C19 ==== Can someone please give me an explaination of the 49G's memory - namely the difference between directories, the VARS key and the CUSTOM key. But there are variables that have appeared that I did not directly create like CST, IOPAR, MODUL and VX. I take it that these are necessary and are created when you use or program the calculator (although I'm not sure how or when). I would like to put my custom programs in their own directory to stay organized. Do all those extra variables have to be in the directory that my programs are in? The manual is very limited in explaining what is going on. ==== Don't worry, be happy - it's fully automatic! Even if you remove them, they appear back! { DATA CALC REPORT } MENU will show a CUSTOM nebu when you press that key. The list above is stored in CST variable in your directory. VARS gives all the variables in the current directory. The huge PDF file http://www.hpcalc.org/details.php?id=3937 will give you an advanced example on using your own custom menus. > Help Can someone please give me an explaination of the 49G's memory - namely the > difference between directories, the VARS key and the CUSTOM key. > VARS. > But there are variables that have appeared that I did not directly create > like CST, IOPAR, MODUL and VX. I take it that these are necessary and are created when you use or program > the calculator (although I'm not sure how or when). I would like to put my custom programs in their own directory to stay > organized. Do all those extra variables have to be in the directory that my > programs are in? The manual is very limited in explaining what is going on. ==== I know its not the HP-58 but hasnt HP hinted at a new calculator sometime next year? What niche do yall think it will fit in? ==== > I know its not the HP-58 but hasnt HP hinted at a new calculator > sometime next year? What niche do yall think it will fit in? HP and Scholastic Present the 2003 Create-a-Calculator Contest: http://www.hp-calculators.com/contest/index.phtml ... Winning entries from grades K-5, 6-8 and 9-12 will earn... ==== > I was trying a few symbolic integration with my HP49G > and this one didn't work sin(ln(x)) the answer should be 1/2 *x (sin(ln(x))-cos(ln(x))) > (The IT89 was doing it easily!) I know this is an old post, but since the TI 89 can do this I wonder why the HP 49G can not do the subsitution itself ==== This was another example that I hoped that HP CAS would do all by itself. Parisse did a nice job, but not good enough to beat TI 89, which seems to use a simple, but efficient table eg. pattern matching and bring textbook answers. This is a weakness in the HP 49G CAS. Maybe RISCH should do risch and let the INTVX do a more traditional work. We have Flash to fill... Just an opinion/observation/question > >I was trying a few symbolic integration with my HP49G >and this one didn't work > >integrate (X/(1-X^4)^(1/2))dX . > >the answer is 1/2*sin^-1(X^2) . > Yes it can... just do it as you'd do it on paper. Put this on the stack: / X > | X > | ------------- dX > | (1-X^4)^(1/2) > / 0 Then X^2=SIN(U) SUBST EVAL Then SIN(U)=X^2 SUBST And that's it. -- > Beto > ==== The 49-rompointer B3 A8 (same as the 48-rompointer B3 A9, documented in the 48FAQ p.110, by JKH) computes the number of days of a month only from begin 1991 till end of 2090. On the other hand, Timeman shows that one can browse through the calendar from end of 1582 (the begin of the Gregorian calendar) till the year 9999. For instance, with Timeman you can have already now a look into the calendar of February 2100 in which some members of this NG may still be alive... February 2100 is the next cumbersome february which many people expect to have 29 days, but it doesn't. The last such february was in 1900. To present a correct calendar for the year 2100 we must have a program that computes the number of days of each month from 1593 till 9999 precisely. CHALLENGE. ite a shortest (a) 49-UsrRPL program (b) 49-SysRPL program, DOM say, whose input is a DATE from the interval 1583...9999, and whose output is the correct number of days of the month to which the input DATE belongs. No argument or input protection is demanded. IMPORANT: Your program must be independent on the date flag -42 ! CONDITIONS: Only supported pointers are allowed, no unnamed rompointers (which probably wouldn't be of any help anyway). Hint: It is possible to compute this number even without precise knowledge of the rules for leap years. Amazing, isn't it? :-) - Wolfgang may not be optiomal as a stand-alone program. Corresponding updates have been made this morning on my site. Ghost days in previous versions of Timeman like the 2/29/2100 were discovered independently by Prange and Otto Praxl. Although such ghost days were showing up in the calendar it was impossible to navigate into these days. ==== > Enter the matrix (A) and use EGV. You get the eigenvectors and the > corresponding eigenvalues. Then S^(-1)AS = D where D is a diagonal matrix > with the eigenvalues as diagonal elements and S is a matrix whose column > vectors are the eigenvectors of A in the same order as their corresponding > eigenvalues forming D. If the original array was n x n, you can enter {n n} > and use DIAG-> to transform the vector of eigenvalues to an actual matrix. > > Hi, > > Can you make an example :-)? Eg. you wanna diagonalize: > > 1 -3 > 4 0 > > Anders > Put the matrix to be diagonalized on the stack then execute << DUP EGV DROP SWAP OVER * SWAP / EXPAND >> ==== +----------------------------------------------+ | Sat, 17.5.03 10:01 p.m. +1200 (NZT) | +----------------------------------------------+ > is it a beta (you know with auto-off after prgm run;-), > or a production unit? > > The latter - s/n CN313 bought at educalc.net (not the same > as EduCALC ;) Good service indeed. I have the same from the same place. Excellent service I agree. Two things I was delighted about with the 12c: 1. My Black-Scholes ran in 8 seconds as against 12 seconds on the 12C gold. 2. The 12c has the *only* version of the HP TVM (on any HP Financial calc, including the 12C gold, HP49G, even the Hp Palmtops) which satisfies this absolutely trivial test: n=1 i=10 PV=1 PMT=1 Solve for FV =1.1 Resolve for PMT. Hehe. What do you expect? 0? Indeed and the 12c gives that. All other HP TVM give very small numbers, but not '0'. OK I know 1E-15 is effectively zero. But this shows ya the 12c lads got something, albeit almost presentational, right. > I'm preparing a review for the next Datafile, but if you > need some info before buying, just let me know. I look forward to your review! > PS: That bug was fixed ... unfortunately! It would be a > nice collector's item. :) But Jordi at least you have some extended precision dates on the back of yours. I still laugh when I see that. Oh - nice leather case too. The 12c has got something - personality. I actually do quite like it. -- Tony Hutchins Wellington New Zealand ==== Ok, finished reading the Nosy manual (I know - RTFM) and found that NosyHelp will check extable compatibility. Sure enuf, my extable was incompatible. Found the right one, and everything is hunky. ==== And keyman, libman, and OT49 as well... ( Wolfgang, if you are listening, you are a monster!! ;-) ) I've played a little with the emacs, looks awesome... development. The grunt-work wasn't hard, the user-interface is kicking my behind. I want Inform screens with checkboxes, but I want the defaults to be determined at runtime. Can't use INFORM because of the checkboxes and can't use InFormBuilder because of the runtime requirement. That's why I'm finally looking at SysRPL but it looks daunting... Anthony Hi, All the pro's in here use latest version # 19-6.... and recommend upgrade to all newbies and neophytes. I upgraded my 49G as well... and I am thankful to say the upgrade went perfectly. I have 2 questions at this time:- 1. Why does hpcalc.org still list a WARNING concerning the 19-6 ??... surely this must frighten the newbies and there seems to be no such need for caution or fear ??? 2. I have observed that the 19-6 has a few more system flags (which I would expect to be the case).... where can I go to read about these additional features of the 19-6 upgrade ???.. (I could not find it on hpcalc... sorry if i am missing the obvious) compared to some of the others on usenet. ==== Raymond Del Tondo schrieb > How about copying all files from the Emu48 directory to the other PC, too... There is a *.ini File which need to be taken into the copy-procedure. Search for emu48.ini and copy it, too and locate it to the new correct place on the new PC. ..Heiko ==== ones) that rely on the the number of days from the first of the given > month to the first of the following month fail in December 9999. I suppose that instead ...in December 9999. I should've itten ...for arguments in December 9999. But of course, barring some amazing breakthrough, none of us will be around in 9999, and I doubt that any 49G will be working then either. And of course, with the current operating system, you can't set the calculator's current date to 9999. -- ==== +----------------------------------------------+ | Sat, 24.5.03 3:08 p.m. +1200 (NZT) | +----------------------------------------------+ Hi Tony Hutchins ! result. > > I could only conclude then that there were severe limitations > with the 12C CPU compared to earlier HP Financial calcs, or > that somehow a lot of HP mathematical knowledge had been lost. Well, well, little did I know!! I just found this great http://mprc.pku.edu.cn/archta/codrefer/0405.pdf itten by Professor W. Kahan. He assited Roy Martin, Dennis Harms and Rich Carone in programming the 12C!! And he mentions that a log transformation was used, along with Newton on 22 November 1983 as well! What a piece of history that I never knew about all these years, but half suspected. -- Tony Hutchins Wellington New Zealand #255 The best way to keep one's word is not to give it. Napoleon ==== My 48SX gave up the ghost a few years ago... Now I wish I had it for some process timing applications. Is there a way to control relay(s) from a 49G over the serial port? Can it talk to Tek DMMs? How about doing both, at the same time? ==== > After i press eval, it comes. Eval Error: Non algebraic in expression > What is my mistake??? Try running CASCFG. ==== I was so tired of complaining about th faulty capacitor of my HP 49G China and getting no results that I decided to ask for it in HP Argentina. Once in my house, I opened the calculator softly. I took out the faulty capacitor, a 6.3 volt 1000uf one. I went to a electronic store and bought a replacement. The new one was a 16volt 1000uf capacitor. You might wonder why I put this instead of the original values??? HP 49G uses the batteries, each one of 1,5 volt. connected in series it is 4,5 volt. I electronics we use a security margen (don«t know how to ite it in english. sorry) of 40%. Then: 4,5 volt * 40% security = 6.3 Volt , yes! the same the original capacitor said. Well, you should know that the volts we read on a capacitor is the maximum difference of potencial it can support. But, why use a 6.3 volt capacitor if we use 4.5 volt??? The security margen is to increase the capacitor life, because they want last for ever. They work with a chemical reaction inside (the dielectric, and other stuff). The capacitor that comes with HP 49G from china has the right size values. But it is of really bad quality, not lasting even 2 years. When the voltage of the capacitor is higher it will last more. More working life. Ok. But isn«t the 16 volt 1000uf capacitor bigger than the original??? YEs, but i have manage to put it in with out any trouble. I put it beside the batteries compartment. I will upload the pictures soon. Finally I would like to thank Jean-Ives for regreting about his first I«m talking about, HP might have a strong pressure above him) and HP of USA and Argentina for giving no rational solution for my problem. Esteban Suarez Marenzi Pd: This is the best part: Hp wanted me to pay u$s 130.- for a replacement. I fixed the calculator myself with a cost of $ 0,6.- (pesos argentinos). In dollars it is u$s 0,2.- ==== > > Mon Dieu! The old HP-TI wars still rage, eh? ;-) > > BTW, substitution is a heuristic, so it may very well fail. But I do > agree that it should have been implemented. It's a question of time: there are so many tricks to integrate, yes one rule could be added, but then you will have someone askinbg for another rule, etc. and I don't have time to work on this now, I'm too busy with giac/xcas (which BTW does it). Everyone will soon be able to add its own rules since the CAS will be released under the LGPL (should be on September). Currently, you can enter SIN(LN(X)) INTVX 'LN(Xt)=Y' SUBST EVAL and you get the answer. Reply-To: Veli-Pekka Nousiainen ==== > > the answer should be 1/2 *x (sin(ln(x))-cos(ln(x))) I was trying to solve this equation for x, and it failed. Is there any way > to get an exact solution for this on the 49? No, numerical only, X=2.19328005074 ==== > > > I was trying a few symbolic integration with my HP49G > > and this one didn't work > > > > sin(ln(x)) > > > > the answer should be 1/2 *x (sin(ln(x))-cos(ln(x))) > > I was trying to solve this equation for x, and it failed. Is there any way > to get an exact solution for this on the 49? Which equation? I don't see an equation... -- Bhuvanesh ==== > > > > > I was trying a few symbolic integration with my HP49G > > > and this one didn't work > > > > > > sin(ln(x)) > > > > > > the answer should be 1/2 *x (sin(ln(x))-cos(ln(x))) > > I was trying to solve this equation for x, and it failed. Is there any way > to get an exact solution for this on the 49? Which equation? I don't see an equation... Sorry for the confusion, I meant 1/2 *x (sin(ln(x))-cos(ln(x))). And please don't give me, The TI-89 can do this one easily. :) ==== The HP49G can integrate anything, as long as you use pencil and paper to help in some cases! Stan >>I was trying a few symbolic integration with my HP49G >>and this one didn't work sin(ln(x)) the answer should be 1/2 *x (sin(ln(x))-cos(ln(x))) >>(The TI89 was doing it easily!) Yes, it can. It is done by IBP twice with the auxiliary function >being simply X 'SIN(LN(X))' >'X' IBP yields 'X*SIN(LN(X))' >'-COS(LN(X))' append another 'X and do another IBP which yields: 'X*SIN(LN(X))' >'-X*COS(LN(X))' >'-SIN(LN(X))' which means: 'SIN(LN(X))' = 'X*SIN(LN(X))' + '-X*COS(LN(X))' + '-SIN(LN(X))' Or, the solution is '(X*SIN(LN(X))-X*COS(LN(X)))/2' which >simplifies to the answer you gave. ``Put on your thinking cap.'' Randolph J. Herber, herber@dcdrjh.fnal.gov, +1 630 840 2966, CD/CDFTF PK-149F, >Mail Stop 318, Fermilab, Kirk & Pine Rds., PO Box 500, Batavia, IL 60510-0500, >USA. (Speaking for myself and not for US, US DOE, FNAL nor URA.) (Product, >trade, or service marks herein belong to their respective owners.) ==== compute this number even without precise knowledge of the rules for leap > years. Amazing, isn't it? :-) Yes, it is. I don't see how, which makes me suspect that I'm missing something obvious. Here's my UserRPL program. It seems to me like a brute-force way to get the answer, but I really don't see any shorter way to do it in UserRPL. Name: DOM Checksum: # 37AFh Bytes, including name: 210. Bytes, not including name: 202.5. Takes: Real number in date format to match state of flag -42. Returns: Zint, correct number of days in the month to which the given date belongs. %%HP: T(3)A(D)F(.); << DUP -42 FS? { FP 100 * } IFT IP R ->I { 4 6 9 11 } OVER CASE POS THEN 30 UNROT DROP2 END 2 =/ THEN 31 NIP END 100 * FP 10000 * DUP 400 MOD OVER 100 MOD XOR SWAP 4 MOD OR THEN 28 END 29 END >> -- Reply-To: raut@math.ful-berlin.de ==== > Hint: It is possible to > compute this number even without precise knowledge of the rules for leap > years. Amazing, isn't it? :-) > Yes, it is. I don't see how, which makes me suspect that I'm missing > something obvious. you do. Your nice UsrRPL-program contains the leap year rules. You will be surprised by solutions which don't explicitely refer to these rules. You have still another chance. Everybody (including the other poster on his Challenge) should ite the number of bytes and CRC of his program (without name). Not necessary to repeot the story on the DATE argument. - Wolfgang Reply-To: dominik@_DROP_THIS_science.uva.nl ==== Hi everyone, here is my entry, which does not implement the leap year rule itself, but just uses the fact that the HP knows it :-) 76 bytes (without name) Checksum #28Fh - Carsten << RCLF DATE ROT DDAYS -42 CF DATE SWAP DATE+ DUPDUP FLOOR 12 == -11.999999 1 IFTE + DDAYS SWAP STOF >> > The 49-rompointer B3 A8 (same as the 48-rompointer B3 A9, documented in > the 48FAQ p.110, by JKH) computes the number of days of a month only > from begin 1991 till end of 2090. On the other hand, Timeman shows that > one can browse through the calendar from end of 1582 (the begin of the > Gregorian calendar) till the year 9999. For instance, with Timeman you > can have already now a look into the calendar of February 2100 in which > some members of this NG may still be alive... > February 2100 is the next cumbersome february which many people expect > to have 29 days, but it doesn't. The last such february was in 1900. To > present a correct calendar for the year 2100 we must have a program that > computes the number of days of each month from 1593 till 9999 precisely. > CHALLENGE. ite a shortest (a) 49-UsrRPL program (b) 49-SysRPL program, > DOM say, whose input is a DATE from the interval 1583...9999, and whose > output is the correct number of days of the month to which the input > DATE belongs. No argument or input protection is demanded. IMPORANT: > Your program must be independent on the date flag -42 ! > CONDITIONS: Only supported pointers are allowed, no unnamed rompointers > (which probably wouldn't be of any help anyway). Hint: It is possible to > compute this number even without precise knowledge of the rules for leap > years. Amazing, isn't it? :-) > - Wolfgang > may not be optiomal as a stand-alone program. Corresponding updates have > been made this morning on my site. Ghost days in previous versions of > Timeman like the 2/29/2100 were discovered independently by Prange > and Otto Praxl. Although such ghost days were showing up in the calendar > it was impossible to navigate into these days. Reply-To: dominik@_DROP_THIS_science.uva.nl ==== I just see there is a possible problem with my program for dates in which the day number is high (29,30, or 31). Here is a modified version (115 bytes) which unfortunately wastes a lot of bytes testing for this case :-(. I guess this can till be shortened ... << RCLF DATE ROT DDAYS -42 CF DATE SWAP DATE+ DUP IP SWAP FP 100 * FP 1 + 100 / + @ fix high day numbers DUPDUP IP 12 == -11.999999 1 IFTE + DDAYS SWAP STOF >> - Carsten CD> Hi everyone, CD> here is my entry, which does not implement the leap year rule itself, CD> but just uses the fact that the HP knows it :-) CD> 76 bytes (without name) CD> Checksum #28Fh CD> - Carsten CD> << CD> RCLF CD> DATE ROT DDAYS CD> -42 CF CD> DATE SWAP DATE+ CD> DUPDUP CD> FLOOR 12 == -11.999999 1 IFTE + CD> DDAYS CD> SWAP STOF CD> >> > The 49-rompointer B3 A8 (same as the 48-rompointer B3 A9, documented in > the 48FAQ p.110, by JKH) computes the number of days of a month only > from begin 1991 till end of 2090. On the other hand, Timeman shows that > one can browse through the calendar from end of 1582 (the begin of the > Gregorian calendar) till the year 9999. For instance, with Timeman you > can have already now a look into the calendar of February 2100 in which > some members of this NG may still be alive... > February 2100 is the next cumbersome february which many people expect > to have 29 days, but it doesn't. The last such february was in 1900. To > present a correct calendar for the year 2100 we must have a program that > computes the number of days of each month from 1593 till 9999 precisely. > CHALLENGE. ite a shortest (a) 49-UsrRPL program (b) 49-SysRPL program, > DOM say, whose input is a DATE from the interval 1583...9999, and whose > output is the correct number of days of the month to which the input > DATE belongs. No argument or input protection is demanded. IMPORANT: > Your program must be independent on the date flag -42 ! > CONDITIONS: Only supported pointers are allowed, no unnamed rompointers > (which probably wouldn't be of any help anyway). Hint: It is possible to > compute this number even without precise knowledge of the rules for leap > years. Amazing, isn't it? :-) > - Wolfgang > may not be optiomal as a stand-alone program. Corresponding updates have > been made this morning on my site. Ghost days in previous versions of > Timeman like the 2/29/2100 were discovered independently by Prange > and Otto Praxl. Although such ghost days were showing up in the calendar > it was impossible to navigate into these days. ==== > > Hint: It is possible to > > compute this number even without precise knowledge of the rules for leap > > years. Amazing, isn't it? :-) > > Yes, it is. I don't see how, which makes me suspect that I'm missing > something obvious. you do. Your nice UsrRPL-program contains the leap year rules. You will > be surprised by solutions which don't explicitely refer to these rules. You have still another chance. Everybody (including the other poster on > his Challenge) should ite the number of bytes and CRC of his program > (without name). Not necessary to repeot the story on the DATE argument. Ok, I've got one that's 84.5 bytes with a checksum of # 12A7h. It usually returns a real, sometimes a zint; is that ok? As you can no doubt guess, it takes a completely different approach to the problem from my previously posted program. Pure UserRPL, and with a minor change (which adds to the size), it should also work on any 48 series. -- ==== I've been beating my head against this, but haven't heard the secret word... I have however significantly refined my original attempt: Ç DUP IF -42 FS? THEN FP 2 ALOG * END IP IF DUP 2 == THEN DROP 6 ALOG * { 4 400 100 } MOD NOT EVAL NOT OR AND 28 + ELSE 1 - 7 MOD 2 MOD 31 - NEG NIP END È I took out the range test on the year (not called for in the requirements), found out that IF THEN ENDs are smaller than IFTs, found out that when doing modulus math don't have to clip out the year completely. Exponent ALOGs are smaller than 1Ewhatever. Got tricky applying the leap year rule. Enter in exact mode. Weighs in at 158.5 and #1D29h Anthony Mohrenweiser ==== > The 49-rompointer B3 A8 (same as the 48-rompointer B3 A9, documented in > the 48FAQ p.110, by JKH) computes the number of days of a month only > from begin 1991 till end of 2090. I've been using the FAQ at http://www.engr.uvic.ca/~aschoorl/faq/, which doesn't have page numbers, and, as far as I've been able to determine, lacks this information. Could you give us a URL where the FAQ you're referring to can be found? -- ==== schrieb > Ç DUP 100 * FP 1E4 * SWAP > -42 FC? Ç IP È > Ç FP 100 * IP È IFTE > OVER 5791 - ABS 4208 <= > Ç DUP 1 - 7 MOD 2 MOD 31 - NEG > SWAP 2 - > Ç NIP È > Ç 2 - SWAP DUPDUP 4 MOD NOT ROT > 100 MOD ROT 400 MOD > NOT OR AND + È > IFTE > È > Ç DROP2 0 È IFTE > È > well... Looks good, really. 297 bytes and I think a lot of potential for a shorter program in sysRPL. Very impressiv are your constants 5791 and 4208. Would you please give a description of the algorithm. Best wishes Heiko ==== Wolfgang Rautenberg schrieb > The 49-rompointer B3 A8 (same as the 48-rompointer B3 A9, documented > in > the 48FAQ p.110, by JKH) computes the number of days of a month only I would like to know, where this rompointer is used on the mashines ? > CHALLENGE. ite a shortest (a) 49-UsrRPL program (b) 49-SysRPL > program, > DOM say, whose input is a DATE from the interval 1583...9999, and > whose > output is the correct number of days of the month to which the input > DATE belongs. No argument or input protection is demanded. IMPORANT: > Your program must be independent on the date flag -42 ! (b) 49-SysRPL programm ROM 1-19.6 Code-Size: 92.5 Crc: 782Ch Output: Number of Days of the month (real) Algorithm: It is calculated the first of the next month and going one day back. Valid for the years from 1583 to 3199 (inclusive). (Above 3199 there might be a change of the rule rqd) Warning: The validity of the month is not checked. With an non existing month the program puts a flashptr on the stack: FlashPtrBkp I have not found in dokus what that does if evaluated :-( It sounds like Backup . My HP49 survived an EVAL for the pointer :-) ..Heiko PS: Carstens prog. (second edition) is wonderfull and for sure shorter than mine in sysRPL. Cathedral ;-) aka DOM %%HP: T(3)A(R)F(.); << :: ' :: %FP %1+ %31_ DATE+DAYS %FP %1+ %-1 DATE+DAYS %IP ; %0 ZEROZEROTWO DOBIND BINT42 SysITE 2GETEVAL :: DUP %IP %100 %/ SWAPDUP %2 TRCXY %- %+ BINT42 SetSysFlag 2GETEVAL BINT42 ClrSysFlag ; ABND ; @ >> ==== > schrieb > > Ç DUP 100 * FP 1E4 * SWAP > -42 FC? Ç IP È > Ç FP 100 * IP È IFTE > OVER 5791 - ABS 4208 <= > Ç DUP 1 - 7 MOD 2 MOD 31 - NEG > SWAP 2 - > Ç NIP È > Ç 2 - SWAP DUPDUP 4 MOD NOT ROT > 100 MOD ROT 400 MOD > NOT OR AND + È > IFTE > È > Ç DROP2 0 È IFTE > È well... Looks good, really. > 297 bytes > and I think a lot of potential for a shorter program in sysRPL. Actually, when I put this program into my 49G, I got 269 bytes with a checksum of # 6AA1h. Perhaps the difference is that I put all of his numbers, except 1E4, in as zints instead of reals. > Very impressiv are your constants 5791 and 4208. > Would you please give a description of the algorithm. Yes, I'm a bit curious about that too. But as far as I've tested, it does work. It seems to me that as it doesn't use the calculator's knowledge of dates, it could be modified to take two numbers for the date (the year standing alone), and work for dates way beyond the year 9999. -- ==== -=[ Monday, 2003-05-19 8:42 PM +1200 (NZT) ]=- There is something very strange about the new 12c platinum. Resolving for i is, one could say, no longer possible. My one displays running for so long I have to press the ON key. I have tried many simple examples now. If indeed it does resolve, before running bateries flat, it must take a very long time. Oh, it did work n=10 i=10 pv=100 fv=0 solve for pmt=-162.745 resolve for i ... it seems to take 40 seconds. On the classic gold 12C it takes 4 seconds. This is, believe me, an example that is favourable to the 12c platinum. Unfortunately this does make it a returnable item, unless one really likes bugs like this. - Tony Reply-To: Veli-Pekka Nousiainen ==== Have you tried IRR with 20 cashflows then change the n down by one and test again, until n=1 > -=[ Monday, 2003-05-19 8:42 PM +1200 (NZT) ]=- There is something very strange about the new 12c platinum. > Resolving for i is, one could say, no longer possible. My one > displays running for so long I have to press the ON key. I > have tried many simple examples now. If indeed it does resolve, before running bateries flat, it > must take a very long time. Oh, it did work n=10 i=10 pv=100 fv=0 solve for pmt=-162.745 > resolve for i ... it seems to take 40 seconds. On the classic gold 12C it takes 4 seconds. This is, believe me, an example that is favourable to the 12c > platinum. Unfortunately this does make it a returnable item, unless one > really likes bugs like this. - Tony > ==== Hi Veli-Pekka Nousiainen ! then change the n down by one and test again, until n=1 I'll get back to you tomorrow on that ;-) No I hadn't tried it till you mentioned it. I can't imagine what has gone ong with interest rate solutions. Maybe they don't use the same initial guess strategy as on the 12C, or .. could be a wobbly exit criterion - that part does seem new. Eventually the answers produced seem OK. In fact more accurate than the 12C in some cases (not that the extra accuracy justifies the extra computation time - convergence is so rapid near an answer): n=E 8 i=1.234567891 E-6 PV=0 PMT=.01 Solve for FV=-1,973,883.396 Now resolve for i - I didn't measure how long it took -maybe 2.5 minutes, but the answer was 1.23456790 E-6. That is MUCH better than the 12C which gives 1.235555343 E-6, not that the 12C was known for great accuracy - and of course that didn't dampen it's popularity. The 12C does it in about 5 seconds as against something like 150 in the 12c-pt. Yup probably all it is is a new exit criterion - to fix it they just need to drop back to the old code, which probably took prcision to the maximum anyway, vis a vis internal precision. But, if this is the case I wonder why the iteration actually does actually stop, eventually. - Tony Reply-To: Veli-Pekka Nousiainen ==== X > In fact more accurate than the 12C in > some cases (not that the extra accuracy justifies the extra > computation time - convergence is so rapid near an answer): > n=E 8 i=1.234567891 E-6 PV=0 PMT=.01 > Solve for FV=-1,973,883.396 12CP: -1,973,883.143 12C: -1.973,883.296 19BII: -1,053,250.72188 49G: -1,053,250.72188 I think that you have already went past the accuracy limit here. Any ideas of an easier calculation? > Now resolve for i - I didn't measure how long it took -maybe > 2.5 minutes, but the answer was 1.23456790 E-6. That is > MUCH better than the 12C which gives 1.235555343 E-6, not that > the 12C was known for great accuracy - and of course that > didn't dampen it's popularity. The 12C does it in about 5 > seconds as against something like 150 in the 12c-pt. X ==== > X > In fact more accurate than the 12C in > some cases (not that the extra accuracy justifies the extra > computation time - convergence is so rapid near an answer): > n=E 8 i=1.234567891 E-6 PV=0 PMT=.01 > Solve for FV=-1,973,883.396 > 12CP: -1,973,883.143 > 12C: -1.973,883.296 > 19BII: -1,053,250.72188 > 49G: -1,053,250.72188 You just used 12 payments/yr on your hp49 instead of 1. I get -1,973,883.39591 on the 49 and 48 ... > I think that you have already went past the accuracy limit here. > Any ideas of an easier calculation? > Now resolve for i - I didn't measure how long it took -maybe > 2.5 minutes, but the answer was 1.23456790 E-6. That is > MUCH better than the 12C which gives 1.235555343 E-6, not that > the 12C was known for great accuracy - and of course that > didn't dampen it's popularity. The 12C does it in about 5 > seconds as against something like 150 in the 12c-pt. > X ==== Wired, on my 12c platinum, it takes less that 3 second to solve pmt for n=10, i=10, pv=100, fv=0 can you give me the serial number of your calculator? > -=[ Monday, 2003-05-19 8:42 PM +1200 (NZT) ]=- There is something very strange about the new 12c platinum. > Resolving for i is, one could say, no longer possible. My one > displays running for so long I have to press the ON key. I > have tried many simple examples now. If indeed it does resolve, before running bateries flat, it > must take a very long time. Oh, it did work n=10 i=10 pv=100 fv=0 solve for pmt=-162.745 > resolve for i ... it seems to take 40 seconds. On the classic gold 12C it takes 4 seconds. This is, believe me, an example that is favourable to the 12c > platinum. Unfortunately this does make it a returnable item, unless one > really likes bugs like this. - Tony > Reply-To: Veli-Pekka Nousiainen ==== > X > > In fact more accurate than the 12C in > > some cases (not that the extra accuracy justifies the extra > > computation time - convergence is so rapid near an answer): > > n=E 8 i=1.234567891 E-6 PV=0 PMT=.01 > > Solve for FV=-1,973,883.396 12CP: -1,973,883.143 12C: -1.973,883.296 19BII: -1,973,883.39591 49G: -1,973,883.39591 You just used 12 payments/yr on your hp49 instead of 1. > I get -1,973,883.39591 on the 49 and 48 ... Oops! That's how I *always* calculate my loans! Set to 1/yr => Both the 19BII & 49G is same as yours Question remains: does the problem already exceed a reasonable accuracy limit at this point? > I think that you have already went past the accuracy limit here. > Any ideas of an easier calculation? > > Now resolve for i - I didn't measure how long it took -maybe > > 2.5 minutes, but the answer was 1.23456790 E-6. That is > > MUCH better than the 12C which gives 1.235555343 E-6, not that > > the 12C was known for great accuracy - and of course that > > didn't dampen it's popularity. The 12C does it in about 5 > > seconds as against something like 150 in the 12c-pt. > X ==== +----------------------------------------------+ | Tue, 20.5.03 06:48 a.m. +1200 (NZT) | +----------------------------------------------+ Hi Cyrille de Bròbisson ! Wired, > > on my 12c platinum, it takes less that 3 second to solve pmt for n=10, i=10, > pv=100, fv=0 > > can you give me the serial number of your calculator? CN31301335, with the last digit stamped almost like an exponent. Solving for PMT is not the problem. *Resolving* for *i* is the problem. Example f CLEAR FIN n=10 pv=100 pmt=-20 solve for i. On the original 12C this takes 15 seconds. Solving for i again takes 4 seconds, as the original 12C uses the existing i as part of its initial guess. On the new 12c platinum it takes 35 seconds.Pressing i again takes another 35 seconds. This implies that the new platinum 12c does not re-use the existing i as part of the initial guess strategy. Tony > -=[ Monday, 2003-05-19 8:42 PM +1200 (NZT) ]=- > > There is something very strange about the new 12c platinum. > Resolving for i is, one could say, no longer possible. My one > displays running for so long I have to press the ON key. I > have tried many simple examples now. > > If indeed it does resolve, before running bateries flat, it > must take a very long time. > > Oh, it did work n=10 i=10 pv=100 fv=0 solve for pmt=-162.745 > resolve for i ... it seems to take 40 seconds. > > On the classic gold 12C it takes 4 seconds. > > This is, believe me, an example that is favourable to the 12c > platinum. > > Unfortunately this does make it a returnable item, unless one > really likes bugs like this. > > - Tony -- Tony Hutchins Wellington New Zealand #105 An ounce of discretion is worth a pound of learning. Proverb ==== +----------------------------------------------+ | Tue, 20.5.03 07:17 a.m. +1200 (NZT) | +----------------------------------------------+ Hi Veli-Pekka Nousiainen ! > > n=E 8 i=1.234567891 E-6 PV=0 PMT=.01 > > Solve for FV=-1,973,883.396 > 12CP: -1,973,883.143 > 12C: -1.973,883.296 No, both those have the correct fraction =.396 The HP92 and HP37E give the same. > 19BII: -1,973,883.39591 > 49G: -1,973,883.39591 > Question remains: does the problem already exceed a > reasonable accuracy limit at this point? Not at all. The problem is in resolving for i. Even on the HP41, a simple FIN program, using Newton Raphson for the i solution, quickly recovers the full 10 digits in i. On the HP41 the user has less precision available than the internal programs appear to use. The problem is with the slowness of the 12CP in resolving for i. I was really surprised to see this. Then I thought - Well, maybe there is a work-around - like on the 12C it uses the i already stored as an initial guess . But this seems to have been removed on the 12CP!!! I never thought anyone would tamper with the 12C TVM code for the i iterative solution, in the slightest. As it is though, it must use an method that converges very very slowly. I can't even imagine what that could be. It definitely needs the old 12C trick of taking the existing i as an initial guess. -- Tony Hutchins Wellington New Zealand #135 If you want to be happy, be. Leo Tolstoy ==== > Question remains: does the problem already exceed a reasonable > accuracy limit at this point? What is 'reasonable'? in this context? 10^8 years is hardly a 'reasonable' number, but you might plug that sort of number in as a lazy man's substitute for 'perpetual bond'. Or do you mean 'reasonable' as in as good as we can expect from the underlying floating point ? Or... calculate the NPV of the error and you'll be shocked! dd. ==== X > The problem is in resolving for i. X Preproduction units do have bugs... ==== >> X > > In fact more accurate than the 12C in > > some cases (not that the extra accuracy justifies the extra > > computation time - convergence is so rapid near an answer): > > n=E 8 i=1.234567891 E-6 PV=0 PMT=.01 > > Solve for FV=-1,973,883.396 >> 12CP: -1,973,883.143 >> 12C: -1.973,883.296 >> 19BII: -1,053,250.72188 >> 49G: -1,053,250.72188 You just used 12 payments/yr on your hp49 instead of 1. I get -1,973,883.39591 on the 49 and 48 ... > 200LX gives -1,973,883.395908565 -- Bruce Horrocks Hampshire England bh@granby.demon.co.uk ==== Sure, many of you have complaint about this wonderful device called hp49g, even me sometimes, but like lovely beings, it has its good things and bad things, but those things are incredible. Just this post to congratulate all the people that have worked on hp49g to make it real, thank you very much to all of you, wherever you are.. J.Manrique Users Club from Gijon #1077 HPCC Member ==== All I want to do is save certain results (store them) so I can recall them later. I have a very old hp41CV where it is quite simple. I get a result hit store assign a number from 01 to 99 and when I want that value again I simply hit recall and the number I assigned as the storage bin. On the hp49g I can't seem to do this. For example: I've entered 5/9 that I'd like to reuse to convert Fahrenheit to Celsius. When I hit STO on the 49g I get the following error. STO Error: Bad argument type. What's the deal. In the manual it says I can store values by assigning them as variables. For some reason my calculator won't allow me to do it. Any suggestions on how to perform this simple task? Tom ==== I don't have a 49g, but I assume it works as a 48 for that. When you want to store whatever is on the first level of the stack, just hit the ' key, and type the name you want to assign, e.g. 'result' then hit STO. To recall it, just type the name you assigned or press the key corresponding in the vars menu. Hope that helped. > All I want to do is save certain results (store them) so I can recall them > later. I have a very old hp41CV where it is quite simple. I get a result hit > store assign a number from 01 to 99 and when I want that value again I > simply hit recall and the number I assigned as the storage bin. On the > hp49g I can't seem to do this. For example: I've entered 5/9 that I'd like to reuse to convert Fahrenheit > to Celsius. When I hit STO on the 49g I get the following error. STO Error: > Bad argument type. What's the deal. In the manual it says I can store values > by assigning them as variables. For some reason my calculator won't allow me > to do it. Any suggestions on how to perform this simple task? Tom ==== > +----------------------------------------------+ > | Mon, 19.5.03 11:55 a.m. +1200 (NZT) | > +----------------------------------------------+ > > Hi David Davies ! > TH > Two things I was delighted about with the 12c: > TH > > TH > 1. My Black-Scholes ran in 8 seconds as against 12 > TH > seconds on the 12C gold. > > DD> You might want to try using a 48 or 49 for that. > > Personally I do use the 49 as well. In Solver I get an option > price in about 3 seconds. Re-solving for a volatility can take > up to 10-15 seconds. > > The 48 and 49 (and 28C) have a big advantage over the 12C with > their built-in UTPN functions. If they were functions - as in usable in the equation iter and quoted expressions - I'd be a happy man! > > DD> Plugging the Black-Scholes equation into a HP48gx will > DD> give you a price in 1/4 sec. and the built-in solver will > DD> spit out an implied volatility in 3/4 sec. Or are you > DD> doing something else? > > No, sounds very similar to what you describe. Except for the > time taken while re-solving for volatility - that can vary a > lot here. Mea Culpa. Should have tested it before I opened my mouth. It would seem that the 49 solver is really slowww... It looks like it is using the CAS expression evaluator and processing expressions rather than numbers at some points. On top of that the UTPN function is slower than on the 48 so even if you ite the whole formula in reverse polish you get solutions in about a couple of seconds or (for really bad guesses) maybe ten seconds. Does anyone with some 'insight' into the 49 ROM know more about this? dd. ==== Hi Tony, > 1. My Black-Scholes ran in 8 seconds as against 12 seconds on > the 12C gold. Hey! When will we see that program for the 12C? In Datafile maybe? Jordi Reply-To: Veli-Pekka Nousiainen ==== > The 48 and 49 (and 28C) have a big advantage over the 12C with > their built-in UTPN functions. If they were functions - as in usable in the equation iter and > quoted expressions - I'd be a happy man! You can make them functions... ==== > > The 48 and 49 (and 28C) have a big advantage over the 12C with > > their built-in UTPN functions. > > If they were functions - as in usable in the equation iter and > quoted expressions - I'd be a happy man! > You can make them functions... Of course, and the first thing I did after trying to use UTPN in the 'obvious' way was ap it in angled brackets. But why on earth was it defined this way? Is it a hangover from the hp28? And why was it left this way for so long (48g/1993, 49g/1999)? dd. ==== +----------------------------------------------+ | Tue, 20.5.03 07:11 a.m. +1200 (NZT) | +----------------------------------------------+ > 1. My Black-Scholes ran in 8 seconds as against 12 seconds on > the 12C gold. > > Hey! When will we see that program for the 12C? In Datafile maybe? Yes, quite likely! It seems worth publishing as although it is a 99 line program, it only uses 4 registers as temporary storage, in addition to having the inputs in the TVM registers. No special constants need to be stored. - Tony ==== > > > The 48 and 49 (and 28C) have a big advantage over the 12C with > > their built-in UTPN functions. > > > > If they were functions - as in usable in the equation iter and > > quoted expressions - I'd be a happy man! > You can make them functions... Of course, and the first thing I did after trying to use UTPN in the > 'obvious' way was ap it in angled brackets. But why on earth was it > defined this way? Is it a hangover from the hp28? And why was it > left this way for so long (48g/1993, 49g/1999)? dd. Life is short ,ROM is full. Originally there was no room to make all functions 'functions' Now I will give you a mission impossible: to a function. Like the UTPN and it's relatives, then HMS+, etc. ==== > Originally there was no room to make all functions 'functions' > Now I will give you a mission impossible: > to a function. Like the UTPN and it's relatives, then HMS+, etc. Well, the way you have to use UTPN at the moment means that the second thing (after the TVM solver) that any quantitatively switched-on MBA is going to try on the machine - i.e. putting the simplest option pricing formulae into the equation iter - will fail. Having to resort to a hack to use straightforward things like a normal probability function makes the firmware look really shabby. dd. ==== Time is critical in this exam. I do not like the format but do not know how to make it display better. I believe it could be make more efficient also. It would be great to see all the results display at the same time. Example input sigma x: -140 sigma y: 205 tau xy: 100 Results Principle stress Sigma pa: 231.89 Sigma pb: -166.89 Max shear stress: 199.39 Angle of orientation of principal stress with principal x-axis -15.05, 74.95 Angle of orientation of max shear plane 29.95, 119.95 program is listed below dashed line -------------------- pSTRES2 Ç ùx ùy úxy PROMPT Ù ùx ùy úxy Ç ùx ùy + 2 / ùx ùy - 2 / 2 ^ úxy 2 ^ + ° + 'ùpa' ÙTAG ùx ùy + 2 / ùx ùy - 2 / 2 ^ úxy 2 ^ + ° - 'ùpb' ÙTAG ùx ùy - 2 ^ úxy 2 * 2 ^ + .5 ^ 2 / 'úmax' ÙTAG úxy 2 * ùx ùy - / ATAN 2 / ' [Eth]px [CapitalEth]axis' ÙTAG DUP 90 + ùx ùy - úxy -2 * / ATAN 2 / ' [Eth]úù' ÙTAG DUP 90 + È È ==== >I am still in doubt which program to use for Laplace and inverse Laplace Why not get one of the Sparcom plug-in cards ??? EE Pro has Laplace and inverse Laplace ==== > Interesting thread again. > > A few points : > > - The 9600 and its very slightly improved version the 9650 are old > > news, I think more than 4 years old. > > You are right: as far as I know the newest Sharp calculator is > the EL-9900N (last quarter of 2002). > Moreover, the EL-9650's display is not much readable, as well as > the display of the EL-9400, and their contrast regulation is useless. > Luckily, the display of the EL-9900N is quite good, comparable with the > display of the HP48GX (my machine of choice). This is true. The screen on the 9600 suffers from a bad readability due to the extra layer of the touch screen. However, the 9400 has no touch screen, so it should have a much better contrast (I don't have this one). > > Last, I used the 9600 a while in the past, and if you never need to > > program anything this is a great machine. The pen interface is a very > > strong point on this product. > > > > Well, let's say it would have been a nice machine, with a decent display... > I used it mainly for its stat functions, but the poor display is a bit > tedious, and IMHO the pen-touch screen is just a useless gadget: > I have been pleased by the fact that the EL-9900N has traded the pen-touch > screen for a good display, yet it features an other useless gadget, > a detachable reversible keyboard. Granted, while I maintain that the touch screen has some value... YMMV ! ==== I am a die-hard HP49 user. My wife is a psychology student and is currently looking for a calculator that will do statistics. I want her to buy an HP, but I'm having a hard time convincing her to get one. I went to their website, and couldn't find anything there that was better than TI's equivalant. What is HP's good statistical calculator? Where can I find it? ==== Who made that one ........the ACO ???? ;-) >... has been discontinued due to defective keyboard ,screen, memory, software and gold paint job! Toby Reply-To: Veli-Pekka Nousiainen ==== X > Sometimes defining a bug is a bit subjective; do we define it as > doesn't work as was asked for , doesn't work as the developer > intended , doesn't work as documented , or doesn't work as the user > wants it to ? Or perhaps some other criterion? If it crashes, hangs-up, runs forever => it's a serious bug! The same with ong answers (not counting the trailing last digits) Leaving carbage on the Stack. Not functioning at all....??? X > Including the Pacific would make California and Italy so small that one > couldn't recognize them as peninsulas on the 48/49 screen :-) A scollable map? X > A control alarm does close the editor, but, as long as last command line > saving is enabled, whatever you were editing is saved and can be > retrieved with LeftShift CMD. If you were editing something that was on > the stack, then the original is still on the stack above anything that > the control alarm put there, as well as the partially edited last > command line version. This behaviour goes all the way back to the 48SX. > Perhaps not the perfect way to handle the situation, but I think that > it's well thought out and a reasonably good way of handling it. As for the requirement that the command lines saves be enabled for this > to work nicely, why would anyone disable them other than temporarily to > get out of a low memory situation? For the equation iter, on the 49G the control alarms run without > closing the equation iter environment, although I expect that any key > presses will just accumulate in the buffer until the alarm is finished. X > But on the 49G, the alarm index and any immediately following valid > objects from the control alarm are entered into the matrix, and any > invalid object and anything following it is left on the matrix iter's edit line. X > Dare I call it buggy? This could stand some improvement; even just going > back to the way the 48 series handled it would be much better. BUG it is, ACO has failed... I wonder if it is possible to start yet another Environment eg. use Virtual Stack for control alarms in the 49G alarm system... ==== shucks.... I guess I made a boo boo ==== EPSON David Banker schrieb im Newsbeitrag > Who has blank ROM cards to be able to put programs on > for the 48GX. ==== Please view the GIF image of maths expression at: http://smendes.com/mathex.gif The guys on the TI discussion board say the result returned by 49G (in Complex > mode) is INCORRECT. I cannot say one way or the other, because I am not that good a mathmatician What do you all have to say ?? Greetings, It is true that the result (#2) from the 49G posted on that site is incorrect. Instead of '-((X-i)/2*EXP(-(i*X)))-(X-i)/2*EXP(i*X)' It should be '-((X-i)/2*EXP(-(i*X)))-(X+i)/2*EXP(i*X)' Note the X+i instead of X-i. HOWEVER, when I ran this on my calculator, it returned the CORRECT answer. So, there must have been a mistake made by the author of that page. Especially since #3 is correct. It is very unlikely for the calculator to result 'incorrect' answers. Perhaps they may not be in the form we're looking for, but the algorithms shouldn't fail for Euler identities. Albert ==== Your result in #2 is INCORRECT in complex mode. The HP49G produces an (X+i)/2 for the second exponential is correct if you do the complex math using Euler's formula. I leave it to someone else to do justice for the third case... !Demeter! ==== So what is the difference between the Custom menu and a director with my custom commands in it? > Don't worry, be happy - it's fully automatic! > Even if you remove them, they appear back! { DATA CALC REPORT } MENU > will show a CUSTOM nebu when you press that key. > The list above is stored in CST variable in your directory. > VARS gives all the variables in the current directory. The huge PDF file > http://www.hpcalc.org/details.php?id=3937 > will give you an advanced example on using your own custom menus. > > Help > > Can someone please give me an explaination of the 49G's memory - namely > the > difference between directories, the VARS key and the CUSTOM key. VARS. > But there are variables that have appeared that I did not directly create > like CST, IOPAR, MODUL and VX. > > I take it that these are necessary and are created when you use or program > the calculator (although I'm not sure how or when). > > I would like to put my custom programs in their own directory to stay > organized. Do all those extra variables have to be in the directory that > my > programs are in? > > The manual is very limited in explaining what is going on. > ==== What is the difference between my custom menu and a directory with my programs in it? > Don't worry, be happy - it's fully automatic! > Even if you remove them, they appear back! { DATA CALC REPORT } MENU > will show a CUSTOM nebu when you press that key. > The list above is stored in CST variable in your directory. > VARS gives all the variables in the current directory. The huge PDF file > http://www.hpcalc.org/details.php?id=3937 > will give you an advanced example on using your own custom menus. > > Help > > Can someone please give me an explaination of the 49G's memory - namely > the > difference between directories, the VARS key and the CUSTOM key. VARS. > But there are variables that have appeared that I did not directly create > like CST, IOPAR, MODUL and VX. > > I take it that these are necessary and are created when you use or program > the calculator (although I'm not sure how or when). > > I would like to put my custom programs in their own directory to stay > organized. Do all those extra variables have to be in the directory that > my > programs are in? > > The manual is very limited in explaining what is going on. > ==== So what is the difference between the Custom menu and a directory I make to store my programs? > Don't worry, be happy - it's fully automatic! > Even if you remove them, they appear back! { DATA CALC REPORT } MENU > will show a CUSTOM nebu when you press that key. > The list above is stored in CST variable in your directory. > VARS gives all the variables in the current directory. The huge PDF file > http://www.hpcalc.org/details.php?id=3937 > will give you an advanced example on using your own custom menus. > > Help > > Can someone please give me an explaination of the 49G's memory - namely > the > difference between directories, the VARS key and the CUSTOM key. VARS. > But there are variables that have appeared that I did not directly create > like CST, IOPAR, MODUL and VX. > > I take it that these are necessary and are created when you use or program > the calculator (although I'm not sure how or when). > > I would like to put my custom programs in their own directory to stay > organized. Do all those extra variables have to be in the directory that > my > programs are in? > > The manual is very limited in explaining what is going on. > ==== Then what is the difference between the Custom menu and a directory I create and store my programs in? > Don't worry, be happy - it's fully automatic! > Even if you remove them, they appear back! { DATA CALC REPORT } MENU > will show a CUSTOM nebu when you press that key. > The list above is stored in CST variable in your directory. > VARS gives all the variables in the current directory. The huge PDF file > http://www.hpcalc.org/details.php?id=3937 > will give you an advanced example on using your own custom menus. > > Help > > Can someone please give me an explaination of the 49G's memory - namely > the > difference between directories, the VARS key and the CUSTOM key. VARS. > But there are variables that have appeared that I did not directly create > like CST, IOPAR, MODUL and VX. > > I take it that these are necessary and are created when you use or program > the calculator (although I'm not sure how or when). > > I would like to put my custom programs in their own directory to stay > organized. Do all those extra variables have to be in the directory that > my > programs are in? > > The manual is very limited in explaining what is going on. > ==== Sorry about that. I had trouble with my newsgroup set up. It's fixed now. > Don't worry, be happy - it's fully automatic! > Even if you remove them, they appear back! { DATA CALC REPORT } MENU > will show a CUSTOM nebu when you press that key. > The list above is stored in CST variable in your directory. > VARS gives all the variables in the current directory. The huge PDF file > http://www.hpcalc.org/details.php?id=3937 > will give you an advanced example on using your own custom menus. > > Help > > Can someone please give me an explaination of the 49G's memory - namely > the > difference between directories, the VARS key and the CUSTOM key. VARS. > But there are variables that have appeared that I did not directly create > like CST, IOPAR, MODUL and VX. > > I take it that these are necessary and are created when you use or program > the calculator (although I'm not sure how or when). > > I would like to put my custom programs in their own directory to stay > organized. Do all those extra variables have to be in the directory that > my > programs are in? > > The manual is very limited in explaining what is going on. > ==== > difference between directories, the VARS key and the CUSTOM key. The calc is always displaying some *menu* at the bottom. The VARS key displays the built-in menu showing variables, which are objects in the current directory; in this menu, unshifted menu keys evaluate variables, left-shifted keys store into variables, and right-shifted keys recall variables (other menus can define plain and shifted operations any way they like, because menus are fully programmable objects). The CUSTOM (formerly CST in HP48) menu displays any menu stored in a variable named CST (or you can store in CST the name of *another* variable which actually contains the menu). Advanced User's Reference manuals show how you can create menus of your own; manuals (or books) itten for the HP48 are often much more detailed about this than HP49G manuals. > CST, IOPAR, MODUL and VX. A command named MENU stores its argument directly into a variable named CST (creating it, if necessary), which in turn, as you read above, is what the CUSTOM [or CST] keyboard function displays when you subsequently re-invoke it. Since said MENU command always stores into the same CST variable, wiping out any previously saved menu, I prefer to store my laboriously developed menus into variables named something else, and to store only the *name* of one such other variable at a time into CST, via commands such as 'MYMENU' MENU IOPAR is automatically created when you either print or transfer data externally; it remembers your serial port settings (baud rate, etc.) PRTPAR remembers printing parameters (line length, etc.) VX, MODULO, PERIOD, REALASSUME, EPS, etc. are created and used by the 49G algebra system (CAS); in old ROM versions, these were created by default in the HOME directory, and once created they will stay there until manually moved, but in later ROM versions, the default directory CASDIR is first created, and then the automatic CAS variables will hide in CASDIR, unless they already exist in HOME. > I take it that these are necessary and are created > when you use or program the calculator > (although I'm not sure how or when). It's all explained in the book of Genesis :) If you don't have that book, there are some manuals from HP which repeat the story :) > I would like to put my custom programs in their own directory > to stay organized. Do all those extra variables have to be > in the directory that my programs are in? If you have (or upgrade to) a ROM version which creates a CASDIR directory, you may then use the Filer to move the CAS variables into CASDIR; once thus moved, they will then stay put forever afterwards. IOPAR and PRTPAR have to reside in the HOME directory, but you may move them to the end of HOME by creating them once (a PR1 command will do that) and then moving all other variables to precede those, via an ORDER command: { X Y Z } ORDER moves the listed variables to the front of the current directory. STO commands store into the current directory only. RCL commands search upwards to higher directories, if necessary, until any variable of the specified name is found. You can also, as in a computer, specify a full path name for RCL (but not for STO), e.g. { HOME MATH X } RCL Note that custom programs have no inherent connection to the CUSTOM menu, unless you design and create your own custom menu(s) to relate to your own programs. My original 48G custom [CST] menu was blank for many months, and I originally employed the seemingly useless CST key only as a means to hide my VARS menu :) [perhaps that's why CUSTOM was demoted to a shifted key in the 49G :] > The manual is very limited in explaining what is going on. Everyone always noticed that -- except the management at HP. Hasn't the day yet arrived when every geek gadget is so intuitive (or packed with built-in help) that no manual is ever necessary? After all, God provided no manual for operating the Earth, Everyone on board our Spaceship Earth can live abundantly and successfully... Humanity has the option to make it. We must choose it before it expires... Whether it is to be Utopia or Oblivion will be a touch-and-go relay race right up to the final moment. When I was born, humanity was 95 per cent illiterate. Since I've been born, the population has doubled and that total population is now 65 per cent literate. That's a gain of 130-fold of the literacy. When humanity is primarily illiterate, it needs leaders to understand and get the information and deal with it. When we are at the point where the majority of humans themselves are literate, able to get the information, we're in an entirely new relationship to the Universe. We are at the point where the integrity of the individual counts, and not what the political leadership or the religious leadership says to do. http://www.bfi.org/operating_manual.htm http://www.bfi.org http://www.mum.edu . ==== Oh, it seems I forgot to mention that you can create all the subdirectories that you want, via 'MYDIR' CRDIR, so you can certainly organize your personal programs, notes, databases, permanent stuff, temporary stuff, etc. exactly as you see fit. Programs that are useful all the time might be stored in HOME, where they can always be invoked by name (names, in turn, can be assigned to places on the keyboard), while specialized programs might be filed in their own separate niches, without cluttering any one directory with all the names at once (I have more than 250 programs in HOME, but very few variables in any other directory). Many common downloadable collections of program functions are structured as non-editable libraries, which are stored completely apart from the directory/variable scheme, and don't appear in any directory menu at all. When VAR menus nonetheless get too cluttered, there exist variable hiding programs which can make selected variables disappear from the visible VAR menu, even though they still exist in the given directory (this also protects them against inadvertent over- iting via accidental manual store operations from the keyboard). Download a good HP48G manual (or AUR ) and read all about it :) [r->] [OFF] . ==== X > Programs that are useful all the time might be stored > in HOME, where they can always be invoked by name > (names, in turn, can be assigned to places on the keyboard), > while specialized programs might be filed in their own > separate niches, without cluttering any one directory > with all the names at once (I have more than 250 programs > in HOME, but very few variables in any other directory). X juuri t .8am .8an takia minulla on kaksi Home-tasoa: HOME ja oma HOME. joilla saan edes hiukan j .8arke .8a tietojen organisointiin. ==== > ... Range (yyyy-mm-dd) from > 1582-11-01 to 9999-11-31 > Size: 58 bytes Crc: B20h > :: > % 1.012 @ date which is independent from flag -42 > DUPROT > ... > > very good program for a beginner in SysRPL (my school :-). It will be > hard even for WH to beat you ... > but not impossible.. take %80 %1/ %1+ (1.012500) instead of % 1.012. Saves three bytes. Werner Reply-To: dominik@_DROP_THIS_science.uva.nl ==== Werners modification of Heiko's SysRPL program brings that one down to 55 bytes. I cannot beat this, but just for fun, here is another 55 Byte SysRPL program. It uses one single unsupported but stable pointer, so it does not conform to the rules. But it uses a different algorithm - so maybe it is fun to see. :: SaveSysFlags 42 SetSysFlag DUP %FP SWAP BEGIN %1 DATE+DAYS 2DUP %FP %<> UNTIL SWAPDROP %-1 DATE+DAYS %IP RestoreSysFlags ; - Carsten ==== > Your UsrRPL-program contains the leap year rules. You will be > surprised by solutions which don't explicitly refer to these rules. The number of days in this month is the difference (DDAYS) between the first date in this month and the first date in the next month. To get the first date in this month, overlay 01 onto the day number part of today's date. To get the first date in the next month, add at least 31 days (DATE+) to the already-obtained first date in this month, then overlay 01 again onto that date. This is done somewhere in this ancient calendar (text mode) program which I dug out from my 48G's memory; given any date, it displays the same calendar as does the Unix cal program, for the month in which that date falls, e.g. for 1.012000 Jan 2000 Su Mo Tu We Th Fr Sa 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 For the sake of the HP49G, three R ->I commands have been inserted, which you must omit for HP48: << STD 0. DATE+ JanFebMarAprMayJunJulAugSepOctNovDec 3. PICK IP 3. * DUP 2. - SWAP SUB + + 1. DUP DUP 5. PICK %T FP %T %T R ->I + Su Mo Tu We Th Fr Sa + { DUP IP 1. ROT %T FP 1. + 1. % + } 3. PICK OVER EVAL DUP DUP 31. DATE+ 4. ROLL EVAL DDAYS SWAP 10.10201 DDAYS -7. MOD 1. + DUP 4. ROLLD R ->I SWAP R ->I FOR d OVER d - 7. MOD NOT { + } IFT CASE d 9. > THEN d + END d 0. > THEN d + END END + NEXT ROT ROT DROP2 >> This 67.5 byte program gives the number of days in any date's month, in any HP48/49 (provided that flag -42 is clear :) << { DUP IP 1. ROT %T FP 1. + 1. % + } SWAP OVER EVAL DUP 6. SQ DATE+ ROT EVAL DDAYS >> [r->] [OFF] . ==== If flag -42 is set, then: << { FP 1. + } SWAP OVER EVAL DUP 6. SQ DATE+ ROT EVAL DDAYS >> @ 47.5 bytes Flag-independent: << -42. FS? { { FP 1. + } } { { DUP IP 1. ROT %T FP 1. + 1. % + } } IFTE SWAP OVER EVAL DUP 6. SQ DATE+ ROT EVAL DDAYS >> @ 105.5 bytes, but no wasted time << RCLF SWAP DATE SWAP DDAYS -42. SF DATE SWAP DATE+ { FP 1. + } SWAP OVER EVAL DUP 6. SQ DATE+ ROT EVAL DDAYS SWAP STOF >> @ 85.5 bytes, but uses DATE twice (can fail at midnite :) 6. 7. * NEG in place of -42. saves half a byte. -42 (integer) in place of -42. (real) saves 4 bytes. A stitch in time saves nine. . Reply-To: dominik@_DROP_THIS_science.uva.nl ==== HA> Heiko Arnemann schrieb HA> I am so sorry, HA> I made a fault, HA> see my correction below. With one unsupported pointer, this goes down to 45 Bytes! Wow. :: SaveSysFlags %FP %5 %+ %27 DATE+DAYS %FP %1+ %-1 DATE+DAYS %IP RestoreSysFlags ; reverse operation is not.... - Carsten ==== Wolfgang Rautenberg schrieb > CHALLENGE. ite a shortest (a) 49-UsrRPL program Hi Wolfgang, Size: 73 Byte Crc: # ACD6h %%HP: T(3)A(R)F(.); << RCLF DATE ROT DDAYS -42 SF DATE SWAP DATE+ FP 1 + 31 DATE+ FP 1 + -1 DATE+ IP SWAP STOF >> I do not like it, because of the seldom case were the programm is exectuted at midnight *and when the month is changing. An expensive challenge to avoid the error:-) Without using DATE twice, I come to 81 Bytes. Validity between. November 1582 and November 9999 (included). I hope, I haven't used unsupported PTR's ;-) ..Heiko Up to now I thougt %1/ is equivalent to << 1 / >> mmpf :-) ==== PPPS: As I read the post by John H. Meyers, I realized that these programs fail if the date is the 31st of any month except December, or the 30th and, for about 3 out of 4 cases, the 29th of January. Oh well. No doubt the easiest solution will be to add code that sets the day of the month -- Reply-To: raut@math.ful-berlin.de ==== It will probably be hard to offer a smaller program in SysRPL as the one by Heiko and Werner. But in UsrRPL, the programs proposed by JMP and JHM are still far from being optimal. You've still another chance. I'll announce the winners only on Friday evening. PS. Modification of a one of JHM's wise aphorisms: After all, God provided no manual for operating the Earth. Hence, why should this have been considered by the creators of the HP49? ==== It will probably be hard to offer a smaller program in SysRPL as the one > by Heiko and Werner. But in UsrRPL, the programs proposed by JMP and JHM > are still far from being optimal. You've still another chance. I'll > announce the winners only on Friday evening. > PS. Modification of a one of JHM's wise aphorisms: > After all, God provided no manual for operating the Earth. Hence, H .9apsist .8a! Onhan meill .8a Raamattu! > why should this have been considered by the creators of the HP49? Siis vitsi! hah? ==== > Pre-production units do have bugs... Is it at all the same processor as the original? (note that new design uses quite a different battery voltage) If by any chance it's a different processor, was the original coding scrapped and re itten, and if so, could even the basic math operations have any detectable differences in results? Going from 99 program lines to 300 or so is certainly different, as is the addition of dual RPN/ALG modes and extra programmable instructions for same, which also must require, internally, either more than 1 byte per program line or else some sort of variable-length internal program encoding (the original 12C had already used up either 255 or all 256 possible merged 1-byte program codes, which was the original reason that storage register arithmetic had to be limited to the first few registers). Could it possibly really be a sheep in wolf's clothing? [r->] [OFF] . ==== > > Pre-production units do have bugs... Is it at all the same processor as the original? > (note that new design uses quite a different battery voltage) No, even the current 12C (Gold) has a new CPU and a new battery system. 3-batt versions are collectables. Now there are first camples with serial numbers and all the bugs that I know of are gone. Gotta find some new (maybe next week)... > If by any chance it's a different processor, > was the original coding scrapped and re itten, > and if so, could even the basic math operations > have any detectable differences in results? Only i and IRR are different > Going from 99 program lines to 300 or so is certainly different, max 399 lines. Note: only 20 registers are used in betas. R0-R .9 > as is the addition of dual RPN/ALG modes and extra programmable > instructions for same, which also must require, internally, > either more than 1 byte per program line or else > some sort of variable-length internal program encoding > (the original 12C had already used up either 255 > or all 256 possible merged 1-byte program codes, > which was the original reason that storage register arithmetic > had to be limited to the first few registers). I don't know - I doubt that. Aren't the GTO's double bytes anyhow? > Could it possibly really be a sheep in wolf's clothing? I think it's wolf/wolf and will eat the cheap sheeps away. ==== > Note: only 20 registers are used in betas. R0-R.9 I didn't spot mention of any more regs in the current downloadable manual; only that R.9 disappears after program line 310 appears, etc. So what's in the actual product, you guys that have 'em ? > Aren't the GTO's double bytes anyhow? Not in the original 12C; as a boring exercise, try counting how many programmable operations exist in that original specification -- don't forget RCL g 12x and RCL g 12/ (which weren't in the original manual), RCL g R/S (weird sequence for IRR with a guess), f 0 thru f 9, GTO 00 thru GTO 99, and all the storage register arithmetic. You can not, of course, get GTO 000 thru GTO 400 into one-byte codes -- the manual keeps referring to 7 lines program memory per sacrificed register, but that leaves something unanswered. In other HP models, the 15C had some 1-byte codes (up to 240) and all the rest were 2-byte codes, although they displayed as one line number, while the HP32Sii has uniform 12-bit (1.5-byte) codes. > I think it's wolf/wolf and will eat the cheap sheeps away. I wouldn't trade one HP17Bii for two platinum 12Cs :) The 17B/19B menus, many more built-in functions, huge memory, and smarter equation solver than 48/49 are worth so much more to me than adding more program lines to the primitive 12C programming system which was so restrictive (no insertion, no labels, no subroutines, only 2 tests, etc., and gee -- no printing, either!) So when will the platinum 48G come out? (how about with a free built-in 256K + 4_MB flash memory bank? :) . ==== > the original 12C had already used up either 255 or all 256 possible > merged 1-byte program codes http://library.hp41.org/Files/00014/qajsiaaa.gif 253 + ALG + RPN + x^2 = 256 ... but what about GTO 100 through GTO 399? New CPU with new bugs, I'm afraid. Read more in the next Datafile. :-) Jordi ==== > >> Pre-production units do have bugs... > > Is it at all the same processor as the original? > (note that new design uses quite a different battery voltage) > > If by any chance it's a different processor, > was the original coding scrapped and re itten, > and if so, could even the basic math operations > have any detectable differences in results? It's an entirely different CPU with a totally new code. Just made it look identical to the old HP12C. I guess differences in some calculations will show more this fact. > Could it possibly really be a sheep in wolf's clothing? What about the reverse point of view ? :) ==== > I guess differences in some calculations will show more this fact. Also difference in execution times and bugs. AFAIK, there have never been found any bugs in the 12C. Jordi -- ==== > AFAIK, there have never been found any bugs in the 12C. I found one, in the original production era; it has to do with trying to enter non-integers as cash flow repetition factors, some values like 1.000000001, which *should* be detected as invalid, but sometimes it is accepted, and then when recalled, the value actually stored (or recalled) may be found to have been interpreted as zero; this is a very obscure bug, which has probably been observed only by idle people who had no real money, and consequently no real cash flows to enter :) But in other areas, the 12C avoided common minor bugs that were present in other models of the same series, such as 15C and 16C. HP must have put their A-team on the original 12C :) [r->] [OFF] . ==== > I found one, in the original production era Confirmed in my two 12C's (1982 US & 1998 Malaysia) Fixed in the Platinum, though (noninteger values for Nj generates an Error 6) But I meant that there have never been any bug reports ... ;-) Jordi -- ==== > HP must have put their A-team on the original 12C :) they more specially put over 5 year of R&D in it! current products (like a 49G) are developped on 1 year time frame max. This makes a word of a difference! ==== > >> HP must have put their A-team on the original 12C :) > > they more specially put over 5 year of R&D in it! current products (like > a 49G) are developped on 1 year time frame max. This makes a word of a > difference! Yes, and past products were made to last, nowadays, most users buy a calc for just their studies and perhaps some field calculus on work. You could spent 5 years on R&D to show a new product, but times have changed. -- J.Manrique L .97pez de la Fuente Users Club from Gij .97n 1077 HPCC Member ==== > > HP must have put their A-team on the original 12C :) they more specially put over 5 year of R&D in it! > current products (like a 49G) are developped on 1 year time frame max. This > makes a word of a difference! Ja kuinka iso on viel .8a ROM koodin kokoero? Taitaa olla l .8ahes satakertainen aika/koodi tehoero? ==== > Note: only 20 registers are used in betas. R0-R.9 > I didn't spot mention of any more regs > in the current downloadable manual; > only that R.9 disappears after program line 310 appears, etc. > So what's in the actual product, you guys that have 'em ? > > Aren't the GTO's double bytes anyhow? > Not in the original 12C; as a boring exercise, > try counting how many programmable operations exist > in that original specification -- don't forget > RCL g 12x and RCL g 12/ (which weren't in the original manual), > RCL g R/S (weird sequence for IRR with a guess), Hei! Tuota en huomannutkaan! > f 0 thru f 9, GTO 00 thru GTO 99, > and all the storage register arithmetic. > You can not, of course, get GTO 000 thru GTO 400 Juuri n .8ain! Siis kaksi tavua hypyille? > into one-byte codes -- the manual keeps referring > to 7 lines program memory per sacrificed register, > but that leaves something unanswered. In other HP models, the 15C had some 1-byte codes > (up to 240) and all the rest were 2-byte codes, > although they displayed as one line number, > while the HP32Sii has uniform 12-bit (1.5-byte) codes. Oho! Tuota en _todellakaan_ tiennyt! 4096 komentoa!! > I think it's wolf/wolf and will eat the cheap sheeps away. > I wouldn't trade one HP17Bii for two platinum 12Cs :) Odotellaan sitten Platinum -malleja noista laskimista... > The 17B/19B menus, many more built-in functions, > huge memory, and smarter equation solver than 48/49 > are worth so much more to me than adding more program lines > to the primitive 12C programming system which was so restrictive > (no insertion, no labels, no subroutines, only 2 tests, etc., > and gee -- no printing, either!) So when will the platinum 48G come out? Kenp .8a tiet .8ais sen, ken kertoisi huomisen? T .8an .8a vuonna?! > (how about with a free built-in 256K + 4_MB flash memory bank? :) Kuulostaa hyv .8alt .8a! Lupaatko varmasti? ==== +----------------------------------------------+ | Thu, 22.5.03 11:00 a.m. +1200 (NZT) | +----------------------------------------------+ > It's an entirely different CPU with a totally new code. Wow, I did not guess that! That is a lot of work. I am impressed that HP have done this. Must have required a lot of testing as well. > Just made it look identical to the old HP12C. Ahha, that's why it seems the same ;-) > I guess differences in some calculations will show more > this fact. Here I only notice that solutions for i in the TVM part seem slower. Possibly the iterations go until greater accuracy is achieved. Do you know if the iteration process uses the method of bisection ? This would be the shortest to code, but relatively slow to converge. If there is limited room for code then I guess this is all that can be done. The great thing is : it *does* converge!! At first I honestly thought it was never going to - oh me of little faith > Could it possibly really be a sheep in wolf's clothing? > What about the reverse point of view ? :) LOL! Actually, what is the grey metal of this wolf in sheep's clothing. I assume it is Al rather than Pt . If there is some Pt on the case I hope it is extractable -- Tony Hutchins Wellington New Zealand #64 Air Force Bingo: B-52...F-111...M-16...F-18...F-117... ==== --------------------------------------------------------------------- ... I'm panning to get the required hardware to run it, along with the best libraries I could fit. But, I find my plain-jane, trusty 48 a little slow. (please, say 'mucho faster') Toby ==== > ... > > I'm panning to get the required hardware to run it, along with the best > libraries I could fit. But, I find my plain-jane, trusty 48 a little > slow. > > (please, say 'mucho faster') > > Toby > > -- Thinking about installing MK? Don't doubt!! It's mucho mucho faster and you'll get many many new features. A little old but read this: http://www.hpcalc.org/hp48/apps/mk/mk.php ==== Seen: > Naturally your most important programs are in FlashRAM. [ROM?] I can backup my whole user memory (as a single object) to flash, or I can copy newly edited individual programs to flash as backups, but can I also edit and run my individual user programs directly from flash? You know, I've been seeing some Sharp PDA organizers which say that they need no backup battery, because their data memory is *all* flash (similar to phone answering machines); I wonder when large-scale programmable calculators, if any new models will ever exist, are going to be like that? Of course, I also read about cases where, when the Sharp PDA batteries get drained, they can't ite(erase) flash any longer, and serious data corruption or loss is claimed, just like those nearly impossible cases nonetheless occasionally reported in connection with well-known people's 49Gs. [r->] [OFF] . ==== > Seen: > > Naturally your most important programs are in FlashRAM. [ROM?] I can backup my whole user memory (as a single object) to flash, > or I can copy newly edited individual programs to flash as backups, > but can I also edit and run my individual user programs > directly from flash? HP49G Flash: Editing cannot be stored directly to flash :2: My.Prog or :F: My.prog will run from Flash I think that they are copied (invisibly to the user) to RAM prior to running. > You know, I've been seeing some Sharp PDA organizers > which say that they need no backup battery, because their > data memory is *all* flash (similar to phone answering machines); > I wonder when large-scale programmable calculators, > if any new models will ever exist, are going to be like that? One can ite to a new location of flash, updating a deleted bit but Flash has to be cleared by erasing a bank (64KB in the Intel Flash insied 49G; OS bank is still 128KB) The PDA should have some working memory anyway. > Of course, I also read about cases where, when the Sharp PDA > batteries get drained, they can't ite(erase) flash any longer, > and serious data corruption or loss is claimed, just like those > nearly impossible cases nonetheless occasionally reported > in connection with well-known people's 49Gs. Wolfgang's 49G ?? My last failure was with the 1.16 ROM. Reply-To: Veli-Pekka Nousiainen ==== > Seen: > > Naturally your most important programs are in FlashRAM. [ROM?] I can backup my whole user memory (as a single object) to flash, > or I can copy newly edited individual programs to flash as backups, > but can I also edit and run my individual user programs > directly from flash? HP49G Flash: Muokattua tietoa ei voi tallentaa suoraan p .8a .8alle :2: Ohjelma tai :F: Ohjelma voidaan ajaa suoraa Flashilt .8a, mutta n ekyll .8a ensin kopioidaan (huomaamattomasti) RAMmiin enne ajoa. > You know, I've been seeing some Sharp PDA organizers > which say that they need no backup battery, because their > data memory is *all* flash (similar to phone answering machines); > I wonder when large-scale programmable calculators, > if any new models will ever exist, are going to be like that? Aina voi kirjoittaa vapaaseen kohtaa flash-muistia ja merkit .8a poistetuksi yht .8a bitti .8a muuttamalla, mutta uudelleenkirjoitus voi tapahtua vasta kun koko lohko on kirjoitettu uudestaan. (64KB 49G:n sis .8ainen Intel Flash; OS kaytt .8a .8a edelleen 128KB) PDA:ssa on toki oltava edes jonkinlainen ty .9amuisti. > Of course, I also read about cases where, when the Sharp PDA > batteries get drained, they can't ite(erase) flash any longer, > and serious data corruption or loss is claimed, just like those > nearly impossible cases nonetheless occasionally reported > in connection with well-known people's 49Gs. Wolfgangin 49G ?? Viimeine flashin pett .8aminen minun koneessani tapahtui 1.16 ROMmilla. ==== > You know, I've been seeing some Sharp PDA organizers > which say that they need no backup battery, because their > data memory is *all* flash (similar to phone answering machines); > I wonder when large-scale programmable calculators, > if any new models will ever exist, are going to be like that? > > Of course, I also read about cases where, when the Sharp PDA > batteries get drained, they can't ite(erase) flash any longer, > and serious data corruption or loss is claimed, just like those > nearly impossible cases nonetheless occasionally reported > in connection with well-known people's 49Gs. The old Newton MessagePad (Apple's revolutionary PDA) uses Flash as main storage. There are many stories of people finding their old Newt in a box forgotten for years, they put new batteries and pop ! Not a byte lost. The Newton was discontinued more than 5 years ago, and it's still going real strong. So much for Flash re ite cycles. The Newt is still the best PDA as of today IMHO... well enough of this off-topic post. ==== I have an hp48 (manufacture in '97) that I bought new and has been giving me problems since day one. It sporadically tells me that my memory is full and I need to purge it, usually in the middle of doing something important, ie. and exam. I have reset, that doesn't work. I have NEVER programmed with the calculator or downloaded anything to it. I know, I know, I should have dealt with it 5 years ago when I bought it, but what can I say. This is very frustrating, to say the least. I love RPN, like the keypad of the 48, and don't think I would need the bigger memory of the gx. I am returning to school at the age of 34 to finish my EE degree, and would like to have a reliable RPN calc. I have a ti-89 that I bought after seeing all the cool things it can do, but I just can't get used to the interface. Funny how the only way I can stand it is to use it in a half assed RPN way by pasting the answers instead of using parenthases. 5 keystrokes to find a square root... well, I'm preaching to the choir here I suppose. I am a newbie to a forum like this, scoured previous posts and didn't see this problem (not to say it hasn't been answered somewhere). Is my memory chip fried? do I need to suck it up and find an old one on e-bay? or should I just learn to use the ti-89 and deal with it? Corey Christianson Pullman, WA ==== Dear Corey: Your HP 48 *screams* it's defective. That's what warranties are for! I know nothing about internals/repairs but I would have certainly tried getting a replacement 5 yrs ago. Maybe some forum people have seen this problem before and can suggest something to do to fix/alleviate it. However, if you want the enjoyment of a fully working HP calculator, your best bet at this point may just be to buy another HP48 or a 49G while they are still available! (Not for long, it seems). Eduardo > I have an hp48 (manufacture in '97) that I bought new and has been > giving me problems since day one. It sporadically tells me that my > memory is full and I need to purge it, usually in the middle of doing > something important, ie. and exam. I have reset, that doesn't work. > I have NEVER programmed with the calculator or downloaded anything to > it. I know, I know, I should have dealt with it 5 years ago when I > bought it, but what can I say. This is very frustrating, to say the least. I love RPN, like the > keypad of the 48, and don't think I would need the bigger memory of > the gx. I am returning to school at the age of 34 to finish my EE > degree, and would like to have a reliable RPN calc. I have a ti-89 > that I bought after seeing all the cool things it can do, but I just > can't get used to the interface. Funny how the only way I can stand > it is to use it in a half assed RPN way by pasting the answers instead > of using parenthases. 5 keystrokes to find a square root... well, I'm > preaching to the choir here I suppose. I am a newbie to a forum like this, scoured previous posts and didn't > see this problem (not to say it hasn't been answered somewhere). Is > my memory chip fried? do I need to suck it up and find an old one on > e-bay? or should I just learn to use the ti-89 and deal with it? > Corey Christianson > Pullman, WA ==== If its a HP48G with 32k ram your stack may filling up with ram. Have you tried clearing the stack or checking what is using the memory. > I have an hp48 (manufacture in '97) that I bought new and has been > giving me problems since day one. It sporadically tells me that my > memory is full and I need to purge it, usually in the middle of doing > something important, ie. and exam. I have reset, that doesn't work. > I have NEVER programmed with the calculator or downloaded anything to > it. I know, I know, I should have dealt with it 5 years ago when I > bought it, but what can I say. This is very frustrating, to say the least. I love RPN, like the > keypad of the 48, and don't think I would need the bigger memory of > the gx. I am returning to school at the age of 34 to finish my EE > degree, and would like to have a reliable RPN calc. I have a ti-89 > that I bought after seeing all the cool things it can do, but I just > can't get used to the interface. Funny how the only way I can stand > it is to use it in a half assed RPN way by pasting the answers instead > of using parenthases. 5 keystrokes to find a square root... well, I'm > preaching to the choir here I suppose. I am a newbie to a forum like this, scoured previous posts and didn't > see this problem (not to say it hasn't been answered somewhere). Is > my memory chip fried? do I need to suck it up and find an old one on > e-bay? or should I just learn to use the ti-89 and deal with it? > Corey Christianson > Pullman, WA ==== > Hi: > > Marking is: > (YB71) > That's a mosfet. > If you look at it with the three pins down and the wider pin up, > the pinout from left to right is G-D-S. > > Greetings > > Steve Sousa > Hi, > > On my HP48G the transistor T1 on the left bottom side of pcb is out! > It's a transistor (with 4 pins) at the same level of the blue self. > If you can read the reference of this component, can you send me it. > ==== I was hoping to get a HP39 but they are no longer available. Same with HP49 and HP48. Australian distributor could give me no indication of availability. Best on offer from HP site is the HP9g or HP30s which both appear to pretty standard off the shelf algebraic calculators. > Hi! HP 48G/G+ might be ideal or, if the HP 39G/40G is cheaper, then that. > The statistic samples are gathered into a matrix so individual entries > can be examined later and naturally the computing precision is better. > One can also find huge amount of statistical programs for these models. > www.hpcalc.org > > It appears the new HP idea of an advanced scientific calculator is a > Casio > knock off. Very sad. I was looking for something to get my son and > daughter but HP's current range has nothing to recomend them. > > > I am a die-hard HP49 user. My wife is a psychology student and is > > currently looking for a calculator that will do statistics. I want > > her to buy an HP, but I'm having a hard time convincing her to get > > one. I went to their website, and couldn't find anything there that > > was better than TI's equivalant. > > > > What is HP's good statistical calculator? Where can I find it? > > ==== HP-39Gs are available on ebay quite frequently. I see several with a buy-it-now price of about $25. Gene -- * These statements and opinions are mine alone and do not reflect my employer's views. * > I was hoping to get a HP39 but they are no longer available. Same with HP49 > and HP48. Australian distributor could give me no indication of > availability. Best on offer from HP site is the HP9g or HP30s which both appear to pretty > standard off the shelf algebraic calculators. > Hi! > > HP 48G/G+ might be ideal or, if the HP 39G/40G is cheaper, then that. > The statistic samples are gathered into a matrix so individual entries > can be examined later and naturally the computing precision is better. > One can also find huge amount of statistical programs for these models. > www.hpcalc.org > > > It appears the new HP idea of an advanced scientific calculator is a > Casio > > knock off. Very sad. I was looking for something to get my son and > > daughter but HP's current range has nothing to recomend them. > > > > I am a die-hard HP49 user. My wife is a psychology student and is > > currently looking for a calculator that will do statistics. I want > > her to buy an HP, but I'm having a hard time convincing her to get > > one. I went to their website, and couldn't find anything there that > > was better than TI's equivalant. > > > > What is HP's good statistical calculator? Where can I find it? > > > > > ==== Classic. Toby > ... > I think you could get used to their nakedness after a short while..... and it > would save you quite a bit of money. ==== However, since even the naked RAM cards are usually inside the HP-48, it doesn't matter that much;-) Raymond Toby schrieb im Newsbeitrag > Classic. Toby > > ... > I think you could get used to their nakedness after a short while..... and > it > would save you quite a bit of money. > ==== I«am VERY sorry if somebody got offended, that was not my intention. Joakim Margell > Hey > > Is it much harder to make the program interface in english? I have found > many interesting programs at > www.hpcalc.org but i can«t use them because, I don«t understand a singel > word Spanish, Portuguese or French. And I hope I«am not the only one... > > > Joakim Then don't buy them. Oh wait a minute, they're free. Maybe those > guys shouldn't get paid to ite these programs. Oh wait a minute, > they don't get paid. You could always ite your own programs, instead of complaining about > the guys who are kind enough to ite the best programs they can and > distribute them to a population who is often unappreciative and overly > critical. You could also learn Spanish, Portuguese or French. Then you could > use the programs that are available already. I have done this with > some programs. I didn't have to become fluent, just learn enough to > translate what little text there was. Math is a language in itself. And finally, instead of throwing out a blanket complaint that will > almost certainly offend a lot of people in this NG, you could maybe > ask if anyone knows of the specific program(s) that you're looking for > with an english interface. This one is probably the most constructive > method. Since there is a lot of people in the world who don't speak fluent > english (world's full of 'em), maybe for some people, iting a > program with an english interface is not only difficult but useless. > I don't speak Spanish, and would therefore have no reason to ite my > programs in Spanish. It would make them harder to use. Why would > someone who speaks fluent Spanish ite their programs in English? In summary: Beggars can't be choosers and there are probably better > ways to accomplish you task. --CS ==== ÁRight on, hermano! Toby > Hey > > Is it much harder to make the program interface in english? I have found > many interesting programs at > www.hpcalc.org but i can«t use them because, I don«t understand a singel > word Spanish, Portuguese or French. And I hope I«am not the only one... > > > Joakim Then don't buy them. Oh wait a minute, they're free. Maybe those > guys shouldn't get paid to ite these programs. Oh wait a minute, > they don't get paid. You could always ite your own programs, instead of complaining about > the guys who are kind enough to ite the best programs they can and > distribute them to a population who is often unappreciative and overly > critical. You could also learn Spanish, Portuguese or French. Then you could > use the programs that are available already. I have done this with > some programs. I didn't have to become fluent, just learn enough to > translate what little text there was. Math is a language in itself. And finally, instead of throwing out a blanket complaint that will > almost certainly offend a lot of people in this NG, you could maybe > ask if anyone knows of the specific program(s) that you're looking for > with an english interface. This one is probably the most constructive > method. Since there is a lot of people in the world who don't speak fluent > english (world's full of 'em), maybe for some people, iting a > program with an english interface is not only difficult but useless. > I don't speak Spanish, and would therefore have no reason to ite my > programs in Spanish. It would make them harder to use. Why would > someone who speaks fluent Spanish ite their programs in English? In summary: Beggars can't be choosers and there are probably better > ways to accomplish you task. --CS ==== > I«am VERY sorry if somebody got offended, that was not my intention. > > Joakim Margell Not offended at all. I am full of work now, but perhaps I could find some time for helping you with spanish... ==== > I«am VERY sorry if somebody got offended, that was not my intention. Joakim Margell > > Hey > > > > Is it much harder to make the program interface in english? I have found > > many interesting programs at > > www.hpcalc.org but i can«t use them because, I don«t understand a singel > > word Spanish, Portuguese or French. And I hope I«am not the only one... > > > > > > Joakim > > Then don't buy them. Oh wait a minute, they're free. Maybe those > guys shouldn't get paid to ite these programs. Oh wait a minute, > they don't get paid. > > You could always ite your own programs, instead of complaining about > the guys who are kind enough to ite the best programs they can and > distribute them to a population who is often unappreciative and overly > critical. > > You could also learn Spanish, Portuguese or French. Then you could > use the programs that are available already. I have done this with > some programs. I didn't have to become fluent, just learn enough to > translate what little text there was. Math is a language in itself. Ha-haa! T .8ast .8a l .8ahtien kaikki neuvoni ovat suomeksi, kunnes pyyd .8at anteeksi! T .8am .8an p .8aiv .8an yleiskieli on englanti, jota tietenk .8a .8an ei ole *pakko* huomioida! > And finally, instead of throwing out a blanket complaint that will > almost certainly offend a lot of people in this NG, you could maybe > ask if anyone knows of the specific program(s) that you're looking for > with an english interface. This one is probably the most constructive > method. > > Since there is a lot of people in the world who don't speak fluent > english (world's full of 'em), maybe for some people, iting a > program with an english interface is not only difficult but useless. > I don't speak Spanish, and would therefore have no reason to ite my > programs in Spanish. It would make them harder to use. Why would > someone who speaks fluent Spanish ite their programs in English? > > In summary: Beggars can't be choosers and there are probably better > ways to accomplish you task. > > --CS ==== > I«am VERY sorry if somebody got offended, that was not my intention. > > Joakim Margell > ... Ha-haa! > T .8ast .8a l .8ahtien kaikki neuvoni ovat suomeksi, kunnes pyyd .8at anteeksi! > T .8am .8an p .8aiv .8an yleiskieli on englanti, jota tietenk .8a .8an ei ole *pakko* > huomioida! > first try at understanding! Toby It looks cool, though... ==== Dear all. (that I dispatched for beta-testing to close people). The entry points were located at the address 4EA37 (TICKSmin) etc... Using an entry point that has an address > 40000 is really bad and I can guarantee that any programs using these entry points will make your calculator crash in any ROM revision (from 1.05 to 1.19-7) The documentation listing such entry points are (but not limited to:) -Programming in System RPL by Eduardo Kalinowski and Carsten Dominic -EXTABLE by Carsten Dominic -EXTABLe by Thomas Rast The problem with any pointers located at an address > 40000 is that it's not guarantee that they will always be visible. The ROM between 40000 and 7FFFF changes constantly. A little explanation, you can skip to the conclusion if you want to. As you probably already know, the HP49 memory map is made of three different views: ROM view 0 (0 to 3FFFF) ROM view 1 (40000 to 7FFFF) RAM (80000 to FFFFF) ROM view 0 can be switched between 4 different flash banks (bank 0 to 3) ROM view 1 can be switched between any flash banks (bank 0 to 15) ROM view 0 in practice is always configured to see the bank 1 of the Flash ROM (except if this flash bank has been damaged, then it will be another one. I'm yet to see any HP49 with a damaged flash bank) Bank 1 contains the RPL kernel (memory management, display, flash support, math library and all reserved entry points) ROM view 1 is switched between different flash bank containing different part of the firmware: Bank2: Main system outerloop, printing, solver, rpl compiler, plotter, matrix, self-test Bank3: Input Form manager, TVM, ChooseBox manager, Stack display, Table view, Catalog, Object Viewer) algebra, constant library etc.. iter, HP49 UI engine + HP49 input forms Bank6: CAS Bank7: CAS2 When you're in the standard stack display the main OuterLoop is running so the typical view is: Bank 1 | Bank 2 | RAM But in the filer it will be: Bank 1 | Bank 0 | RAM In the CAS it will be: Bank 1 | Bank 6/7 | RAM In the Equation iter it will be: Bank 1 | Bank 5 Etc... The entry points with an address > 40000 can be in any of these banks, and are not visible at all time. All the entry points referenced in the different documentations are from Bank 1. Probably because it's the most called bank and also because if you start a program from the main stack (like a program searching for reals) it will see Bank 1 | Bank 2 But start the same program from the Filer, it will be: Bank 1 | Bank 0. Eg: If you run any programs using these totally invalid entry points from the Filer, the equation iter it will crash: 100% guarantee If you create a user function calling one of these programs and run any math commands on it: it will crash 100% guarantee. Conclusion: Never use any entry points located between 40000 and 7FFFFF. These entry points are reserved for the system only. Any documentations referring to these entry points should be corrected ASAP. There have been several bug reports of system commands crashing when called from the Filer: it was always the same reason: by mistake we called some entry points between sub-system banks. Hopefully we found all of them. You've been warned. ==== > ROM view 0 (0 to 3FFFF) X > Never use any entry points located between 40000 and 7FFFFF. Can you please give even more general information of the banks and combinations. Just for fun! I'd like to know more, more... ******************************************************* BTW: Is it possible to change the bank switching code to handle 64KB banks the same size as the Intel flash chip banks are. This would allow a much finer grade of usability and even a boost of operation speed. Just an example (I don't know if this is rational at all) *0-64: _resident_ OS basic bank including bank-switching code, etc. management [switched only with bootstrap code: the permanent boot-block] 64-128: rest of the RPL kernel | or ??? 192-256: Stack display, Catalog, Object Viewer | UI engine, input forms, etc 256-320: RAM | CAS | rpl-compiler, plotter, matrix. self-test | or whatever 320-384: RAM | CAS | InputForm, TVM, ChooseBox | or whatever *448-512: _resident_ 64KB user main RAM Who needs 256KB RAM? Maybe this could be selectable by user: the RAM could be divided between IRAM/ERAM or main/port 1 RAM freely between 64-128-192-256 (320??) for main RAM use A) big RAM, causing a lot of bank switching and long delays in GC OR B) small RAM and virtually no (or very little) bank switching and also faster Garbage Collecting (-: How do you like my ROM 2.03 idea? Others? ==== Are you still slightly annoyed when the CAS starts asking you to change modes and delete your VX variable every time that you invoke a CAS command, even if the VX variable has nothing to do with the command you want to use (but perhaps some other variable does, as with ISOL and SOLVE, for example, but the CAS doesn't even know it)? Okay, here's another version of XC, a program which executes any other CAS command, but it never asks you to change modes or delete variables. Example of use: 123 'X' STO 'X' STOVX @ create 'X' and set VX to 'X' 456 'Y' STO @ create 'Y' and store something 'X+Y=5' 'Y' { SOLVE } XC @ solve this equation for 'Y' By the way, this version of XC even cures the old problem that if you had the same variable created in more than one directory in the heierarchy (e.g. in both HOME and a current subdirectory), then the CAS command would formerly fail anyway, even if you let the CAS delete *one* of those variables for you, but now even this is fixed. Download and run the following to create the new XC program (and another function VRENP which it calls in turn), but BEFORE you even start downloading, please first do HOME 256 ATTACH, otherwise this may not compile correctly. You need ROM version 1.19-5 or later to use the RCLVX command; substitute #C4007h FLASHEVAL for each RCLVX if your ROM is older. %%HP: T(3)F(.); @ include when downloading using Kermit @ Did you do HOME 256 ATTACH even before you started downloading? @ 'name' -> (rename all occurrences of variable in current path) D9D200036234133D9D20881309EB5079E6031BC39136251B502A170D9D20 03D53B7970A3B63CA620400960018A30B2130F046385230B2130B2130 + DUP BYTES DROP #DB72h == { H -> } IFT 'VRENP' STO @ if VRENP now contains a string, @ then there was a data transfer (or manual typing) error << IF OVER TYPE 6. == THEN OVER ELSE RCLVX END -> v. << v. VRENP RCLVX VRENP PUSH -120. SF IFERR EVAL THEN 1. ELSE 0. END POP v. S~N NOT S~N VRENP RCLVX S~N NOT S~N VRENP { ERRN DOERR } IFT >> >> 'XC' STO @ End of downloadable file Since the built-in calculator menus and equation iter, etc., will not be able to use this directly, you might want to make yourself a custom menu of your favorite CAS commands, so that each one will be executed via this fixer, when invoked from your own custom menu, e.g.: { { DERVX << { DERVX } XC >> } { INTVX << { INTVX } XC >> } { SOLVE << { SOLVE } XC >> } { TEXPAND << { TEXPAND } XC >> } { } @ Etc. STOVX } MENU @ becomes the CUSTOM menu in current (and lower) directory If you ever find any variable accidentally renamed, just put its current weird-looking name on the stack and invoke VRENP to restore the original name. SysRPL source for VRENP (58.5 bytes, #AAFEh checksum): ( ID -> rename all occurrences in current path ) :: CK1&Dispatch SIX :: DUP ID>$ ' xNOT EvalNoCK $>ID BEGIN :: 2DUPSWAP @ NOTcsdrpfls FPTR2 ^RENAME TRUE ; NOT_UNTIL 2DROP ; ; [r->] [OFF] . ==== XC is a program which executes CAS commands so that they don't delete variables, don't ask about mode changes, and don't leave any modes changed after executing. The use of PUSH and POP in the version just posted may noticeably slow it down, because those commands keep appending and removing data saved in a permanent list in CASDIR. Just use TEVAL to time the do nothing program << PUSH POP >> to see how much time this may add in your own calculator, which may vary according to how much is stored in HOME. So here's a faster alternative for XC which avoids PUSH and POP: << IF OVER TYPE 6. == THEN OVER ELSE RCLVX END RCLF PATH -> v. f. p. << v. VRENP RCLVX VRENP -120. SF IFERR EVAL THEN 1. ELSE 0. END f. STOF p. EVAL v. S~N NOT S~N VRENP RCLVX S~N NOT S~N VRENP { ERRN DOERR } IFT >> >> 'XC' STO Note that it still uses VRENP, found in the preceding post. You are also welcome to re ite it all in SysRPL to make it faster still, if you're that much a Type A person :) The basic reworking of this version over the version posted last year is that the actual deletion of variables, ordinarily forced by the CAS, is now replaced by the temporary hiding of the same variables, via renaming. This renaming is done over the entire heierarchy of directories in the path, matching the coresponding search and formerly forced deletion on the part of the CAS. We also don't now distinguish commands which use variable names supplied on the stack instead of VX, as is actually most common; last year's version temporarily changed VX to correspond with the possibly different variable name on the stack, whereas we now rename both the VX variable *and* any variable named on the stack (this doesn't take much time, because we use names that remain exactly the same length, so that nothing moves in memory). The temporary names we use for hiding variables (instead of purging) are actually the bitwise complements of their original names, which have a very low probability of matching any other existing variable names (if one did match, the final result would only be a renamed existing variable, which you could rename back yourself). This scheme almost always works out fine, giving us the desired proper results, but with no more hassle. [r->] [OFF] . ==== > > Do you mean a user interface like the 48 or 49? That would mean > re iting a lot of things: SOL, keys, parser... : although the 48 > palparse is in the 38/39 firmware, unfortunately it is not complete > (that would have been a waste of memory!), the new elsieparse is used > instead. Consider buying an HP48 :-) > > > HPCC #1046 Buying a HP-48 is to expensive. Someone interested in making a simplified RPN program for the HP-39/40g? ==== I have finished my game Dope Wars, which is the adaptation of the PC game with the same name. Include a full manual (french & english versions), enjoy! You can donwload it on my website: httt://www.noda.online.fr HP48&49 versions will follow soon... ==== Why would some non engineer, scientist , or numbers geek be interested in rpn, the large stacks of storable values, or integral approximation?? Bri > A problem that I bet is common to everyone in this newsgroup: trying > to explain the power of HP calcs to someone who isn't into math or > calculators. Most people seem to think that a calculator, no matter > how powerful, is just a glorified four-function type. If you try > explaining, their eyes glaze over right away. I was trying to describe the HP49 to a woman here at work, and she > kept saying, Oh, so it's like a Palm Pilot. Mmm, no. I gave up. Bill Speechless Markwick ==== A problem that I bet is common to everyone in this newsgroup: trying to explain the power of HP calcs to someone who isn't into math or calculators. Most people seem to think that a calculator, no matter how powerful, is just a glorified four-function type. If you try explaining, their eyes glaze over right away. I was trying to describe the HP49 to a woman here at work, and she kept saying, Oh, so it's like a Palm Pilot. Mmm, no. I gave up. Bill Speechless Markwick ==== > Why would some non engineer, scientist , or numbers geek be interested in > rpn, the large stacks of storable values, or integral approximation?? > They're not. What they're interested in is why *we* are and why we spend so much time tapping at the little box. Very difficult to explain. Sometimes I think it's best to tell them it's the latest Gameboy and let it go at that. Bill ==== approximation. Geeks are interested, and others well, nope. Brian > Why would some non engineer, scientist , or numbers geek be interested in > rpn, the large stacks of storable values, or integral approximation?? > They're not. What they're interested in is why *we* are and why we > spend so much time tapping at the little box. Very difficult to > explain. Sometimes I think it's best to tell them it's the latest > Gameboy and let it go at that. Bill ==== At the time when HP was having only RPN machine, explaining HP was the same as explaining RPN. I told people that RPN is like they were taught how to do it in school. For example, when adding two numbers, you do: 1 - ite the first number 2 - ite the second number under the first 3 - Add the two numbers It translate in RPN with: 1 - Type the first number 2 - Press the ENTER key to tell the machine you have finished entering the first number. This action puts the number in the stack (similar to what will happen if doing it manually on a piece of paper) 3 - Type the second number 4 - Press the required function key ( + in this example). Then I explained to them how the stack works by keeping the results so you can use them when solving more complex equations. Then I finished by explaining the following procedure that can solve any equation (in numeric form, of course): 1 - Type a number. when finished typing, go to step 2. 2 - If you can perform a single argument function on that number,(like SIN, COS, SQUARE ROOT, LOG, etc.) then do it. This tells the calculator you have finished entering the number and to perform the function. The result will be flagged as such and will be pushed on the stack by the next typing of a number key. This step can be repeated as often as required for imbeded functions. Go to step 4. 3 - If you dont need to perform a single number function, then press ENTER to tell the calculator that you have finished entering the number. The number will become like a result and be flagged as such and will be pushed on the stack. 4 - Type a second number. 5 - Do step 2 if required. 6 - If you need to perform a two number operation on the numbers present on the stack, do it. Again, the result will be flagged as such and will be pushed on the stack by the next typing of a number key. 7 - If you dont need to perform a two operation on a number, then press ENTER to tell the calculator that you have finished entering the number. The number will become like a result and be flagged as such and will be pushed on the stack. 8 - If not yet finished, go to step 1. This can be simplified as follow: 1 - Type a number. 2 - If you can perform a function with that number, do it. 3 - If not, press ENTER. 4 - Go to step 1 untill you complete the calculation. Then I explained how the four level stack could present some limitations with very complex equations and how to solve these by starting in the middle of the equation, when possible. Or how to use the registers to store intermediate results. With the advent of more powerfull machines and stacks limited by RAM, this became obsolete. I also explained that, with RPN, you dont need to bother with the precedence of operations like doing multiplication and division before addition and subtraction, etc. To conclude, I think, that even with simple maths, outside of the engineering, the scientific or the finance worlds, a simple RPN machine would do marvel because, those using it would prbably lack the mathematical understanding of operation precedence and the like. But I am pretty certain that they could grasp easily the concept of RPN with a 4 steps procedure as stated above. Just ideas. Jean (Johnny) Lemire from Richelieu, Quebec, Canada ==== Surveying problem: knowing the state plane (x and y) coordinates of two points, find the azimuth from one to the other, and save the azimuth in a variable. What I did was set the calculator to polar mode, input both [x,y] coordinates, and subtract. The answer was displayed in [r,theta] form. Good. But grabbing the angle and performing a calculation on it or storing it was awkward. When I tried to EDIT the vector, Matrix iter displayed it in rectangular format. Also tried the VEC-> command, and it too produced the elements in rectangular form. I found a way to get the angle, but it's clumsy. 1. Press the up arrow key. 2. Press EDIT softkey. 3. Move to beginning of angle, do BEGIN. 4. Move to end of angle, do END. 5. COPY. 6. Press CANCEL twice. 7. PASTE the angle to the command line. Is there an easier way to get the angle part of a two dimensional vector? -- For my real address, change the last letter in the alphabet to the first letter (two places). Reply-To: raut@math.ful-berlin.de ==== > Therefore, if you use marks randomly, you will fill eventually all 9 > slots and the data will sit there. I believe that hardly anybody will > be able to remember more than 3 jump marks, so I followed Wolfgangs > suggestion to limit the number. I think your method of storing with a jump 3 lines of context is OK. 3 jumpmarks should suffice even for very large source text. I did not yet test the new Emacs but had some problems with loading extable2. This lib has 110,308.5 bytes (CRC 1F24). In order to save time, I normally load files above 5 KB with the HP49G Connectivity kit via XMODEM. It has an acceptable transfer rate of about 830 b/s. That works fine with Emacs, but not with the huge extable2. I got it as a HP49... string on two different HP49. Hence, perhaps some hint should be added in Emacs.txt for beginners how to manipulate large objects close to 128 KB, perhaps slightly corrupted by the transfer procedure. Suppose one has at least 130 KB free RAM (not counting port1 and port2). Applying ObFx from OT49 as recommended in the 12 devices for the HP49 by Tim Wessman does then only work if 1. the corrupted string is on the stack, 2. the file extable2.lib is purged, 3. ObFx is applied, 4. the lib is immediately sent to Port1 or Port2, not first returned to HOME. - Wolfgang ==== > > I am really sorry about the Greek post... > > It won't happen again :-) > > Yes-yes-yes IT SHOULD happen again > > BUT > > give us the English translation > Why is that so? > Are you interested in learning Greek? > Or are you just interested in the structure of > the Greek language? > As English is the current international language > and as this is an international news group > I wish everybody who wishes to use their own native > language (other than english) here > would at least provide an AltaVista translation of it. > PS: I read the New Testament in FIN/ENG/GRK You read Greek!?!? This is a very pleasant surprise! Anyway, I admit that it is not appropriate to ite in Greek in an international NG. I apologize again and I will avoid doing so in the future. No-one denies that English is an international language nowadays. Just like Greek used to be some 2400 years ago, during the period that Alexander The Great lived ;-) Alex Markatis Civil Engineer GREECE ==== > > Anyway, I admit that it is not appropriate to ite > in Greek in an international NG. I apologize again and I > will avoid doing so in the future. Of course, if we assume that this is not an *inter*national, but instead a *multi*national NG... :-) > No-one denies that English is an international language > nowadays. Just like Greek used to be some 2400 years ago, > during the period that Alexander The Great lived ;-) Hey, statistically 100% percent of the empires/kingdoms that used Greek as the official language have vanished from the face of the earth (disregarding architectural leftovers). So it is probably a wise choice the move to English. -- Alexandros Andreou, ee4299 at ee.teiath.gr. Undergraduate student, Department of Electronics, TEI of Athens. Life's not a bitch; she's a BOFH. ==== > No-one denies that English is an international language > nowadays. Just like Greek used to be some 2400 years ago, > during the period that Alexander The Great lived ;-) > > Hey, statistically 100% percent of the empires/kingdoms that used Greek as the > official language have vanished from the face of the earth (disregarding architectural > leftovers). So it is probably a wise choice the move to English. It was a joke synonomate, BUT I don t see any harm in refreshing our historic memory from time to time, even by joking. I really think that the people who read this NG are not monoxnwtoi , that is, they don t care ONLY for their Calcs, but they are interested in knowledge in general. Should I be ong with this, then I d better stop reading this NG, since I am a supporter of the olistikis theorisis of life. Besides, sometimes the situation here in Greece makes we (Greeks) have left :-( What s ong with sharing these memories with a few million people :-) Perasmena megaleia kai dihgwntas ta na klais Alex Markatis Civil Engineer GREECE-HELLAS ==== > I really think that the people who read this NG > are not monoxnwtoi , that is, they don't care ONLY > for their Calcs, but they are interested in knowledge > in general. Should I be ong with this, then I d > better stop reading this NG, since I am a supporter > of the olistikis theorisis of life. I hope you're right about the general readership, although one can wonder about quite a few who express themselves in iting here, since it receives so little overt support, while its direct opposite, the off topic -- has no business being discussed in this newsgroup mentality gets loudly and often expressed, especially by folks who come out upon any excuse to actually attack anything which might widen and broaden our complete experience and perspective on our relationship to the whole of our universe. The educational institutions where I work also consider the holistic approach, the integration of all knowledge, to be essential, especially that level of integration which brings us full knowledge and experience of our own complete self. This is why every class on any particular subject always concludes here with a segment called connecting the parts of knowledge with the whole of knowledge, where the whole of knowledge is inclusive of the knower himself, at his deepest level of self perception and self knowledge. What else can possibly unify the world into something where everyone supports everyone else, instead of where small subsets act only for themselves? There is nothing new about the value of full, holistic education, including the relationhip between all objects of knowledge and the knower who studies them as if they were separate from and entirely unrelated to himself; here, for example, is Marcus Aurelius speaking of this, nearly 2000 years ago: This thou must always bear in mind, what is the nature of the whole, and what is my nature, and how this is related to that, and what kind of a part it is of what kind of a whole; and that there is no one who hinders thee from always doing and saying the things which are according to the nature of which thou art a part. Very little is needed to make a happy life; it is all within yourself With best wishes from http://www.mum.edu . Reply-To: Gregory Warsewicz Savage ==== [ snip ] This is quickly getting Way Off Topic. Alexandros Andreou > Hellas Hey! I though this was forum for geeks, not Greeks.? ;-) or rather... (-: ==== > Who is qualified enough to judge people and > categorize them into thinking and not-thinking ? > Who and why is qualified enough to propose, vote on and pass laws in the Parliament? > The lack of Morphosis , the lack of Paedeia (education) > is the reason for all these. > All people are shaped (morfvnontai) into something, with which education has nothing to > do. However, you just defined `us' and `them' the same I way I usually do: the people that > are educated and the people that are not. You are saying that education doesn t shape our character? Then what are we all going to school for? Only to learn reading and iting? Education determines an important part of our morphosis (not 100% of it, though). But, I am talking about morphosis and paedeia here, not school education or specialization in a working field. > By `educated' I do not mean `people that are enrolled in Uni courses'; I have, and am > willing to bet that so have you, met people that hold degrees of high importance but are > essentially uneducated, and vice versa. That s exactly my point here! How do you define thinking vs. talking ? University? RPL? Height? Colour? Footbal team? What are your citerias? Who are you to separate people into two groups and then blame the other group for the ong course of the society? Are you not responsible, too? Not even a little? > I am not a follower of the ``we are all responsible -- no-one is able to set the criteria > as to what makes a $THING'' line of thinking. This can lead one to paths such as ``all laws > are void'', ``police has no valid power'' and in general anarchy, in the ``lack of order'' > sense of the word. There are people responsible for a situation and people irresponsible for > the same situation. Surely not equally responsible, but responsible! And I realy can t see the thread that leads from this argument of mine to irrelevant conclusions like all laws are void , anarchy etc. Just for the record: I am NOT an anarchist. On the contrary, I really like ORDER and I like being a member of an organized society, in which everyone and everything has its place. Therefore, I m sure that you can understand how suffocated I feel sometimes here in Greece, where everything happens by chance , without any organization. But, aren t you an anarchist when you question the right of the members of the Parliament to legislate? Instead of blaming them and only them for their (tragic) mistakes, blame also the un-educated people who are willing to exchange their vote with a job at a ministry, without caring who they vote for. But the fact that they cannot appreciate the value of their vote, is also my responsibility. Because I never taught the forthcoming generations that nepotism is not a right thing. No man is an island, integral, on his own; he is a part of the mainland... Every man s death lessens me too, for I am one with mankind. So, never send someone to ask for whom the bell rings. It rings for you. --- John Donne > This is quickly getting Way Off Topic. I know, but it doesn t matter! Read the post by John H Meyers above :-) Alexandros Markatis HELLAS ==== Use AltaVista or other computerized translators, but with some reserve. They are far from accurate with even simple sentences. For example I used AltaVista with: I tried to explain that concept to my friends Translated in French I obtained: J'ai essay .8e d'expliquer ce concept .88 mes amis Which is excellent. Then I used the the french result to go back to english and I obtained: I tried to explain this concept with my friends Which is quite different from the original. Imagine with more complex sentences. Use with care. Jean (Johnny) Lemire from Richelieu, Quebec, Canada ==== I'm currently porting Usinagaz to hp38/39/40, but I'm having troubles with silly things like login & password: how can I ask the user to enter it? I've itten a DoInputForm with Text fields of expected type 2, and yet, although one can type ' ' thanks to the CHAR tool, one can only get syntax errors... It might work with some Notepad variables, but how can I get them back? -- Samuel Thibault Reply-To: Samuel Thibault ==== > I've itten a DoInputForm with Text fields of expected type 2, and yet, > although one can type ' ' thanks to the CHAR tool, one can only get syntax > errors... OBTYPE_NA as legal object types, while I was using { OBTYPE_CSTR }, which would require the user to be able to enter a string, which is not the case. OBTYPE_NA shortcuts any compilation, so that I really get a string :) -- Samuel Thibault ==== Could you either send it to me also or give me an address where I can get it? I'll put both of them on my site so that the location will (hopefully) have some stability instead of HP's here now, gone tomorrow policy. I already have the 39/40G manual (in pdf) that I have added extensive hyperlinking to on my site (and corrected a few errors that crept in), as well as the CAS manual in english. PS. Why two versions of the CAS manual? Would anyone want version 2 if version 3 is better? > > French CAS manual, version 2.0: > http://www-fourier.ujf-grenoble.fr/~degraeve/jean.pdf > > French CAS manual, version 3.0: > http://www-fourier.ujf-grenoble.fr/~degraeve/jeanmark.pdf > > if I can find it. > > HPCC #1046 > > ==== > I get such a rush when trying to do something with my nifty little calculator. > I extend my love to all the ACO members, the people who surround this great > community injecting small gems of wisdom into it and everyone in general. I'm > in a hippy (HP...) ``me loves universe'' mode right now. > That's great! After reading so many stupid and insulting flames, it's wonderful to see something like that. And if it's chemically induced, pass it around! :-) Bill ==== I was wondering if there is a snip_it area for user RPL and/or sys Rpl residing somewhere within this messgae area perhaps? ==== > > While it's not on your list, I've found that un 'unknown' yet very > powerful mathematics software is 'Derive'. It not only handles > numerics to arbitrary (user defined) precision, complex variables, > matrix/vector data types, but also handles symbolics very well. > Further, it costs just about $100.00 (maybe $150, don't recall > offhand). It's not quite as pretty as far as user interface to Mcad > and the others, but it certainly is more affordable. You can download > a fully functional trial (I believe 30 day) version from their > website. You can find it via your fav search engine. A dealer down > in Miami, Fla discounts it a bit, too. > > TomCee > One can run the DOS version of DERIVE on a Hp200LX which is extremely handy, IMHO. And there is MiniMat which one can sometimes find which was the academic precursor to the commercial version of Matlab and also will run just fine on the 200LX. Do a Google search for computer algebra or symbolic algebra and you'll find that there are all sorts of other interesting programs out there--REDUCE, Fermat, YACAS ( yet-another-computer-algebra-system ), Maxima (nee Macsyma) and our good friend Parisse's own XCAS among them. Roger Metcalf ==== I don«t think I have to say more... Read topic to answer my question... i.e. I have f(x) = 10^(2,55-0,15*X), where X is ASTM (standard) grain size, varying from X= -1 to X=15. I want a straight line on my HP, if it«s possible... Martin J¿rgensen. begin 666 Martin Jorgensen.vcf ==== I don«t think I have to say more... Read topic to answer my question... > i.e. I have f(x) = 10^(2,55-0,15*X), where X is ASTM (standard) grain size, > varying from X= -1 to X=15. I want a straight line on my HP, if it«s possible... Martin J¿rgensen. How about plotting log(f(x))? ;-) Of course, it makes it more difficult to trace... Good point, though. Have fun, Gerald. ==== I have download : RF v2.4 - Redundance Fighter compressor from http://www.hpcalc.org I send the file to the HP48GX. but when i want to install this lib (Lib : 1789...) I have only a string with the first letters are HPHP48-E@+ *..... instead of Library 1789: .... If someone have an idea ??? Cedric ==== Use this little tool : http://www.hpcalc.org/search.php?query=objfix cedric skrev i en meddelelse I have download : RF v2.4 - Redundance Fighter compressor > from http://www.hpcalc.org I send the file to the HP48GX. but when i want to install this lib (Lib : 1789...) I have only a > string with the first letters are > HPHP48-E@+ *..... > instead of > Library 1789: .... If someone have an idea ??? Cedric ==== > Also updated is my UTOOL library for units handling - I added some > support for relative units. Those are units like degree C or Gauge > pressure which are defined relative to some arbitrary zero point and > normally make conversion a pain. This is likely to be the final > version of this library - cannot think of anything to add. Great work! It is exactly what I need (capability of relative units, i.e. conversion between gauge - absolute pressure). Kenji nagahamaknjNOSPAM@mx10.ttcn.ne.jp ==== > I have an HP-48GX with a crack in the LCD screen. Otherwise > everything works fine. Can anyone tell me if it can be fixed or even > if it would be worth it? If not does anyone have any ideas as to what Jesse, Courtesy of French baggage handlers and my failure to adequately cusion my HP-48GX, I also have damage to my LCD screen (dark spot and missing horizontal lines that obscure part of the screen). I haven't found a way to repair it yet, and I think the cost of doing so, if such a repair is even commercially available, would not be worth the cost of a new calculator. If your calculator is unusable with the damage, you have engineering know-how, and have time to spare, you can probably crack the case open and might find a suitable LCD screen of the right size to replace the damaged one (or with some trimming of the plastic case). Afterall, if the calculator is unusable, you have nothing to lose but time. As for me, I don't have the time to figure out how to break the connections and interface to a new LCD, so I plan to buy a new HP-48GX and use my old one as a one-line non-graphing calculator. Ryan ryan_shepperd@yahoo.com In-Reply-To: <3E572670.3070302@yahoo.com> ==== > >> I have an HP-48GX with a crack in the LCD screen. Otherwise >> everything works fine. Can anyone tell me if it can be fixed or even >> if it would be worth it? If not does anyone have any ideas as to what > > > Jesse, > > Courtesy of French baggage handlers and my failure to adequately cusion > my HP-48GX, I also have damage to my LCD screen (dark spot and missing > horizontal lines that obscure part of the screen). I haven't found a > way to repair it yet, and I think the cost of doing so, if such a repair > is even commercially available, would not be worth the cost of a new > calculator. If your calculator is unusable with the damage, you have > engineering know-how, and have time to spare, you can probably crack the > case open and might find a suitable LCD screen of the right size to > replace the damaged one (or with some trimming of the plastic case). > Afterall, if the calculator is unusable, you have nothing to lose but > time. As for me, I don't have the time to figure out how to break the > connections and interface to a new LCD, so I plan to buy a new HP-48GX > and use my old one as a one-line non-graphing calculator. > > Ryan > ryan_shepperd@yahoo.com > > hi, I have just repaired one yesterday. With 2 hp48, one hp48g with a good lcd screen but with a destroyed motherboard and one hp48gx with a broken screen, I made a working hp48gx. I have just exchanged the motherboards and the trick was done. HPSB ==== Press and hold [ON] whiile pressing [+] for darker or [-] for a lighter display. >I need to adjust the display on my 48G. Can't find the manual, so can >somebody pls tell me how to adjust brilliance and contrast? Any help much appreciated! >Mads >Odense, Denmark > ==== Press and hold [ON] whiile pressing [+] for darker or [-] for a > lighter display. > Mange tak for det, Torstein :-) VH Mads ==== Try [ON]-[+] [ON]-[-] As I know, only contrast is adjustable. Ricks Madrid, Spain > I need to adjust the display on my 48G. Can't find the manual, so can > somebody pls tell me how to adjust brilliance and contrast? Any help much appreciated! > Mads > Odense, Denmark ==== I just got an HP-41 HP-IL based Thermal Printer (82162A) but it came with no battery pack. I was wondering if anyone knows the specs for the NiCads used in this printer? A colleague of mine has rebuilt several battery packs for several HP calculators and printers and says if I can get the specs. we should be able to fix something so as to get the printer working again. It looks almost brand new except for not having a battery pack. Harold A. Climer Physics/Astronomy Lab Instructor U. Tennessee At Chattanooga ==== use four Panasonic P-170 SCR (KR23/43) . Each off these has a capacity of 1700mAh at 1.2V . This type of battery suits perfectly into the cblack holder. Raymond Harold A. Climer schrieb im Newsbeitrag > I just got an HP-41 HP-IL based Thermal Printer (82162A) but it came > with no battery pack. I was wondering if anyone knows the specs for > the NiCads used in this printer? A colleague of mine has rebuilt > several battery packs for several HP calculators and printers and says > if I can get the specs. we should be able to fix something so as to > get the printer working again. It looks almost brand new except for > not having a battery pack. > Harold A. Climer > Physics/Astronomy Lab Instructor > U. Tennessee At Chattanooga ==== Better use the 2000 mAh Ni-Cad just available now. Raymond Del Tondo a .8ecrit dans le message use four Panasonic P-170 SCR (KR23/43) . > Each off these has a capacity of 1700mAh at 1.2V . This type of battery suits perfectly into the cblack holder. > Raymond Harold A. Climer schrieb im Newsbeitrag > I just got an HP-41 HP-IL based Thermal Printer (82162A) but it came > with no battery pack. I was wondering if anyone knows the specs for > the NiCads used in this printer? A colleague of mine has rebuilt > several battery packs for several HP calculators and printers and says > if I can get the specs. we should be able to fix something so as to > get the printer working again. It looks almost brand new except for > not having a battery pack. > > > Harold A. Climer > Physics/Astronomy Lab Instructor > U. Tennessee At Chattanooga ==== After reading what you have itten... I would bet its the RAM. But consider this, unless you are tooled to remove the IC's... and can get replacement ones... or anonther calculator to butcher... what's the point of proceeding ? I agree with the guy who says get another one and be done with it. Thats what I would do, if I was in your position...... that's what I have done for other defective items I have owned that were far more expensive than hp calculator. Personally I cant be bothered with the hassle... I get pleasure out of USING the technology.. not repairing it. ==== I'm not intending to repair if it is too complicated. If it's related to hardware, it is almost imposible to repair. But anyway, I have a non working calc and still want to know what has failed. I get pleasure out of using technology, but also get pleasure out of learning from it. You are absolutely right to get a new calc. I'm looking for a HP49. I guess I am going to get more from a 49 rather than a 48. Mariano > After reading what you have itten... I would bet its the RAM. But consider this, unless you are tooled to remove the IC's... and can get > replacement ones... or anonther calculator to butcher... what's the point of > proceeding ? I agree with the guy who says get another one and be done with it. Thats what I would do, if I was in your position...... that's what I have done > for other defective items I have owned that were far more expensive than hp > calculator. Personally I cant be bothered with the hassle... I get pleasure out of USING > the technology.. not repairing it. ==== Mariano, I was not REALLY able to help you... but you are quite welcome anyway :) About your NEW CALCULATOR.... I have BOTH an hp48 and an hp49 coming because I couldn't make up my mind... and I wanted to use a plug-in ROM which I like. However it seems from reading the posts that the hp49 is the better (more powerful) of the two.... better except for the quality of the keys and the placement of the enter key... thanks for posting in the forum and let us know how you like your new purchase when you get it ==== I'm looking to trade my HP-41cv with all covers, and manuals and electronically sound with the exception it has need of the repair outlined at this link: and Hp-82104A card reader boxed and looks as new for perhaps Garmin GPS III plus or Garmin GPS V? This is what I have, for the right item you can get it all: 1) an HP-41CV that does comes on when the front and back are held tight in hand. 2) 3 manuals for the HP-41CV, owners handbook, standard app's, and experienced user manual 3) boxed card reader w/manual and blank cards plus card holder 4) Brown HP vinyl case and quick reference guide (I may find the black leather HP case, I'm looking) 5) 3 modules with holder - stat 1 HP 00041-14001 with manual and quick card, hp-41 advantage 5061-7285, memory module hp-82106a with quick ref card If anyone would be interested in this, or perhaps some other trade And yes I can send pictures. I'm ujb on ebay, you can check my feedback. ujb ==== in hpcalc.org in my 49g, using the same method that I use to normally update the ROM of the 49. But when (I suppose) the process was about to to finish, he appeared in the screen of my 49 a message that said a bank had to be erased, that I went to the Erase Banks menu, as soon as I returned to the previous menu and that I chose Download Pack to finish the transference. I went to this menu and I erased bank 0 (I did not dare to erase nothing else), but continues appearing the happy message and I cannot install the ROM of the 39/40 in my 49g. I could not make another thing, then I installed the normal ROM of the 49, process in which I did not have any problem in concluding it. Some of you has tried to do this successfully?. You have managed to install the ROM of 39/40g in the 49g?. Hopefully you can help me in this. Sorry for my ---horrible--- english. Greetings from Valparaiso - Chile ==== 39g: $40 49g: $190 About 5 of each. Martin Cohen ==== I have an old HP41cv, it works just fine, but I have problems with the battery pack. Is there a solution to exchange the battery pack with some more normal batteries or something like it? Please help, I don't want to lose the calc. Have a nice day and thanks for answers:) Borut Korpar borut.korpar@uni-mb.si ==== Note that you have now the 450mAh NI-MH to put in the normal batterie pack (without the recharging inside). b4tbf2$7ct$1@planja.arnes.si... I have an old HP41cv, it works just fine, but I have problems with the > battery pack. Is there a solution to exchange the battery pack with some > more normal batteries or something like it? Please help, I don't want to lose the calc. Have a nice day and thanks for answers:) Borut Korpar > borut.korpar@uni-mb.si ==== What is the difference between a hp49g+ and the hp49g and who sells the ==== as far as i know, there is only one HP49G model produced. ==== > What is the difference between a hp49g+ and the hp49g and who sells the Hmmm! As far as I know the model with a + is the HP48G+ as opposed to the HP48G. The + model has 128KB memory instead of only 32B. !Demeter! Reply-To: Veli-Pekka Nousiainen ==== www.cynox.de HPSpeed1, Double Speed Module for HP48G/GX and HP49G, 30.00 EUR, which is currently installed on my (other) HP 49G, which I may sometimes refer as HP 49G+ - or whatever > What is the difference between a hp49g+ and the hp49g and who sells the options in the cell expression (menu>Format>Show Expression) appopriate ____________________ ==== > The complex number pairs as (r,angle) or (x,y) are restricted to > numeric only. Please change to numeric first before combining to > complex. You may like to use use Approx. and Numeric to key in > complex numbers without entering symbolics. Yes, I see now. Out of frustration, I went ahead and purchased the Urroz HP49g book. What a great help! It has already paid for itself. My HP49g is finally useful now. :-) -- Titus Barik (barik@ieee.org) ==== Most HP49G clocks are not very accurate. One second a day may not seem to be much, but will sum up to a minute in two months. Additionally, the temperature of the HP49G, and therefore the usage of the calculator, influences the clock speed. ClckAdjst is a library, which provides tools for automatic clock adjustments and much more. ClckAdjst provides fast and easy setting and adjustment of the system clock. It handles changes of the time zone and automatic daylight saving time changes for USA and Europe. It provides an accurate automatic clock adjustment and shows the daily deviation of your HP49G clock. Next Sunday we will switch to daylight saving time in Europe. In the US the switch will be done one week later. Download of the Lib is available on: http://users.belgacom.net/EAA/Heiko/HP49/download.htm Update on Erics Site is in progress. Comments apprechiated! Heiko ==== Heiko Arnemann :: > > Most HP49G clocks are not very accurate. One second a day may not seem > to be much, but will sum up to a minute in two months. Additionally, the > temperature of > the HP49G, and therefore the usage of the calculator, influences the > clock speed. > ClckAdjst is a library, which provides tools for automatic clock > adjustments and much > more. > > ClckAdjst provides fast and easy setting and adjustment of the system > clock. It handles changes of the time zone and automatic daylight saving > time changes for > USA and Europe. It provides an accurate automatic clock adjustment and > shows the daily > deviation of your HP49G clock. I checked the package and saw that basically it works by correcting time with alarm routine. Altought it is little more sophisticated. I am using the kinda similar method currently, but was thinking of somewhat different approach for this task. Wouldn't it be possible to incorporate the time correction routine into System Outer Loop? It would be easy to implement in Java Stack Replacement program, or in MetaKernal. Easier in the first one, because sources are freely availailable. How come there are no sutch solutions yet? Best wishes Robert Tiismus ==== thanks for testing and your comments. > I checked the package and saw that basically it works by correcting time > with alarm routine. The basic idea behind the package is, *not using an alarmroutine. It can be done by those who prefer... The clock need to be adjusted only, when the accurate time is required! The corrective action can be done whenever you want, e.g. directly before showing the time or calculating something with the system time. Say, you travel to NewYork, make the change for the TimeZone, and run the corrective action (AutoADjust AAD it ist called in the package). This will check, if their is daylight saving time in NewYork and makes the setting of the clock, taking into account that some inaccuracy has been counted up, since the last adjustment). > Wouldn't it be possible to incorporate > the time correction routine into System Outer Loop? It would be easy to > implement in Java Stack Replacement program, or in MetaKernal. Easier in > the first one, because sources are freely availailable. How come there > are no sutch solutions yet? This is not my world:-( It would be nice to make AAD a little faster, by using sysRPL or ML. But I am not familar enough, with. Than AAD could be started together with the switch-ON. Heiko ==== Heiko Arnemann :: > The basic idea behind the package is, *not using an alarmroutine. > It can be done by those who prefer... > > The clock need to be adjusted only, when the accurate time is > required! The corrective action can be done whenever you want, > e.g. directly before showing the time or calculating something with > the system time. Say, you travel to NewYork, make the > change for the TimeZone, and run the corrective action > (AutoADjust AAD it ist called in the package). This will > check, if their is daylight saving time in NewYork and makes > the setting of the clock, taking into account that some inaccuracy has > been counted up, since the last adjustment). Yes, I read about this from the documentation. But this method is not universal. I mean that the existing time display routines or other time related programs can not use this method by default. They must be recompiled :( The alarm method is more universal, but it has weaknesses. For example: if set to every night alarm at 4 am. then you have to be aware that your work will be interrupted if you do it too late at evening or too early at morning ;) And between the corrections there is always the moving discrepancy between the real time and the calculator time... >>Wouldn't it be possible to incorporate >>the time correction routine into System Outer Loop? It would be easy to >>implement in Java Stack Replacement program, or in MetaKernal. Easier in >>the first one, because sources are freely availailable. How come there >>are no sutch solutions yet? This was a rethorical question. I was just wondering. By all means it was not directed at you or your software :) Best wishes Robert Tiismus ==== ==== > Is that also working in ALL the models (including 28) the same way? Ok, ok. I was reasonably certain that they worked on the 28 series (I couldn't see any reason why not), but now I've tested them on both a 28C version 1BB and a 28S version 2BB (as far as I know, this was the only version of the 28S), and they do indeed work exactly as intended. Provided, of course, that LASTARG is changed to LAST and -55 DUP SF CF (in two of the programs) is changed to 31 DUP CF SF . Previously tested on a version E 48SX, version R 48GX, and revision 1.19-6 49G, with 2 and -55 entered as 2. and -55. on the 49G, although 2 and -55 also work. I don't see any reason to suspect that they wouldn't work on other ROMs on the 28 series, 48 series, or 49G. I'd almost forgotten how nice the 28S is, very well designed (with the notable exceptions of the lack of any input method other than the keyboard and the battery cover and batteries), although I wouldn't want to go back to programming it on a regular basis. By the way, the 48 series and 49G accept LAST as a synonyn for LASTARG , but it always decompiles as LASTARG . Perhaps that might save someone a few keystrokes when typing in the command. Does anyone know of a built-in shortcut for LASTARG; shorter than the PRG ERROR menu or the catalog, that is? It seems a strange thing to leave out. Of course, the program that returns *nothing* in the event that the last arguments buffer (Is that the correct term? For the lack of a better term, that's what I've been using.) is empty may well be considered buggy by those accustomed to a LSTx key that always returns *something*. > If so, great work, Mr. Bond! I've been wondering whether a variation that leaves all of the contents of the last arguments buffer as they were is possible. I can't think of any way to do it in UserRPL, but perhaps it could be accomplished in SysRPL, if anyone cares to play around with it. -- ==== Try the following; note that it restores whatever was the most recent > object in the last arguments buffer back to the last arguments buffer. You're welcome Bill. I'm glad that someone finds it to be useful. I was mostly just playing around. I did remember that I'd seen a version of LASTX somewhere but hadn't bothered to look it up. I had previously itten a version of LASTX that mostly worked, except that in the somewhat unlikely event that the last arguments buffer were empty, it would return *nothing* to the stack but put the real number 0 in the last arguments buffer. I was never quite satisfied with that, so I decided to take another look. And then a couple of more looks. If you don't mind, which version do you prefer? For myself, I prefer the version that, when the last arguments buffer is empty, puts the real number 0 both on the stack and in the last arguments buffer. -- ==== > > If you don't mind, which version do you prefer? For myself, I prefer the > version that, when the last arguments buffer is empty, puts the real > number 0 both on the stack and in the last arguments buffer. Yes, that's the one I prefer. Actually, I don't use LASTX much, but it sure is useful when you need it. Bill ==== Veli-Pekka Nousiainen schrieb > Who needs [SYMB] anyway??? You are right Veli. It just took me 2 Minutes to find SYMB, because since years I use it for EVAL (which is mentioned above the key :-) For these rare Moments one can switch of usr-mode or I use Keyman , to get longpressed (and doubble-klick) funktionality. ..Heiko ==== I Would NEVER get rid of my 20 year old HP-41C! Never Never Never ==== Is there a way to (or software that will) isolate a variable in a linear inequality on the HP48 (e.g. 2*x-4 > 0 isolate x.). If not, is this something that the HP49 CAS can do? Ollie. ==== The HP 48G ISOL cannot do it, but the HP 49G commands SOLVE or SOLVEVX will give 'X>2' Is there a way to (or software that will) isolate a variable in a linear > inequality on the HP48 (e.g. 2*x-4 > 0 isolate x.). If not, is this something that the HP49 CAS can do? > Ollie. > ==== So Erable or ALG48 won't do it either? > The HP 48G ISOL cannot do it, but the HP 49G commands > SOLVE or SOLVEVX will give 'X>2' Is there a way to (or software that will) isolate a variable in a linear > inequality on the HP48 (e.g. 2*x-4 > 0 isolate x.). If not, is this something that the HP49 CAS can do? > > ==== Hang on... I tried it on my HP49 emulator and got a 'solvx error: non unary operator'. Is there a trick to doing this? > The HP 48G ISOL cannot do it, but the HP 49G commands > SOLVE or SOLVEVX will give 'X>2' Is there a way to (or software that will) isolate a variable in a linear > inequality on the HP48 (e.g. 2*x-4 > 0 isolate x.). If not, is this something that the HP49 CAS can do? > Ollie. > Reply-To: Veli-Pekka Nousiainen ==== Get the latest 1.19-6 ROM > Hang on... I tried it on my HP49 emulator and got a 'solvx error: non unary operator'. > Is there a trick to doing this? > The HP 48G ISOL cannot do it, but the HP 49G commands > SOLVE or SOLVEVX will give 'X>2' Is there a way to (or software that will) isolate a variable in a linear > inequality on the HP48 (e.g. 2*x-4 > 0 isolate x.). If not, is this something that the HP49 CAS can do? ==== the SX and GX communicate fine over the infrared link. And, by the way, they communicate even faster over a wired link, if you get or make the appropriate cable or adapters. > although the 49G has an IR window, there is NOTHING about IR in the > documentation... can you comment on this ? It might look sort of like a window, but it seems to be quite opaque and doesn't have any IR transmitter or receiver behind it. I suppose that they considered it best not to even mention that bad news. > however, the 49G and 48GX communicate well over the wire link... but its slower > than the IR. Slower than the IR? If you've managed to communicate with the 49G by IR, then please let us know how to do it. If you mean slower than communicating between two 48s, well, perhaps if you used binary transfer between the 48s but ASCII with the 49G, or weren't using 9600 bps with the 49G, but otherwise, I find it just a bit difficult to believe, but perhaps possible; the 49G is indeed slower than the 48GX at some things. > Please comment on the Connectivity Kit F1897A........do I or do I not need it ====