A26 Don't say bombs out because this is exactly the opposite. It prevents bombing by trapping the error condition. (Bomb would be a warm start). My HP49 just solves this fine and finds P=1E12. If you already have a variable with some value (perhaps a complex one) thenthis will be used to evaluate P-R^(-N),==== Now that ACO is dead, I think it's appropriate to evaluate somehow whatACO has accomplished. Obviously, I do not intend to write a complete report- just to mention some points and, hopefully, start some constructivediscussion.OK, let's start at the beginning. The following paragraph is an excerpt ofthe 1997 HHUC Conference Report, included in the Nov/Dec 1997 issue ofDatafile, written by Wlodek Mier-Jedrzejowicz:Chris [Wallin, head of the new HP calculator operation] spoke with greatenthusiasm about the past history of HP calculators, the many waves ofinnovation brought to calculators by them, and the occasional less successfulproduct. [...] Calculators have been left to one side lately, and HP Australiawanted to take over, because of Australia's culture of innovation.He mentioned various technical innovations to come from Australia [...] andsuggested that this gave his team in a special advantage in designing a newrange of HP calculators while retaining their high quality. I think that atthis point Chris lost some sympathy among some listeners - he seemed to besuggesting that the successful calculator team at Corvallis had somehow lostthe ability to innovate - whereas we all know that the team has actually beenthe victim of changes and lack of interest by senior management. Nevertheless,his statement that HP will continue to design and produce innovativecalculators earned him a cheer.Which innovations have been made? Aside from the excellent metakernel andthe CAS, I'm afraid that...Flash ROM definitely couldn't be considered state of the art, and thesuccessful HP6S is anything but innovative - Can you believe that underNORMAL light conditions one cannot see the shifted key labels on the bluemodel? The 39G/40G are nice rebuildings of the wonderful HP38G - a trueinnovative model! I think that the 30S doesn't deserve much comment, andabout the new HP10BII, well, fortunately the rubber hasn't been placed onthe keys!As well as Xpander, I know that there were many more projects that werecancelled for different reasons. It would be enlightening to know moreabout them.What went so wrong? Did you know that some members of the Corvallis teamare still working on handhelds ... for TI!!!? I'm sure many of them wouldbe delighted to make the goose fly backwards again.Like Jake and Elwood rejoining the band :-)Just my two cents.Bye.-[blah blah snipped]Unfortunately, you talk with very few element of what has been done. Thefact is, very little information has been released.ACO hasn't been closed due to performance, or lack of profit.But before I go into details, let see about what HP-ACO has achieved overthe past 4 years:I will not talk about consolidating distribution channels, selling HPcalculator in countries where it was not available before, creatingeducation material etc... No, just pure hardware workLet's hope I won't get into trouble for this. But I honnestly don't careanymore.In order of appearance, some of them may have been released in the sameamount of time:1)HP48G+2)Morgan, project cancelled never released3)HP12C, total redesign of the internals and CPU. A huge work, shit load ofmoney4)HP17BII, HP19BII: Internal redisgn (due to component obsolescence)5)Total redisgn of the IR printer due to obsolescence of components6)HP49G7)Firmware Datalogger (financed by HP, hardware produced by HP)8)HP6S9)HP39G, HP40G10)HP10BII11)HP30S12)Xpander: project cancelled, one month before released as big HP bossesdecided that HP would NOT go into education anymore. Xpander was not just acalculator, it was a total new concept of educational tools. Hugedisappointment within ACO.Really nice device, a complete PDA for young people. Supposed to be releasedin two weeks time. Didn't get the chance for it. But the product is almostfinished. Hopefully it will be distributed one dayOther PDA: Carbine, based on the same architecture as Calypso (software,hardware) but it's a very tiny PDA, looks like a mobile phone. It's the bestlooking PDA I've ever seen in my life, and the screen supposed to be usedwas fantastic. Tiny, light, powerful, long battery lifeOther than these projects (all cancelled) ACO investigated how we could gointo wireless, so I worked with many others for the past year on developpinga GPRS/GSM/EDGE plateform, result: the world first embedded single-chip (notpeople were making comment on why I was the author of the port of a CThe project one more time was cancelled and the prototype was working..So the result is? Sure, few has been released. ACO's fault ? Don't think so.In this division, I've seen some of the most talented people in my life.I've worked in various companies, but I had never seen such a level ofcompetencies.ACO was profitable, and in fact, could finance itself just with thecalculators sale (and it's not a small profit, ACO by itself would be amulti-million dollars company).Why did they close down ACO then ? that's a good question.But management the week before said: we had to cut our workforce by 4%. It'smore popular I guess to get rid of a small division, in a country far faraway, a long time ago (oops that's Star Wars), than in the USACO is a small division in HP organisation, and doesn't really fit intoCarly's model where you have to be either first, or close second. Threewords: profit, profit, profit.But don't worry, HP will continue to sell calculators. The one redesignednot so long ago. What about graphicals ? Well, there are stocks.Unfortunately, the Saturn is becoming obsolete this year, and nothing hasbeen done to redisgn it, the bet was on Xpander. And redesigning a customCPU cost over 1/2 million, too much in a market where short term profit isthe key (even if analysis shows that big profit would have been made afterone year only)As Forrest Gump would say: And that's all I have to sayIf now somebody ask me if I'm sad for not working for HP anymore? No, andthe fact is : I'm relieved.The only problem is: it's not HP, HP is a great company with great product.Unfortunately, it's the way the world goes. Pt marketing people at the headof the most brilliant companies and see what happen:Q: What is this division doing?A: the sell the product XQ: Do they make profit ?A: Sure, they own 85% of the marketQ: Are they growing over 16% to make our share-olders happy ?A: Sure they do for at least 2 yearsQ: What is this division doing?A: The do R&D and work on a replacement of product X.Q: Do they make profit ?A: Uh? No, it's a R&D facilitiesQ: Sure... So they don't make profit ?A: noQ: Well, our figures show that we could save $500 millions by reducing ourworkforce by 7000 people. We also need to increase our growth otherwise theshare price will drop. Our CEO also needs to keep flying on his/her privateplane with the private hairdresser. Let's close this departmentA: But what about in the future ? What will we sell?Q: Well, make sure you buy a lot of stock now, sell it next year. You won'thave to work after that, so why care about the future of the company ?terms: customer orientation, which is an euphemism for shareholder orientation, shareholder meaning someone who buys some stock and wants to sell it with profit in a short time, nothing similar to the traditional long-term shareholder who had more or less relation with the company. The CEO of RedHat said it; with the .com madness, many of their shareholders only knew the NASDAQ symbol of the company; they even didn't know the complete name. Anyway, I think many companies going into this nonsense will sooner or later regret. I see that HP wants to go to a market which can offer high short-term profits (PCs) but in which you just have little possibilities of keeping some leadership. HP already abandoned instrumentation, a field in which they were one of the leaders, and with an enormous barrier for any newcomer. Moreover, I don't think there may be a drop in the demand for biomedical instrumentation. However, PCs are more powerful with time. Is there a limit? When will users get as much power as they think they need? A new PC can perform huge tasks, such as encoding a movie in DivX, etc. If the power multiplies by 10, for example, will we find tasks to keep the computer busy so that the user wants to buy another? I think there will be a limit. And if that limit is reached, PC dealers won't be able to keep that growing figures stock market analysts like. Stock-market oriented strategy is not good; look at the UMTS fiasco in Europe. European operators have paid enormous amounts in the English and German auctions for the UMTS licenses, and now they are facing financial trouble. The market is almost saturated; they will *not* find new customers. And, are existing customers willing to pay more for the most advanced technology? I don't think so. Mobile telephony is already expensive, and there are market studies showing the maximum amount of money the average european is willing to pay for telephony, digital television, etc. I guess the people responsible for the bids in the UMTS auctions had those studies in hand. However, they bidded as stupids. Why? Companies quitting the auctions saw they stock prices dropping, and nobody wants to see prices dropping. Something similar is happening in the computer industry. HP seems to want to manufacture PCs and printers, and *only* PCs and printers. PCs are a product in which you have a very small margin to innovate and keep an edge over the competition; the main argument is price, and there are many asian manufacturers selling low-quality PCs, but Brave words, big opera.......;-)Hmmm, excellent? Well, according to the discussions here, with so muchcomplaining and confusing, I would say: The parts of the built-insoftware of the HP49G are very good, but the integration of them (asBhuvanesh correctly says, it should be an integrated environment) in awhole is less thoughtfully implemented.Overall physical quality has declined. The keys of my HP49G startlosing color. Not to speak about the quality of finger power exerciseswhen I (try to) press the keys.Many projects? YOu mean there were more of them? I only know about theXpander. Can you tell us about other projects that were cancelled?Really? For TI? Hey, could it be that the next super calc will comefrom TI? Boy, I start thinking about assimilation of myself by thedark side, which after all might also be not so dark at all, now thatthe Jedis are there.Who are those guys? Perhaps two Jedis of the old kind infiltrating thedark side? ;-)product.headWhat happens then you say? Then we take over. Simple as that. Those HP ====This makes me sad -- and angry. :( IMO there's something fundamentally wrong with the present computer industry. It's know-how and creativity draining out ... ACO's demise is just another example. The big companies have been cutting down expenses on R&D for quite some time now, and those elephant mergers sailing under the flag of globalization only make everything worse. Innovate??? Nah, it's Make Money Fast, on a grand scale.Those marketdroids can only focus on their quaterly reports anyway, they have no idea what innovation means. The last century has been called the century of the scientists and engineers; now we apparently have the century of the managers. Let's see how long they can run the show. ;-)Anyway, JY, let me take the opportunity to thank you for all the good work you have done with ACO. I'm sure that with all your knowledge and ====Hmm, for me it works quite well. Take an equation, simplify it in the EQW, copy the part you need, turn it into a function, make a graph, all with a few keystrokes. If I compare this to a PC - trying to work with both a symbolic software and a numerical package, and a visualization software can be quite a nightmare.I think that many of the problems people have with the 49G are actually due to the lack of decent bundled documentation, in particular, a tutorial on how you actually put this plethora of functions to work for concrete problems. Urroz' books can help there, but they still don't ====comp.sys.hp48:142725 comp.sys.hp.misc:29114 comp.sys.handhelds:80307 misc.forsale.non-computer:108843To each his own... Ben Myers====Where can I find the hp49 version of Statpack?/Jonsite. It'sparticularly thosebugs reportedreceipt, buthand, if any====chip on glass (cog) vitrim tm series (SEIKO)slightly larger than HP49 display 70.7x51.4x2.2 dot 240x1605 volts .225x.225 dot size $55 add backlight = $6.88evaluation kit available to customize application ($199)anything better out there? while back I was looking into potential new calc displays ( lookedat the one you mention ). Did a lot of searching for various LCDmanufacturers but I found that the Seiko/Epson devices were among thefew that had the right combination of resolution + size. Of course,many companies will manufacture custom displays but then you'retalking big $$$. :) Anyway, these might be of some interest to you : http://www.eea.epson.com/EEA/LCD/publicFiles/lcdmodule_e_0109.pdf Beware of line break - huge link ). Digikey also has some displays but I think the only graphic typedisplays ( made by Optrex ) are too large: http://info.digikey.com/T013/V5/SectO.pdf ( page 13 )and don't forget Yahoo :) :====thank you for your input, i'll look them rigth away.====I'm looking for Sylvain Gamot's FXMIT and MIDIPLAYER packages for the gone and I couldn't find the software on hpcalc.org. Does anyone know where his page has gone, or where I can find that software? Or, better yet: Sylvain Gamot, if you read this it would be nice if you could drop ====Bad style to follow up on one's own post, but anyway. Looks like these programs never worked, I should've done a google groups search earlier. I would see the paper clip reset more like an On-C. I have never lost You can actually fix this problem without clearing your memory(probably too late now ;). Here is a program by Robert Tiismus thatdoes the trick : http://www.hpcalc.org/hp48/utils/memory/p0fix.zip .==== http://www.hpcc.org/hp39_40/lp.zipUseful links: http://www.hpcc.org/links.html#hp3940I have been trying to unsuccessfully display a string on the screen inML. Firstly I tried $5x7. No go. So I have tried usingMINI_DISP. The following Code causes the screen to fill with rubbishthen a warm boot. Can anybody tell me why?IT is getting quite annoying that it is not working.ASSEMBLE NIBASC HPHP49-CRPL::$ Test StringCODE GOSBVL =SAVPTR * SAVE RPL Pointers A=DAT1 A * Pointer to String D1=A * Point D1 to Address D1=D1+ 5 * Skip Prologue C=DAT1 A * Read Length C=C-CON A,5 * Remove Length field from size CSRB A * Divide by two for number of char D1=D1+ 5 * Skip length field GOVLNG =MINI_DISP * Display String GOSBVL =GETPTRLOOP * Reset RPL PointersENDCODE( now freeze the screen, else you won't see anything )SetDAsTemp;Below is the documentation for the use of MINI_DISP. As you can seemy program sets the followingD1 = first char of string (after skipping prolog and lengthD0 = Top left hand corner of screenC = Number of characters determined directly from the length fieldST11 should be set to 0 as default% MINI_DISP: Assembly function% purpose: Display a text on a 131*x screen in mini font% at a 4 pixels alligned position% Inputs:% D0: point on the screen where you want the chrs to% be displayed% D1:point on the first chr of the string% Ca: Number of chr to display% ST11: 0/1 normal mode/video inverce mode====Can you make sure your entries are correctly resolved ?Also, change the GOVLNG MINI_DISP for a GOSBVL MINI_DISP, MINI_DISP is a====you should use GOSBVL =MINI_DISPWhy would you even w-- oh. You have a HP48G?define hp code. define hp command. There are PC(and on-calc, but nevermind) development tools for MLand SysRPL -- and you can write UserRPL with any asciieditor (but why? *that* you can do fine on the calc).For UserRPL just get the HP49G debugger and restrict yourselfto HP48 commands. It has a debugger in the PRG RUN menu thatyou can use. For ML and SysRPL... I dunno, but if you'reseriously wanting to deal with those you ought to at leastget a GX anyway.OK, forget ML (and probably also SysRPL). define 'ROM' flash====any feedback on TINI (as platform)?(from agilent calculators thread...)they have new and more powerfull TINIs comming out====On Sat, 10 Nov 2001 10:57:03 -0600, Jonathan Busbybyte serial internally I should say. It still has a 4-bit external==== Like Raymond said, there is no such thing as a permanent solution.Eventually, no matter what processor you use, the gulf between itsperformance and the current status quo will eventually grow too wide.The key is in making sure it takes that a relatively long time tohappen. For example, the 48GX was introduced in 1994 and at the timeit was about 25 times slower (In terms of clock speed alone) than thehigh end PC processor (Pentium 100) and that was acceptable back then.Now it's about 500 (!) times slower than the current high end ( interms of Mhz - long live Athlon! ;) - the Pentium 4. It's been 7years and now the Saturn speed is starting to slip past the borderlineof acceptability ( some might say that happened long ago ;). Aredesigned saturn on an FPGA might give us a speed increase of 25-50x.So, assuming Moore's Law is correct and processor speed doubles every18 months then we'll be in this situation again in say, 7 years :D -coincidence? ;). ( assuming of course that by then we don't havesingle chip massive parallelism which is a distinct possibility ;)That's enough time to satisfy me :). And I guarantee that if HP + usdesigned a backward compatible calc that came anywhere near to thatspeed increase it would pacify all of our whining and ranting (==== Actually, It seems I was looking at the 68000 cycle times when itsexternal data bus is operating in 8-bit mode ;). 16-bit mode levelsthe playing field somewhat. Keep in mind though, that the Saturn isable to operate on any nibble aligned field in a register that startsat the least significant bit ( and some other fields. The cycle timeis roughly proportional to the field's nibble length. There areexceptions for some fields ). The 68000 can only operate on the fieldslisted below which necessitates multiple instructions for other fields( making it slower than the Saturn ). Also, note that I've left out64-bit operations on the Saturn. For example, it takes 11 cycles to doa 64-bit add.Here is a sampling...Operation: Saturn Cycles: MC68000 Cycles:binary ADD8-bit 4.0 4.016-bit 5.0 4.032-bit 7.0 8.0bitwise OR:8-bit 5.75 4.0 16-bit 6.75 4.032-bit 8.75 8.0register to register move:8-bit 4.0 4.016-bit 5.0 4.032-bit 7.0 4.0shift right ( 1 bit ):8-bit 7.0 8.016-bit 8.0 8.032-bit 10.0 10.0HP48GX cycle counts by Mika Heiskanen and Dan Kirkland : http://www.hpcalc.org/hp48/docs/programming/cycles.zipM68000 8-16-32-Bit Microprocessors User's Manual by MotorolaCorporation : http://e-www.motorola.com/brdata/PDFDB/docs/MC68000UM.pdfNOTES: (1) All cycle counts use general purpose registers only. (2) All Saturn cycle counts assume the opcode is on an even address. On the calculator there will be a higher cycle count for odd addresses but this is not due to the Saturn but to the memory controllers in the Yorke ASIC. (3) Saturn only supports single bit or four bit shifts. The MC68000 does not have this limitation. I apologize for any inaccuracies in this post.Hope this helps...====On Sat, 10 Nov 2001 10:57:03 -0600, Jonathan BusbyOops. This is if the 68000's external data bus is operating in 8-bitmode ;). For 16-bit mode the cycle counts would be 4 and 8====On Sat, 10 Nov 2001 03:31:56 +0100, Raymond HellsternA=A+1 A measured with CTIM of the Hack library on my 48GX-Rtakes around 5 cycles. Compare this with the 68000 which takes 8cycles to do about the same thing ( via an ADDQ.W #1, D0 say ).Although on the 68000 you're limited to 8, 16, or 32 bits and the32-bit case takes even longer ( 12 cycles ). If you increased theinternal data bus of the Saturn to 16 bits then this would probably becut in half ( as all evidence points toward the Yorke Saturn beingbyte serial ).-- before replying. I was just assessing the hypothetical situation in which you wouldit's already been ported to ARM :).Ok. I thought you were talking about *converting* all of the sourcesfrom sys-rpl/ML to C ;).I agree. Anyway, the minor changes I wanted to make were to theSaturn *architecture* not the actual Saturn as it is now. Who knows ifHP was even using an HDL when they designed it. The current Saturn isa leveraged design from the Clarke, which was a leveraged design fromthe Lewis etc. Were HDL's even in widespread use in the mid to late80's ? Even if they were I doubt whatever high level representationwas used is amenable to the type of changes I want to make as theyprobably weren't thinking about the Saturn 15 years into the future;).I wasn't saying ARM assembly is difficult ( I'm somewhat familiar withit ). I mean I used to program in x86 assembly and that is asconvoluted as you can get ;). Anyhow, I was just trying to say thatthe ARM architecture is a whole lot more complex ( and consequentlyfeature rich which makes it easier to deal with at the assembly levelin many respects ) than the Saturn's.I meant it wouldn't be a walk in the park if we were implementing aCPU with ARM like features. :)A little self bashing is good for the health ;).Hehe. I wondered why that gate count was so high. Silly me ;).Ok. No one's planning on reimplementing the XScale on an FPGA ;). ButI think the gate counts for the processors I mentioned are goodestimates since they contain very little if any on board ram (besidesthe registers). At any rate, they certainly make putting a Saturnclone on an FPGA sound feasible.Well I don't have the prices right now for the XC2S30 but for the100000 gate XC2S100 it costs around 10 USD in large quantities. Thelower end ones are bound to cost significantly less.I know. Like previously I was just imagining the hypotheticalsituation in which you did have to write one from scratch. At anyrate, even if you wanted to create a new compiled language a lot ofthe most difficult parts would have already been taken care of. Thisis because with GCC you can redo the front end (lexical analyzer,parser etc. which can be written with the help of lex and yacc so mostof the grunt work is taken care of in those stages also ) and reusethe other more complicated parts ( optimization stages, codegeneration etc. ).Ok. But what I want to do wouldn't have anywhere near that much ram.It seems you want to design an engineering powerhouse or a high endworkstation calculator. All I want is what we have now but better andfaster in terms of hardware and software. Perhaps they'd be differentenough to peacefully coexist within the market? ;) ====I apparently read this too fast ;). Yes, it would be a matter of recompiling BUT... unless those entrypoints are emulated to the strictest degree it will involve majorprogram redesign. Emulating user-rpl is the least difficult althoughI'm not implying it would be easy either. Now, when you enter sys-rplland you get a lot closer to the innards of the OS and really itnecessitates emulating (or in this case reimplementing - even worse )a large portion of the higher level OS routines such as displaymanagement, libraries, directories and all the myriad of built inprogram helper routines. Obviously this is much more involved than theuser-rpl case. For the ML case, well, really the only way you canprovide true compatibility is to emulate the whole machine whichreally makes the former attempts at emulating the OS at a higher levelpointless. A lot of ML programs delve into the depths of the systemand if its behavior changed in the slightest bit it would be crashtime :). Even an address change in ML can cause problems at the sourcelevel so as to make a straight recompile impossible. I mean, portingstuff from the 48 to 49 is hard enough not to mention a system thatonly incompletely emulates these :).Well, considering that to have full compatibility you'd have toreimplement the entire OS ( and emulate the hardware ) so as toduplicate its behavior to the satisfaction of most programs, I'd have====On Sun, 11 Nov 2001 00:39:56 -0600, Jonathan Busby...Another complication (among many) would be that if you're staticallycompiling assembly programs into some other form, what is going tobe done when the program decides to modify itself - directly orindirectly (as many do)? Clearly this is a major problem. Plain old====Obviously I'm having a little trouble with my usenet provider. ;) (All the messages I sent in the last 3 days went to the bit bucket )--====Many has. We need to optimize any existing OS for the XScale, or we go theother way, implementing an OS kernel ourselves.No, that would be a bit tideous :-)I'm not sure of that either - even though it was NEC ;-)Yeah :-)True.You think? ;-)LOL :-)No? ;-)Good estimates - I agree.That's pretty cheap too - we need to get started :-)====But they are. We start at ML (fairly easy), then we compile each SysRPLmnemonic at ML level! Then we go by all the USerRPL words, this time atSysRPL level and ML where necessary. Pretty easy.about.We do not have to emulate the hardware anymore than taking it intoWell,I took the cycle count information from SASM.DOC...So maybe CTIM is not exact, or not adjusted?Here's an excerpt:A=A+1 fs - Increment A-fs = A opcode: E4 cycles: 7====On Mon, 12 Nov 2001 17:30:03 +0100, Raymond HellsternSASM.DOC is for the SX Saturn. Dave Arnett said it was redesigned in====Ok. But I thought this was supposed to be a clean roomimplementation of RPL. If not, you'll have to license the source codefrom HP ( assuming it isn't GPL'd ). Unfortunately, it isn't that easy. Many ML programs access thehardware directly ( especially the display and associated registers ).Whatever you want to call it, it's still emulation. Although in thiscase it would be in a different form than say EMU48 since the assemblylanguage would be pre-translated into native code. But even with fullemulation, there are problems such as self modification. Really whatyou end up doing is implementing a full blown emulator like EMU48even if it doesn't look like it superficially.====I don't think we'll have time for that at first.I assume it is. If not, we're FUBAR anyway.I see every ML instruction look in a global jump table where the registers,I just downloaded the 39/40 ROM in my 49G, and I am wondering which program====Tried it. Seems the 39G program is the one to use.--Thierry Morissette====dgeve:====Lo people !but a still have a newbie question .....Can anyone tell me how to do text formatting on hp49====X==== I think calculators can be a very valuable marketing asset for Agilent. HP calculators have been the first contact with HP for engineering students, and I guess (this is only speculation) that it has had a strong influence on the preeminence of HP workstations (compared to Sun) in engineering ====Not that I abandoned hope, but with all critisism against calculatorsthere are 2 things possible:1) We wont see an HP calculator again.2) HP is aleady working at some future model, which is going to be theI love RPN and quality HP calculators. I've never owned a TI thatlasted more than a year before it went from an expensive toy to acheap piece of shitty plastic with the worlds poorest keypads (notcounting the HP-31/32/33/34 switches which were on a par with TI's). I've been an HP user since 1977.The raw computing power in a Pocket PC (even my $199.00 Jornada 525)so overwhelms the processor speed and memory capability of an HP-48GXor an HP-49G that I don't see a real need to have a dedicated piece ofhardware called a calculator in the future, though for sentimentalreasons I'll continue to keep a 48GX at home, one in the office andback them up with my HP-15C, HP-42S, two HP-28Ss and when the enemy isfinally at the gate and I'm too old to read the LCD screens, my HP-67with its non-continuous memory and its bright little LEDs. After that,I'll fall back to my Post Versalog and my K&E Log Log Decitrig. :-)I expect that it will be a lot easier to have a killer calculatorliving next to WORD and EXCELL in a fast Pocket PC than in any other====(snip)Other than disbanding ACO, what have they done? (snip)The ROM's future is uncertain right now. JYA has mentioned that he willtry to see it GPL'd, and a lot of people have supported this with theirvoices here if nothing else. If the ROM goes GPL, it is basically freedfor public use, modification, and redistribution; so a good hacker couldfix a few bugs, add a few features, optimize a few routines, and releasethe hip new 20-x-jrf7Dec2001 ROM, and communicate his improvements. Youcould have patch sets, and development tools, and modular ROMdistributions, and groups of people working on different parts of the ROM,and all *kinds* of such neatness. We'll have, basically, our own operatingsystem to improve and play with on hardware we've bought with our peers.Anyway, I think you overestimate the place of the ROM; the 49G ispretty extendable any way you look at it, so it's not as if youare *doomed* if the ROM is static and opaque. Case in point: Inotice that nobody has bothered to put a periodic table in ROM, yet!(and there's no graphical memory viewer, but just *look* at how manyof those things people've made!) If you're a programmer, all existingcode is a jumping point toward future improvement and developement andusefulness -- no matter what happens. Sure, in the year 2043 you mayhow to crack his library, and you can add the feature you always wantedto it! (This is a joke, of course; by 2043 we will have wristcomps.)This is certainly a danger, and as time passes it is a certainty.When those braincomps and wristcomps come, my 49G use will certainlyplummet. The community will shrink too -- and, relatively, it'salready shrinking, been shrinking, and will only shrink faster; TI issucking up more and more people and few will care to hunt for othercalculators when they A) do not know enough to know what they lack, andB) are satisfied with this ignorance. And as you've seen and as youcan see in yourself, a lot of people seem to be taking ACO's departureas a deathtoll for the 49G; some of these will leave the community anduse other calculators. More may never come, when the word is spreadabout how the 49G is 'unsupported'.So, if your concern is with the 49G community, I hope I've properlyfed your fears. Hey, just by having a calculator you are way in theminority! And do you have a computer? Whew! Better not instigateFortunately, you have either practical or ... nonpractical use ofyour 48GX, and have some idea of what the 49G will mean to you.If a TI-92 or somesuch will fit your needs and your personality, andyou think the benefits of support and community outweigh the====If all we want to do is add and subtract, we don't even need the raw processingpower of the HP48!Pocket PCs have their place. They have a nice front end for a majority ofusers. They can do more things out of the box. But that is also theirlimitation. In order to be as user friendly as possible, developers must try tothink of everything everyone might want, so that each person can use the fewthings they need personally. The rest of the features are unused, wastingstorage space, memory and sometimes processing power. It is this overhead that makes the Jornada 548 barely able to emulate the 4MhzSaturn processor at normal HP48 speed. Perhaps rather than calling it a calculator, we should start referring to it asa personal graphics device or something like that. A device with a basicfront end for the average user, with access to a powerful lower level languageand also directly to the machine instruction set, enables users of varyingcapablilities to take full advantage of the units abilities.As to whether or not such a device is necessary, that will depend on what itCAN do, not on what it DOES do. And how impressively it does it. If youdescribed PONG to someone 25 years ago, no one would have bought it. (I sawthis great game. It's like tennis, only you spin a knob to hit the ball backand forth. ) It was only when you played it that you had to have one :)==== Personally I do not want anything else that runs using anyMicrosoft software and it's bloated code. I do not need anothercomputer that crashes. I have two already and that is enough. The reason we need faster processors and larger hard drives andmore memory is that software developers have no incentive to designefficient code. If you read the latest issue of PC Magazine; faster processorshave not really gotten us a far as they could, because of this factor.At least the designers of the 48 and 49 had the memory limitations toforce them to become more efficient in their coding. I am afraid the lure of faster processor and more memory in anew HP calculator will lead to the same problems, especially if HPuses a Windows derived OS as a shortcut to get the calculator tomarket faster and does not write its own independent and optimized OS. I do not need a PDA since I have no need for spread sheets onthe run and I do not go to meetings.====XOnly the Saturn CPU chip Yorke (along with the display pixel count)It's the software that counts.A new ROM 1.19-7 will appearANDin the meanwhile this group will find an answerSpending $18.4 billion to purchase Compaq when HP is already facingshrinking revenues for one. It actually makes me wonder if HP will gothe way of DEC... (marginalized until someone who doesn't understandthe value buys them out)That would be an excellent development of course! That is a great wayto continue the evolution of the platform and to make the HP49 staycompetetive. However, if my read of this is correct (that thehardware will cease to be manufactured at some point?) then it doesn'tdo a lot of good. People's HP49's will eventually break, wear out,get lost, etc. and without the ability to replace them, they'll haveto revert to something like TI.Of course (I've been a Debian developer for years). There is aActually, I was not at all concerned about the ROM. I was concernedabout the hardware platform. Seeing things like Java on my HP48GX Iknow how expandable the system is without even modifying the ROM.I am quite sure that the HP49G will more adequately fit my needs. Theconcern is this: if indeed the HP49G has now become a dead end as faras HP is concerned (this much even seems uncertain), everyone is goingto be forced over to someone else sooner or later. How soon? I don't====Just to remind you that HP aren't spending any money at all on buying Compaq since it is an all share offer. The money comes from issuing new share-capital which means that all existing shareholders see their shares diluted in value by a fraction that is roughly equal to $18.4bn divided by the total number of HP shares currently issued.-- Bruce HorrocksHampshireEngland====The theory of such purchases is that the shares of the buyer are notdiluted. After the purchase, there are more shares in the combinedcompany, but the post-acquisition per-share value is comparable to the====Sorry if you read this post for the second time. I have to repost itbecause of an unidentified problem :-(Le 2 Nov 2001 02:41:09 -0800, gtsiros@yahoo.com (George Tsiros) a216crit :gx. It is a great idea, but you are not the only one who have thisproject. There is 3 french projects like this one : - ShellOS was a multitasking OS for the HP48Gx. There is a demoavailable (I think you can find it on hpcalc.org). Unfortunately itseems that the authors can't work on this project anymore, and thatthay have lost the sources :-( - SOS (Saturn Operating System) was a project started by myself. Itis a multitasking OS. It was designed for the HP49G but could easilybe adapted to the HP48. Actually there is only a little part of thekernel done (and sorry, the documentation is in french, but it couldbe translated), so there is no demo available. You can download thesources here : http://clement.pillias.free.fr/hp/SOS.zip or here : http://clement.pillias.free.fr/hp/SOS.tgz Actualy I don't work onthis project since I want to finish Doom first. - WinSOS is a project similar to SOS, started by Yves Brisseaud (theone who did the MASD mode for emacs recently). The project alsoinclude a graphical user interface made by someone else (sorry I don'tremember his name :-( ). This project is not more advanced than SOS,====Has anyone else experienced difficulty in obtaining these books:Science and Engineering Mathematics on the HP49G Vol1 and Vol2?According to the web site (BookSurge), all books ship within 48 hours. I ordered these books over three weeks ago. I can't get much of aresponse to my inquiries.==== I ordered mine on the thirtieth and got a message on Friday, after aninquiry, from a fellow named Rick Jonesrickjones@greatunpublished.com, who said it was to be shipped on==== with all the commotion going on about creating the next best superpowerful calculator I can't refrain to at least speak my mind. We haveto acknowledge that the entire industry is shifting and for a reason.I'm not saying that it is a good or bad decision and I'm happy to seeinitiative and creative thinking to look for alternatives from usersthat do not want to settle with the corporate decision making process.What I do not understand is why everyone thinks that the power has tocome from the unit at hand. I read many debates about which CPU touse, at how many Hz it should run and different compromises toconserve power. There is a lot computing power around us and we don'tuse it. Nevertheless, we want to have the most powerful numbercrunching machine close to us. For the most part I like to agree, I'ma performance freak myself. However, this paradigm is NOT practical.At least not now that communication technologies abound. I believethat the next best calculator (or any portable computing device forthat matter) should take advantage of that and run as a dummyterminal.essence, everywhere I have an internet device I can access the powerof a robust tool that is sitting on my home machine doing nothing mostof the day on my idle P3 1Ghz w/ 256mb RAM. The question is, If Ialready have that, why do I need to worry about how much power I cansqueeze from a 4 Mhz Saturn CPU, StrongARM or any othermicrocontroller? They can not compete. Therefore, we should weworrying about how we can harness that power and make it accessible.PDA’s with internet access are great for displaying information,they have nice big LCD’s (plots converted to gif’s,formulas, etc on a browser). The port of the wrapper could easily bedone for Maple, MathCAD or any other of your favorite applications. Irealize that you still need to have a powerful math application, anexpensive PDA with web access, have your computer set-up as a webserver always connected via cable-modem or other connections. But thatis why the industry is shifting, in the near future this will not bean issue, just like cell phones are today (that’s another idea,how about accessing MAPLE with your phone). This is just a prototypeand in the future could serve hundreds of users (license issues mayhave to be resolved). Business could offer the service and chargesmall fees for its use, via a monthly payment or a one timesubscription, so user don’t have to worry about having thelatest most powerful calculator gadget on the block using valuablereal state on your pocket. Anyone with a small HTML capable devicewith a connection will enjoy the same capabilities, no matter if youpaid $600 for your color LCD PDA or the simplest most featurelessinternet capable organizer.This is just for consideration, it has its limitations and problemsbut I believe it is a legitimate alternative that more people shouldstart looking into. The corollary of all this is that applicationdevelopment should focus on robust OO languages available for PCplatforms (like the C++ CAS being developed by B. Parisse for example)under GPL licenses and then create ways to access its power withOf course, this alternative would be out of the question for examslike the EIT, PE, GRE, SAT, etc, but this is more for everydaywork/education on mathematics, engineering, physics, etc. You would====Sorry, I totally disagree. I am an engineer and allthough building sitesare getting pretty sofisticated I doubt there'll be an Internet Connectionon the 10th floor when the concrete hasen't even been poured.My meetings also rarely have computers or IP connections though thedevelopers are usually flashing the latest PDA's and phones. Actually Ihave a little snigger when the money men hold up the meetings for 5 minutesto enter the next meeting time in their calander. Paper is still best forsomethings.As for wireless technology a lot of remote sites don't even have MobileWhen I'm out of the office 90% of the time all I want to do are simple backof envelope type calcs so something small with instant On is a must. Idon't take my laptop to meetings or site since it's bulky, fragile and toslow to boot up. Anything that needs a laptop is probably to complex to be====I've crossposted from alt.lang.basic for the writer.Please reply to me jwwed@supanet.com-Original post-how do I write an HP48 emulator in dosHow 2 do it?PS: i've not crossposted to comp.sys.hp(?)because i don't have access toit.--Week ago I bought myself HP49G. Well, I can't say that I'm too happy to read====In both calculators use ASCII transferIn the 49G: [MODE] [F3] [DOWN-ARROW] [DOWN-ARROW]Check the Approx Mode by pressing [F3][F6] [F6]Now your 49G is in the 48G Compatibility ModeI hope this helped!Veli-PekkaPS: you're not related to O. BLaden, are you? ;-)====Ok. I have tried to use this method to display characters on thescreen. But it does not seem to work. Anyone know what is wrong withthe following code?ASSEMBLE NIBASC HPHP49-CRPL::CODE GOSBVL =SAVPTR * Save RPL pointers GOSUB ASCII NIBHEX 6464ASCII C=RSTK D1=C LC(5) 00022 D=C A====Here is thedocumentation for the entry $5x7: http*A Name: $5x7 - Write 5x7 pxl data for string at Ith char in Row**** Category: DSPUTL**** Abstract: Writes 5x7 bit patterns for given character string** starting at specified address (S) offset by a character** position index (I). Data for subsequent pixel rows of** each character is offset by (W) nibs from the preceding. D0:S (Start of row)** B[A]: I (Chr Pos'n (0 - ..))** C[A]: N (String Length)** D[A]: W (Row Width in nibs)**** Exit: 5x7 Bit Pattern for string written.** XM: Font Table Flag (0=5x7, 1=5x9)** CC iff next chr is at even location.**** Alters: CPU - A[A],B[A],C[S,A]; D0,D1; P, CARRY, SB,XM** RAM - Destination**** Calls: PtBits(0), sub8/10(0), Set8/10(0)**** Stack Levels: 4 (RSTK=C GOSUB go?Covered CON(5) =cPt2Bits)**** Notes:** (1). The 5x7 font for Chr Pixels is at =PIX5X7, 16 bytes/chr** The 5x9 font for Chr Pixels is at =PIX5X9, 20 bytes/chr** --** (2). Generally, the starting address S is the start of** pixel data for a row of the display. I is an** option base 0 index specifying where the pixel data for** the 1st character will start in the given row. The** address X where the 1st character will start is:**** X := S + 1.5*I (each char takes 6 bits = 1.5 nibs)**** The width W is the number of nibs that must be added to** X to obtain the address where the 2nd row of pixel data** for the 1st char should be written.** --** (3). Applications - Writing string pixel data:**** a) Into a grob.** b) Directly into Display Refresh RAM**** Implied by (a) is substring replacement operations.** ie; ability to replace a specified number of characters** in a grob with others.** --** D0:S (Start of row)** B[A]: I (Chr Pos'n (0 - ..))** C[A]: N (String Length)** D[A]: W (Row Width in nibs)As you can see, your WIDTH is incorrectly set, it should be zero if you wantto display at the beginning of the screen (cf the documentation====This is what I'd do (tested, but not in Jazz syntax):ASSEMBLE NIBASC HPHP49-CRPL::CODE GOSBVL =SAVPTR GOSUB ascii NIBHEX 6464 * NIBASC FF should also workascii C=RSTK D1=C LC(5) #22 * 22h = 34d nibbles D=C A B=0 A LC(5) #2 * FF is only two characters GOSBVL =$5x7 GOVLNG =GETPTRLOOPENDCODE( now freeze the screen, else you won't see anything )SetDAsTemp;Thomas-- ====In case this documentation does not exist for the public in such athorough form, is there a chance we could get a copy of it? This wouldbe wondereful!Rgds,-Al-- -Al ArduengoIf at first you don't succeed, redefine success. -- snip-- -Al ArduengoSometimes a cigar is just a cigar. -- Sigmund Freud -- Yep, you are right about speed. I still wonder though, what speedcould be reached if the hardware would be a little more modern.Hmmm, and is it as easy to use? I mean is it: buy it, get it out ofthe box and use it? I have the (perhaps wrong) impression thatPocketPC and GPL software are not as easy to get them running. Or do Ilive in the past (again)?Now, *that* is a good idea! I hope that all people from thescientific/technology world can stand together, to give the companiesreason for making a math-PDA at last. :-)Well, if everything is pre-installed and works before delivery then itwould be fine. But it would be not so good to expect from the user tobe a PC expert to install and configure everything.Yes, yes, yes. :-)I meant something that doesn't need a standard OS (windows-yacks!!!) for other software to run. Something that has its own(easy) OS and doesn't need 100 configuration files, only to tell youthen that program A can't run because file B is not there. (While yousee that file B is there ;-) )What is Maxima? Math-software?Yes, and a very strong touch, it seems. Do you think that the Jediswill return?What do you think it would be like? I guess it would be more in thedirection of the question, is A.I a living thing or not? If it knows====I've been asking lots of questions here lately, mainly because I'm stillin the process of configuring my Calc to suit my needs as an engineeringstudent.My intention is to post most of my config here, and maybe get some newideas and feedback from you! (Before commenting the names of my dirs andvariables, keep in mind that it's just names that fits my taste... I'msure everybody would come up with something different.)OK, about my Calc: It's a HP48GX with the newest Rom and a 512 Kb memorycard. Hmm, anything else? I don't think so, let's get on with it!- INSTALLED LIBRARIES -Port0: (68K Free)257: UFL (full version)905: QPI1081: GxTools 5.11696: AllMemPort2: (104K Free)769: Tetris (yeah!)807: TicTacToe1111: Sokoban1213: LemmingsPort3: (56K Free)744: SymVector909: Alg48 4.2911: SpecFun (Alg48)913: InteGr (Alg48)1494: NeoPoly 6.5Port4: (57K Free)788: Erable v. 3.117789: Arit (Erable)790: Geom (Erable)791: Prep (Erable)792: LinAlg (Erable)Port5: (123K Free)System Backup- HOME DIRECTORY STRUCTURE -First, I have 6 subdirectories for CST-menu: Numeric, Games, Utilities, Applications, Computer, ConversionNumeric is equivalent to the Calc's MTH menu, just more advanced, and with the commands I use (more about that later).Games is for entertainment! :-)Utilities is for extensions to the calc, not found in the MTH menu (e.g. Limits, Diff. Equations, Integrals etc.).Applications is for specific programs (none installed yet).Computer contains user units from bit to Tbit and byte to Tbyte.Conversion contain radian, binary and HMS conversion commands.Next is 4 subdirs for customization: Formulas, Functions, Text, LibrariesFormulas is for custom equations from textbooks, which can be solved with the equation solver.Functions is where I store my custom programs (mainly vector/complex commands, and my RD10 program).Text is for storing small messages and ToDo notes - nice and handy.Libraries is primarily for storage of downloaded/uninstalled libraries but currently I only have a backup copy of Keys and Flags there.Then there's 2 variables EQ and VX, and the user unitsbyte/bit-Tbyte/bit.And of course CST-menu and parameter files.- USER KEYS -OK, now it gets interesting.The very first custom element is, that the keys A-F haveshift-assignments.This is virtually 12 free user keys, as I very rarely use shifted menus(and if I must, I can turn off user).For my basic config, these 6 have the letters A-F left-assigned, andX,Y,Z,N,T and my function dir ({HOME Functions}) right-assigned.Next up, alpha-right H and J is assigned to -OO and +OO (that'sinfinity).Then right-[STO] is STEQ and left-[<-] is AllMem; right-[EVAL] is QPIand [EVAL] is my RD10 function, which converts to numeric, and roundsthe answer to 10 digits (QPI can still convert correctly to symbolic).For alpha keyboard, right-V, right-W and right-X is assigned to theirnormal functions (root, y^x and 1/x), so you can use them while in alphamode.The [ENTER] key has DUP assigned to the left and EVAL assigned to theright (I rarely use Equation/Matrix), these commands are more handy forprogramming purposes...Finally, right-[DEL] has array to list conversion, and right-[BS] hasarray (2D) to complex conversion.In addition, the 9 number keys have some key bindings in user mode, forinstance 6 (units) have unit factorization (left) and SI-unit conversion(right). In some cases I have to turn off the user keys if I want toenter the standard applications, but I prefer command line operationsall the way.- CUSTOM MENU -Finally, there's the custom menu. Apart from the basic commandsmentioned in the dir section, there are some more submenus in theNumeric and Utility parts.The Numeric menu contains 6 submenus so far: Vector, Matrix, Polynomial, Algebra, Real, ComplexThe first 4 are special, since when you enter these menus, part of thekeyboard is also remapped.Vector and Matrix has commands from Alg48 assigned to the +, -, *, / and1/x buttons. Besides, the right-[ENTER] is now the Erable command EXPANDinstead of EVAL.Polynomial has commands from NeoPoly assigned to the equivalent buttons(y^x instead of 1/x), and Algebra returns the keyboard to normal.The Utility menu contains 4 submenus, namely: Simplify, Limits, Integration, Diff. Equations.Here I've put different Erable commands for these subjects... Forinstance, my integration menu is:- FINAL REMARKS -For the custom menu, I've made GROB's to represent:1: The 4 first labels in the Numeric menu, because programs areexecuted.2: The first label in every dir-menu, so that you can tell CST-menufrom VAR's (the first tab is slighty longer).I would like to enhance my configuration even more, but I still need todo some of the menus... I've also thought about installing EQSTK, it'smy impression it will work well with this config.But please, if you have any comments, do not hesitate to respond, I'mXYeah !!!X====The specs that have been suggested here for a new, much-improved 'better-than-HP' calculator included things like a nice low-power ARM processor, lots of memory, smaler physical size ... perhaps we'd all end up with something like this: http://www.theregister.co.uk/content/54/22699.html(No, this is not intended as a serious suggestion for choosing a base for such a product - but perhaps lessons can be learned, ideas borrowed, from this device.)====Hmm, I haven't seen this before, but I can comment a bit on similarities: Pogo Virtual Platform ;-)CPU: 75 MHz ARM 1 GHz ARMRAM 16 MB 96-512 MBInterface: Stylus KeyboardMP3: Yes YesScreen;Resol.: 320x240 320x240Colour: Yes No (grayscale)XOK! For the third generation that is!1) Make Saturn XP = X=10=#Ah timers the P=Power eg. new process & lower voltage2) Make Jupiter, a HW enhanced Saturn XP====X====X====X====I think that full Sys-RPL support should exist, or do plan onrewriting a whole mathematical software suite for the new system?goingCASFirst Emulated, later native. We can keep the emulator.1) Saturn XP = Saturn in, say 0.25 micron process & lower voltage====but it's already available in the debug2 program...new(andit150isbit====As if I need another Hp calculator, but the BullS*** we get from theupper management at Hp is horrendous. At least we as a group can tellwhen they lie to us. Just watch for their lips to move.The following answer is pat and sounds good, but now for the reality. Look around. All the major outlet stores are clearing out inventoriesof their Hp stock of calculators (aside from the new junk ie 10BII) That is all I find if I find Hp. I cannot find Hp32s, 20s, or 48G'sanymore. I know that a few stores still carry old stock, but Target,K-mart, Walmart, Officemax and Staples only carry the 10Bii. Therefore the message below doesn't ring all that true to me.Therefore, my query, aside from a few vendors on the WEB, where do yougo to buy Hp scientific calculators. If there is no major outlet, themessage above is truly BS of the first order. And since I suspectedas such before I posted, I can't/won't believe much from Hp uppermanagement until refuted and proved otherwise. I await such proof,and will gladly humble myself and apologize for my skeptism. Untilthen however, I say:Carly Foirna and her lacky sidekick Iain Morris are two Jerk CEO'sthat have contributed to Hp's withdrawal of a Market that Hp used toRULE. Poor marketing and follow the leader (Ti) mentality have put Hpin the backseat on a downhill skid that probably no one can stop.No, I don't have the answers either, but I feel that Hp's abandonmentof this area will cost the company prestige and market aligience inthe future.My first Hp product was an Hp15c that is still with me to this day (Iused and abused this calculator for nearly 10 years, prior to that Iwent through a Ti every year). This calculator introduced me to a==== So they can keep their highly inflated and undeserved salaries andput people like JYA and the other members of the ACO out of work.====/(&%#!ó&%I took for real...PS: You ASSUMEable rref-er to:====48's successor. Please, no hurries!By the way, why has the new HP-12C version (that with one CR2032 batt)engraved the word Agilent on its CPU?Bye.-====Well, perhapsat Corvallissure it couldI suspect that's in the natural order of things. When theydesigned the 48SX they were doing something new. The features camefrom the designer's imaginations. Probably no-one else knew eoughabout the possibilities to interfere.But when they were working on the 48GX the marketers and executivesalready had a working model and had already probably heard amillion ideas and probably all had their favorites and did whatthey could to see that they were implemented. The designers musthave had less control with that never having been stated.With the 49G they also had some TI successes to compete with andthat influenced the list of features, too. Less and less controlby the designers.I'm just guessing at all of this. I have no slightest idea whatwent on at HP. But I've led programming projects that weretotally new ideas and no-one really understood the possibilitiesand I know how I got left alone. If I wanted a suggestion I had toask for it.Then, doing the second version of those same projects, everyone hadSome I couldn't. The second version might have been more usefuland more tailored to the user's needs, but it was always lessgraceful, more confused, less coherent and more trouble tomaintain.It's just how things work in business. The designer has a lot moreability to express his own ideas in his own way in the firstedition of something very new than at any other time.Of course business has developed a partial cure for the designer's====As there is so much talk at the moment about new directions I'd like to ask:Just what is the perfect calculator each of us expects? If somebody usesthis for market research then great I'll have a replacement for my 49 in afew years.If you assume it's going to be 2-3 years before a new calculator gets to themarket we can assume that the PDA's will by then be fully functional windowsdevices. You'll able to run apps lick Mathcad, Maple etc with the onlyrestrictions being screen real estate and hardware interface (touch screenlimited keys. Any new calculator would have to compete with these and besufficiently different in functionality to compete.So here is my stab at the basics:Size:Pocketable, smaller than HP49, size of HP42 if possible. The 48/49's are tobig to be carried anywhere in your pocket.Whats acchievable will depend on interface and screen. One option could bepull out or flip out screen that retracted displays single line of resultlike a cheap calculator.Power:Uses standard batteries say 2xAAA. Batteries life of 2 months minimum.Screen:HP49 seems reasonable but higher resolution. 2 colour to keep cost and powerdown.User Interface:Has to use a keyboard for rapid input of calcs. RPN of course but Algebraicwill be needed since RPN will have even less of a profile than now.48 and 49 seem to have it about right for an advanced calculator. You can'tget away from the need for a numeric keypad plus extra's for commandfunctions, advanced features and menu. Touch screen just won't cut it as Idon't want to stop to pull out a stylus in the middle of a calc.Memory:1.5 Mb in the 49 seems to be heaps. With memory so cheap just give itaccess to memory flash cards or a heap of internal memory.CPU and OSCan't comment except it should be capable of emulating the HP48/49functionality with improved speed.A dream would be to have the capability to run the same apps as your Desktop====XtobeX====Maybe a shrunken Xpander?Raymond====gets to thefunctional windowsthe only(touch screenthese and beI have Derive, Mathcad and Matlab for my HP 200lx. The version ofDerive I have was designed with the lx in mind. I'm not sure aboutthe others but they work just fine. I don't have any need forthese and my math is limited so I can't judge how they compare tothe HP49. But they are certainly available in an HP49 size packagewith a screen about 3 times the size of the 49 screen and a QWERTYkeyboard with a numeric keypad. (too small for touch typing). Andgood old HP sturdiness and reliability.The 200lx is almost exactly the same size as the 49G. It will fit(snugly) in the HP48 soft case. It's a PC that uses 2 AA batteriesand gets about 40 hours from them and weighs 11 ounces includingbatteries. Rechargebles are fine. There is a recharger built in.It only knows how to charge standard Nicads but there is freesoftware downloadble to control the charging in various ways. Itcan recharge the 1600 mah NIMH just fine.It also has Lotus 123 version 2.4 (complete) in rom, along with alot of PDA apps, more powerful ones than are available in newerPDAs. Dos 5.0 is in ROM. Also Quicken, CC:Mail and an HPbussiness calculator (which talks to Lotus). Several full sets ofinternet software are available. Some free, some not. Severalcards for wireless access are avaliable, as are modem cards. Ithas infra-red. A group of users are now looking into the possibityof Bluetooth. The power supplied to the PCMCIA port is limited sothis may or may not work. With a Parallel port PCMCIA card itworks fine with a Zip drive.It also runs Microsoft Flight Simulator 4.0, Word Perfect 5.1,dBase 3+, Paradox, most older compilers of most languages and aworld of commercial games. All that and it fit's in my pocket. Italso has a PCMCIA slot. I have a 160 meg flash card. It's my A:drive.====it sounds great! now, if i could put my maple7 (o mathematica) in it...may be, the next HP49G will come with a crusoe chip :)oh yeah, and running linux tooBTW: HP49G still rules!!!!!!!!!when you take the thought of others, reality becomes ready made====toI can easily fit a HP49G in the pocket inside my jacket.beAnd it'll be cheap quality too. Fold lines destroy cables.Duracell AA MN1500 are 2450 mAh, so two months on two of these will allowyour device to use 116 mW @ 5 volts (continous use; 106 hours a month).Built-in Lithium Ion will be better though, both for the environment and foryour wallet.powerIt'll have to be a little larger, so you won't have to use a magnifyingglass to see on it. Else you could use a video camera display - 2 & maybe400x300 pixels.You seem to have big pockets;-)A 15C or 42S has the maximum weight I want to carry,aside from their nicer form factor.resultMaybe.====The weather in Oz is to good to be wearing jackets. Even in winter onlywear them on bad days. :)I want something for a shirt pocket, HP49 is to big. Be thankfull I haven'tasked for a Calc suitable for the back pocket of pants when wearing a====Yes, I have. AFAICT, its docs are all here's how to do