwXcTtjFlQ0oJGNFCij Arghh... forgot the email scramble, sorry... The survey email was sent from: hp AT survey DOT com Mike Mander X-RFC2646: Format=Flowed; Response 49G CAS problems C1) demanding mode changes interactively, C2) guessing the wrong variable that really needs purging, C3) failing even after you let it delete an existing variable, because another same-named variable exists in a higher directory, 49G EQW problems E1) Units are not supported 49G All other problems A1) X-RFC2646: Format=Flowed; Original I think it's the second one I heard runours that it means: Virtual Polish Notation others claim it's a new HELP system that connect directly to VPNs brain and PUSH the answer to your calcs stack as a text string Due to the format VPN thinks sometimes you get GROBs but mostly just Externals, so it's not that compatible.... [VPN] posting-account=vamiSQ0AAAB6ebA-YINuYo1FZ8MBY1RJ X-RFC2646: Format=Flowed; Original X I though that that command IS recursive that's why there are two of them one works from inside out and the other inwards just try the other command, please [VPN] I thought they were supposed to r=work recursively, too, but apparently not. Try the MATCH commands on 'A+B+C' non-recursively, as I did, and you will either get 'A+B ADD C' or 'A ADD B + C', depending on which one you use. Apparently the MATCH commands treats 'A+B+C' as '(A+B)+C', so that one MATCH command works only on the outermost (lowest priority) plus sign and the other only on the innermost (highest priority) plus sign, but neither does more than one replacement per execution. Actually, when one thinks about it, that is the way they ought to work. There may be times when you only want one replacement, but when you do want recursive replacements you can easily program them as I have done. The MATCH commands are HP48 commands, and the hp48 has a marvelously designed operating system. X-RFC2646: Format=Flowed; Original X *********************** I would still do: as it is more general or the SREPL (again --. as the replacement) *********************** X *********************** I tried and the vMATCH seems to change only one at a time while you get more via ^MATCH, but even then at least two runs are needed *********************** Well, there are two commands for a reason. [VPN] Visualize an algebraic as defining an inverted tree (or root) structure for the order in which operations are to be performed, with bottom operations taking precedence over others on the same upward branching. Then the uparrow MATCH operation acts on all of the lowest instances of the particular subalgebraic and the downarrow MATCH on all the highest. If there are more than one lowest (or highest) of a particular type, all of then are done in a single operation. It is only when one of that type has priority over another that one is not done. Short answer is: I don't know. I've tried a couple of things but I can't cause an Undefined name error to occur. I tested it with a simple equation 'X+D(Z)|(Z=1)' etc. with D = { 1. 2. 3. } so my EQ variable is: { 'X+D(Z)|(Z=1)' 'X+D(Z)|(Z=2)' 'X+D(Z)|(Z=3)' } and then pressed APPS, 1 (Plot functions), 2 (Plot window), pressed ERASE and then DRAW from the menu using the default parameters. This gave me a graph with 3 parallel lines as you would expect. Can you try this and see that you get the same, please? At least then we know that it is not something fundamentally wrong with the technique or the setup. My flag settings are: { # 8657110000d # 0d # 4504836611506176d # 0d } -- Bruce Horrocks Surrey England I don't have an APPS button on my 48g+. You are probably using a 49g+, correct? Has anyone tried this method on a Hp48 and made it to work? Sorry, I'm not looking closely enough - that plus sign always makes me think of the 49G+ despite owning a 48G+! Anyway, it works fine on the 48: disregard the APPS menu bit and simply do RS-PLOT to display the plot menu, then press the ERASE and DRAW soft keys as before. -- Bruce Horrocks Surrey England While Boris' suggestion is quite elegant, I notice that it slows up the graphing by the indirect referencing of the parameter values. The following method avoids the slowing of the graphs and, once the program is entered and saved is easy to use: This assumes that the algebraic, with parameters, is stored in 'eq', note the lower case to distinguish it from reserved variable 'EQ'. And that each parameter value or list of values is saved under the name of that parameter, with all lists of the same size (number of elements). It also assumes that the graph type (FUNCTION, CONIC, etc.) and other graphing settings have already been made. Program: (Note the header %%HP: T(3)A(R)F(.); for copying to hp49+s %%HP: T(3)A(R)F(.); << eq @ recalls equation 'eq' to stack DO { '&x+&y' '&x ADD &y' } |^MATCH @ |^ is an uparrow character UNTIL NOT END @ DO loop replaces each + with an ADD EVAL @ replaces parameters in equation on stack with values. @ If any lists were used, creates a list of equations. @ Errors out if lists of different sizes were used. STEQ @ stores equation or list of equations in 'EQ' ERASE @ erases any prior graphs DRAX @ draws coordinates DRAW @ draws the equations stored in 'EQ' PICTURE @ Enters the usual graph editing mode Now save the program as, say, 'draw'. Example: 'A*X^2+B^X+C' 'eq' STO creates the equation to be graphed 2 'A' STO { 3 4 5 } 'B' STO { 0 1 6 } 'C' STO sets the parameter values or value lists. Note lists are all of the same size (number of items). Now executing the 'draw' program creates a list of equations { '2*X^2+3^X+0' '2*X^2+4^X+1' '2*X^2+5^X+6' }, stores the list in 'EQ', graphs the result and leaves you in the usual graphic mode. After exiting the graphic mode, you can change the parameter values and lists any way you like (as long as each parameter has a numerical value or value list stored under its name and all lists are the same size) and executing 'draw' again will do the rest. If you plan to do this sort of thing often, some sort of program like this is a great labor saver. You might want to save the program in your home directory or attach it to a key (using Keyman, for example) or put it in a library of your own if you want to keep using it. posting-account=p96ZUA0AAACXwuRub9Hyvsg2HJdsq2Lr You cannot use the alarm feature of the 49g+. Delete all alarms. And cross your fingers for the next ROM release (which hopefully fixes the alarms). But it appears unlikely . . . with they if So how did these differ from the GX? X-RFC2646: Format=Flowed; Original Calculator Comparison 5 http://www.hpcalc.org/details.php?id=3012 and Calculator Command Comparison Sheet 1.19-6 http://www.hpcalc.org/details.php?id=5265 [VPN] posting-account=p96ZUA0AAACXwuRub9Hyvsg2HJdsq2Lr Why can't HP make a 49g+ with the 48-series keys? My 48gx is over 10 years old and I never had a keyboard problem! X-RFC2646: Format=Flowed; Original Because the new HP has no respect for the customer.. they've entered the new age of confidence trickster business practices. They had a go at making a calculator that was attractive in some ways... built like a piece of crap in other ways... hoped that the market would go for it.... and were wrong... FWIF It's also possible that the old HP calculators were made so well that the market would simply not pay enough for them to make the business of building them a paying proposition. I'm sure that 14 years ago people didn't complain about lousy HP Junk... but many many people voted with their wallets by buying a cheaper calculator that did nearly as much. The problem is that HP has tried to invent the 6 month life, disposable calculator for 200 bucks... I think not... Al... posting-account=p96ZUA0AAACXwuRub9Hyvsg2HJdsq2Lr OK you're right, but anyhow, this is what I'd like to see: 1) Fix all the bugs in ROM 1.23 (especially alarms, but also the CAS (digamma, etc.) 2) Go back to the 48-style keys. The business model right now (make crap that lasts 6 months) will not work in the long term. I bought my first HP calc in 1978, and have been a loyal HP user ever since. However, my kids won't buy HP calcs, if quality stays this way. Why should they? HP needs to build on their base. Continue to make quality products. It's the only way to stay alive. 10 the new making a crap in wrong... that of people with their disposable posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo Your slow , clunky was the CPU , not op sys , don't give FettedKernel credit for hardware he did not do .. MetaKernel is slower and clunkier and more obtuse , all things considered than a 48 op sys ! Why bother with uclinux? The Samsung ARM9 has a MMU, so it can run a full linux (not likely with Jean-Yves I doubt even HP knows it :) Jean-Yves x-mimeole: Produced By Microsoft MimeOLE V6.00.2900.2180 X-RFC2646: Format=Flowed; Response Sorry, I should have said HP or whoever did it for HP. I didn't mean to leave anybody out... ;) Claudio posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo It's also fast programming to create your app using stack for it's handy then if the HLL is smart and simple as FORTH is , it automatically assigns these PUSHs/POPs to locals ! Slick eh .... Your Source looks great ! No more annoting what this or that stack value is . And it does it autmatically ! I have been revu'ing Forth as an OP SYS , and figured how to make it multi-tasking and secure w/o using Intels hardware fault system features. Forth pays no speed penalty , for it in-lines in seconds , both automatically and interactive . It can also compile script files on the fly . When all user threads come from script files , instantly compiled , there is no access to anything fragile . Your app cannot MEM-ALLOC , it can't write anywhere it chooses , it can't access anything but what the OP SYS allows it . All the biggies do this , but they have loopholes for the bloat and convoluted kernel threads using C+ . Linux hacked into Forth is 100K instead of 1500 ! Text and images and non-code has no place in this discussion of safety , so i'll guess LINUX 2.4 has about 1 megabyte of code . Forth at 1/10 the size ,, for MODULAR !! But at run time Forth can in-line all critical device drivers without any complicated methods ! So it takes less than a second to optimize And it takes one day to figure Forth !! Try that in C+ !!! you're crazy man Don't think so. Most CAS (if not all) are written in P-code for a reason. When you actually compare the Erable CAS and its equivalent XCAS written in C++, Erable performs very very well (if not better all things considered) Many of the most used Saturn Assembly subroutine have been rewritten in ARM already. Jean-Yves Weren't you the one who told me a rewritten Factorial function would only be ~30% faster? ;) Seriously, I've never tried to write a CAS. I'm just extrapolating from all the other benchmarks I've tried. P-Code? Do you mean Pascal Intermediate Code? Some quick googling says that Mathematica is implemented in an object-oriented variant of C, and as you said XCAS is in C++. I'm not sure what you mean. I haven't tried XCAS, nor have I tried writing one. I'll take your word for it. best, Al If it's purely about dealing with long numbers then a 100% ARM ML would see a great improvement. When it's about a full CAS , managing objects, polynomial etc... I doubt the improvement would be that extraordinary. By P-CODE, I mean using a dedicated interpreter/language Of course you can rewrite everything in any language, but C++ CAS would used) with 2MB of ROM.. Hopefully those limits will disappear at some point where the technology is mature enough. JY posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo OBJECT ORIENTED is at odds with MODULAR programming . Thus it is just another way of slowing and obtusing s/w . The support is a claim of safety .. If you write good code it is safe code . JYA uses the term P code for it's WIZARD factor posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo A proper USERRPL can on-the-fly complile to SYSRPL . I can write a script type USERRPL that compiles on-the fly in < 1 second . The speed is for the modularity , C+ would not be able to do this . The idea is to write a direct one to one with SYSRPL , except USERRPL has sefety that may take more time to run . And many things in USERRPL that you may see as prolific , would drop out as run time , they are there only to remind you what to do . And if you used the stack to write your app the SYSRPL would assign all your stack stuff to locals AND rewrite your source showing the LOCALS ! It clarified your src !! ha ha ha Does that shake you in your boots JYA ? But when you have debugged (and the op sys has debugged by running it once) then it needs not the safety slow down , so it nixes them and runs as fast as SYSRPL . Or another way of looking at it is all scripts are compiled to genuine SYSRPL ! In Forth there is an analog . All apps are a list of addresses or scripts There is NO CODE in your app !!!! But at run time < 1% of the kernel thread is using these addresses !! No slow down !! At the mid level in Forth there is direct pointers , so no need to interpret any code at run time . Now try to make a Virus using Forth !! You have no access to anything access to what Forth wants a user to have ! I hate SYNTAX errors !! Why can't the stupid op sys figure whats needed and add the damnned thing and shut up and JUST DO IT How many deaf OP SYS syntaax errors can you think of that are unnecessary ?!! We need not to TYPE , we need the OP SYS to do that work for us .. Here the =LIST reminds us that level 1 stack must be a list AND if it is not , it skips the whole proggy and if USER flag 10 is set , run the rest of the stream . But for programmers trying to make USER proggys WIZARD food ..... That's weird...it showed up in Google groups, but not at the comp.sys.hp48 site. Bob X-RFC2646: Format=Flowed; Original No, unnamed entries = internal entries, and therefore they might (and they do) change from rom to rom. To support all calculators you should stick to the named and documented functions that are in extable. Think about this: a user programming in SysRPL would have to use xREPL, right? So why should your converter come up with something better? Your converter will be good, but don't try to do it even better than a (human) programmer or you will never finish it :-). Claudio they to Dang. I was afraid that would be the case. Unfortunately, the named commands in extable don't even begin to cover all the commands beyond the arg checking, as I have found. I'll probably have to make another version... Really? I try to avoid using UserRPL commands completely when I'm writing SysRPL programs. I always use the appropriate internal version of xREPL. For example, Repl works much faster on grobs, and system binaries are *much* smaller than user ones. Bob X-RFC2646: Format=Flowed; Original X What about, Bob, if you change the wordsize? [VPN] I can do that? Please elaborate! :-) Bob X-RFC2646: Format=Flowed; Original BASE eg. [Left-Shift] [ 3 ] [NXT] and there thay are: STWS and RCWS that applies only to user binaries and saves memory naturally in nibble decrements while the wordsize can be anything from 64 to 1 bits [VPN] posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo I tried searching this NG with key = Xmodem Connectivity Kit which is found at the link below No match ... __________________________________________________________ I can't use EXTABLE . I dnloaded to 49G+ and it sits there , doin nothing .. I tried attaching it , i read the book . I turned 49 ON-OFF-ON ... still just sits there and does nothing . I did manage to COPY it to a dir and then it lit up the activity light , but still useless . Why don't these WIZARDS who do the extable , write some instructions ? I can make my own , but it would save some time if i could not reinvent the wheel . does colum swaps, deletes , moves .... I have 2 columns of #'s i want to swap . I have been searching for an editor to replace NOTEPAD for years , none available .... I did find an old one but it was too quirky in normal stuff , could not remember the odd named cmds , it was not intuitive enough . posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo HP was bought out . It is not the H.P. you know . It is an impossibility to turn a crony into a competitive business . To be inventive and competitive is a threat to their methods of gain ... posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo The excitement is gone ?.... The competition is gone . The normal free market competition is gone . Killed by the U.S. gov't in the name of safety , security , helping America ..... B.S. !! What it's doin is putting the best Americans out of work for the cronys and Affirm' Actions ... posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo And you sound just like the other crazy guys poluting the news group... Jean-Yves ---------------------------------------------------------------------- Not like polluting the world with your amatuerish ,METAKERNEL posting-account=fgGaTw0AAAD45Fx2_7bZCc6hcNf--wQo ha ha ha it does in the 48 .. After JYA got his hands on the 49 , everything is changed for the better ha ha ha ..... I have both ,and the 49 op sys and new improved cmds are dog droppings posting-account=9wLUeQ0AAACd9G5v3T6VPoZctF9pMIKW Watch out, guys.... The virus strikes back on 11.03.05..... Newsgroup full again of nonsense by our fellow We##y.... posting-account=iyqLeQ0AAACkE0etKTOSEgfyEmkmLQlQ I hate to bring up dead topics, but I had a similar problem. I went to math class, removed my precious HP49G+ from its case and (!!!) SAW A FLECK OF DUST!!! It really annoyed me. So, today, I did a little experiment. I found that by tapping my calculator very lightly on the palm of my hand on one side and then the adjacent side, I was able to wiggle the speck of dirt out of the calculator's screen. It was pretty neat. Tap lightly if this is still your problem! Good luck to everyone who has this problem. Has anyone tried the KEYTIME command to help with the keyboard problem. try using n= 300. or read this: ____________________________________________________________________ Hello the anti-bounce procedure of the HP49. a key bounce occurs when a key pressed down 'bounce' on the connection and cause the software to detect the key low twice instead of once. On the HP48, the problem was solved using a delay loop in the interrupt system. When a key was detected as down, the system waits 1/10 of a second, checking the key, in order to ensure that they where no bouncing. This is one of the reason why the HP48 was going slower when a key was pressed. on the HP49, the anti bounce system is different. When the system detects a key as down, it push the key in the keyboard buffer, and save the current time in a special variable. If the system detect the key low a second time, before pushing it in the if this equation is true, it's means that the key has been pressed twice. else, it means that the key just bounced (and the key is not pushed in the keyboard buffer). You can adjust this constant using KEYTIME (The argument given to keytime is a time delay in 1/8192 of second). Try for example to set KEYTIME to 4096 and to press the same key as fast as possible, you will see that you can not have more than 2 valid key per second. ----- Original Message ----- the with NEVER the with NEVER Holly Cat, Great idea. I'll give it a try. Unfortuately I'm beginning to suspect is is a physical attribute of the key not always making contact. I'm afraid to say it, but it may be this is as good as it gets. Tom by second, is as 's CNA 's CNA posting-account=iyqLeQ0AAACkE0etKTOSEgfyEmkmLQlQ My enter key seems broken. It clicks in several ways when you press it (depending on which angle you press it from). But for some reason, it functions a heck of a lot better! I found that it presses easier and is very accurate. X-RFC2646: Format=Flowed; Original ClockSpeed Adjustment Tools - Overclock your 49G+ 2.00 http://www.hpcalc.org/details.php?id=6081 [VPN] How old an iPaq? All recent (last three generations?) iPaqs use ARM compatible processors - required by MS. If this one is recent, it should run the current emulator fine... Checking it seems like this is recent. All current Pocket PC running Windows Mobile use ARM compatible processors - just load and run the current emulator. Pete M. Wilson Gamewood, Inc. wilsonpm@gamewood.net