HP Forum Archive 14

[ Return to Index | Top of Index ]

OpenRPN - A Dissenting View Message #1 Posted by Steve S on 15 Apr 2004, 7:50 p.m. ...Since I've written in the Forum recently about new calculator design activities, I won't bore any of you by rehashing old material. But I would like to articulate a couple of thoughts, just for the record. In terms of presentation, a tip of the hat to Nelson for suggesting a clever writing approach! <Thoughtful Mode On> 1. Reading the existing discussions, it still seems to me that most people here are simply looking to resurrect a slightly improved version of an existing HP design, like an HP-41++ or an HP-16++. I think it would be a terrible waste of time and talent to settle for such a thing! It's the 21st Century! Can't we collectively think of something a bit more clever than how rounded keys should be or building it out of aluminum? 2. Rhetorical Questions: If HP thought this way, would they have ever leapt from HP9800 to HP-35? Or produced the HP-75 / 71? Or even the HP-48? Didn't all of us look to HP to make quantum leaps in product design and capability over the years? Why should we settle for less? <Thoughtful Mode Off> <Personal Preference Mode On> 3. I like crisp, double-shot keys as much as the next guy. 4. Keyboard label clutter beyond a certain point is undesirable. 5. More display area is ALWAYS, ALWAYS better. 5a. Color displays are better than monochrome. 6. A folding design (a la LX100) is an acceptable form factor for a dedicated calculator. 7. Well-implemented soft keys are a good way to preserve access to functions while avoiding clutter. 8. An original calculator design is more appropriate as a project than a warmed-over redesign of an "old classic." <Personal Preference Mode Off> <Crazy Mode On> ...In my mind, it's long past time to think beyond calculators as we know them today. My ideal design would be a slate, 8.5 x 11 or somewhat smaller, that you would WRITE numbers, calculations and equations on. Hand writing recognition would convert these to characters and solve the problems. Graphs would be interactive and would pan and zoom just as CAD systems do (and as easily!). As each "page" filled up, a new one would scroll up, and all previous pages would be saved, JUST AS YOU DO WITH PAPER. Well, OK, that's crazy talk. How about this: An LX-100-sized case with a large color LCD screen, flanked by two or four (2 left and 2 right) columns of soft keys. The other half of the case would have a conventional arrangement of one-per-function keys with, say, one or at most two shift keys. And YES, put the damn ENTER key right where it belongs! Calculations acummulate and scroll on the screen, saving to one of those nifty small disc drives. As various functions and modes are invoked, labels along BOTH right and left edges of the screen change to "label" the adjacent soft keys and maintain the one-per-function approach. USB connectivity assumed, and windows-like file manipulation to load and swap libraries of functions and data. Now, THAT would be a calculator worthy of our combined talents..! <Crazy Mode Off> ...Hey, wait a minute! This damn key is stuck down! Oh, no.....

Re: OpenRPN - A Dissenting View Message #2 Posted by Hugh on 15 Apr 2004, 11:53 p.m., in response to message #1 by Steve S Steve, First, I would like to say that I appreciate your input. Dissenting views are at least as important as votes of confidence and support (often more so). Now, allow me to respond and explain some of my logic behind the direction of this project. Indeed, a large portion of what has been seen thus far may appear insignificant or that we will be slightly modifying older designs. But there is quite a bit more involved. The key to innovation is looking closely at what has been done before and finding ways to embellish upon it. So we're accordingly starting with a solid base from which to work. People have voiced a nearly unanimous desire for something extremely durable with a timeless appearance and tactile keys. In terms of software, this project is an open book. I have a strong feeling that with the community developing what the community wants... everyone will be suprised by the level of innovative ideas. For lack of better words, give it a little time. I just hatched this egg, so give it a chance to grow and develop (even the case design may suprise you). I agree that a 8.5x11 tablet would be a great way to do math. I also think that I speak for a lot of people around here when I say that it's nice to have a solid scientific calc in your front shirt pocket for straight number crunching. Looking over you post as a whole, I'd say we're more on the same page than not. Again, thanks for your opinion! I hope to hear more of them as the project progresses. Best Regards, Hugh OpenRPN

Quantum Leaps? Message #3 Posted by Wayne Stephens on 16 Apr 2004, 7:31 a.m., in response to message #1 by Steve S Thanks for the dissenting opinion, but I thought we were having a free-form discussion about building "the calculator that HP refuses to build". That would, by definition, require that it be a CALCULATOR, not a handwriting recognition tablet. I especially didn't realize we were trying to make quantum leaps in anything. Now I'll have to dust off my Physics texts... Seriously, I don't think most of us are necessarily trying to generate any great advances in technology. We're trying to use materials we can acquire, and to take the best features from the best calculators, and put then all into one package. In fact, all some of us WANT is an upgraded/updated/super fast HP 41/42/43?, etc. Perhaps I am small minded, but my impression was that we were trying to design/build something WE want, not something we think someone else might want, or which might be some new breakthrough. I know that's why I've entered the discussion. Take care. Wayne.

Re: Quantum Leaps? Message #4 Posted by Steve S on 16 Apr 2004, 8:54 a.m., in response to message #3 by Wayne Stephens <<I thought we were having a free-form discussion about building "the calculator that HP refuses to build". That would, by definition, require that it be a CALCULATOR, not a handwriting recognition tablet.>> ...A handwriting recognition tablet, if it is optimized to perform calculations, is, BY DEFINITION, a calculator. It's just uses a different interface... <<In fact, all some of us WANT is an upgraded/updated/super fast HP 41/42/43?, etc.>> ...Yes, I know. My point is, if that's all you figuratively want, then simply emulate a 41/42/43 on a Pocket PC. Project over. My experience is that it's really hard to design and build real hardware, and make it work (nearly flawlessly. Reliable computation is essential for a calculator, no??). Spending so much time and energy simply to build a faster HP-42 is a trivial exercise. Why not exert nearly the same time and energy trying to build something a BIT more ambitious?? <<my impression was that we were trying to design/build something WE want>> ...Yes, that is correct. If we design and build it, won't it BY DEFINITION, be what we want it to be? And if all that anyone wants is a fast HP-42, then that's OK. I'm just suggesting that we could all aspire to something a bit better. And I did say this was a dissenting view...

Re: Quantum Leaps? Message #5 Posted by Wayne Stephens on 16 Apr 2004, 9:48 a.m., in response to message #4 by Steve S I HAVE an emulator on my Pocket PC. It's okay, but on my CALCULATORS I want a KEYBOARD. Be as ambitious as you want; but please do not try to make the argument that a Pocket PC or a writing tablet interface is acceptable on a calculator for people who have stated over and over again that they WANT a keyboard. I still think you are misunderstanding the reason for the exercise. I don't think that a bunch of "regular" folks planning, designing and (hopefully) constructing a sturdier, faster, better calculator based on the classic HP designs of the past is a trivial exercise at all. I for one do not have access to modern, high-tech research or manufacturing facilities. I also don't have thousands of dollars to pay someone else to build prototypes. I am, however, prepared to volunteer to participate in hand assembling whatever device we end up with. I think it will be quite an accomplishment if we can get anything built at all. Take care. Wayne.

Re: Quantum Leaps? Message #6 Posted by Bill Wiese on 16 Apr 2004, 3:11 p.m., in response to message #5 by Wayne Stephens Steve Quote: Yes, I know. My point is, if that's all you figuratively want, then simply emulate a 41/42/43 on a Pocket PC. Project over. Uggh!! Calcs are for 'playing with numbers'. Real keyboards on real calcs let you do that quickly, gracefully and effectively - and mentally transparently. I see Palm people using their calcs and pulling out the stylus and pecking and it's just not smooth. Quote: My experience is that it's really hard to design and build real hardware, and make it work (nearly flawlessly. Reliable computation is essential for a calculator, no?). Spending so much time and energy simply to build a faster HP-42 is a trivial exercise. Why not exert nearly the same time and energy trying to build something a BIT more ambitious As I've written before, the hardware and firmware are (relatively) trivial these days. Hardware's way more reliable, and we can readily write firmware that is too. There are probably 100+ off-the-shelf microcontroller variants that have adequate specs for the job - and that can support a C compiler if you don't wanna write in assembler. Lotsa cheap LCD displays out too. Many of us know how to scan keyboards, drive LCDs, and understand microcontroller power managment, etc; some of us have a grasp on BCD floating point math & CORDIC issues - at worst case we could just emulate existing HP transcendental code. Building a tolerable-looking case, keyboard+keycaps, etc. is the hard part. Building a million of something is sometimes far easier than building 250 of 'em - other than the money ;) Bill Wiese San Jose, CA

Re: OpenRPN - A Dissenting View Message #7 Posted by Grant Goodes on 16 Apr 2004, 9:23 a.m., in response to message #1 by Steve S The reason that people are discussing an HP-41+ or HP-16+ (I don't even think most of us are going even as far as a "++" version) is that those calculators represent in some sense the pinacle of calculator design for their areas of application, and as another poster said, HP refuses to make them anymore. This whole idea that if it isn't a quantum leap ahead in technology then it isn't worth considering is just so much 21st Century BS, in my opinion. Yes, HP _did_ make massive strides when going from their early behemoths to the petite HP-35, and forward to the HP-41 series, but those were the early days of the development curve and I would claim that as the technology reached a certain level that we started to see a honing and polishing rather than a vast rethinking. And it is simply inappropriate to throw out all those years of incremental fine-tuning just to adopt the latest buzz-word technological gim-crackerery. For what it was designed for (napkin-size programming tasks and number crunching on the fly), calculators, particularly HP calculators, are just a fantastic tool. And I DON'T want a built-in cell-phone and digital camera (though a plug in peripheral port with such options would always be welcome). It's a calulator, not a hair-drier. And in that area HP developed considerable expertise, and refined the product to the point where further improvements (aside from more memory or faster CPU, lower power consumption, etc) become tough to make. In fact, I would argue that modern "improved" devices are just cheaper to make (as if we care: I was happy to pay for the quality I got in 1981, and am appalled at what I get today) and have features which I frankly don't want (like a colour display, or speech-to-text, or whatever) and have a crappy physical interface (cramped and low-quality buttons). Bottom line: What _I_ would like to see developed for RPN should provide a high-quality physical platform (case quality, keyboard physics, and display legibility matter VERY much) to ressurect the already very highly developed HP calculator IP (mathematical routines such as SOLVE and even just trig functions that work), and go from there. grant..

Re: OpenRPN - A Dissenting View Message #8 Posted by Wayne Stephens on 16 Apr 2004, 9:35 a.m., in response to message #7 by Grant Goodes Thanks, Grant. Once again, someone else has said it better than I. Take care. Wayne.

Re: OpenRPN - A Dissenting View - comment to Grant's comments, about cellphones... Message #9 Posted by Bill Wiese on 16 Apr 2004, 3:29 p.m., in response to message #7 by Grant Goodes Grant, Your comments about cellphones are well-taken and apropos to this calc discussion. I don't need/want all those extra features. Camera? Hah, I have a real one (Sigma) for real digital pix. And it works at night too ;) They keep making phones smaller - there's one Siemens model that's almost suppository-sized. The only innovations for me that count in cellphones are (1) battery life/current drain (2) ease of use of keyboard (3) ease of storing #s while in middle of call (4) LCD viewability/angle, etc. & sunlight readability (5) not dropping calls-in-progress (sometimes a system-dependent configuration - how bad does the call get before dropping out?) I wish we could get today's low-power electronics in a thin Motorola mid-1990s "flip-phone" (*not* StarTac) style. (Those were GREAT keyboards.) That way we'd have a nice big battery for lotsa talk/lotsa standby and fewer recharge cycles. Digital (nonanalog) cellphones (and their networks) need also to have something I call a 'survival vocoder'. Typical cellphone voice compression methods compress 3kHz bandwith speech down to 8-13kilobits/sec depending upon system in question. Supposedly these voice coders are trying to be 'near-toll-quality' when there's a good signal path. When the RF path btwn phone & cellsite isn't good (too far, obstructions, noisy channel, adjacent channel interference, etc) the vocoder doesn't get all the bits it needs to reproduce the speech adequately. When more than a few bits can't be error-corrected, you get all sorts of squeals & artfacts. The cellphone companies believe they should just drop your call rather than have you suffer. But I'd much rather have the phone & cellsite drop down into a different low-bit-rate (800-2000bps) military-style voice coder when the RF channel's signal/noise ratio (SNR) drops appreciably. Sure, it sounds raspy & buzzy (ever hear tapes of US military pilots talking on some of their radios?) when compared to a properly operating higher-bit-rate coder but it may get thru fine when the channel can only support 1/4 the usual bitrate. And this is not too hard to do - just some different DSP firmware. Of course, Harvard MBAs would never think of that. Bill Wiese San Jose CA

HP-15S Message #10 Posted by Karl Schneider on 17 Apr 2004, 9:59 p.m., in response to message #7 by Grant Goodes Grant -- Well stated! Amen, brother! Philosophically, there's nothing I could add to that. Regarding details of implementation, I do have one idea: HP-15S A resurrected 15C, with the following improvements: Faster (Pioneer speed at least -- about 12x as fast) More RAM -- at least 100 numbered registers (with 98 allocatable) Make "LN" unshifted, rather than "10^x" (as on Pioneers) Make programming keycodes show "A"-"E" when appropriate (as on 16C and 20S) "ENG +" and "ENG -" functions to shift engineering-format exponents by +/- 3 (as Casios can) Glare-free, dust-resistant display window (as on later Pioneers) Scratch-resistant, durable bezel with printed model insignia (as on 12C Platinum) 12-digit display and higher computational accuracy would be nice. However, this would require a redesigned display unit and, quite possibly, extensive revisions to ROM microcode. (Note: I wonder if KinHPo even has the microcode. If not, maybe Eric Smith could provide them his 14-kword scan if requested...) True, this product spec does not provide for data communication or AOS, but its form factor with 7-segment display and fully-utilized keyboard doesn't allow for much else. I have some other ideas for a "12B" and a "32Siii" -- what the 12C Platinum and 33S ought to have been. However, KinHPo had other ideas... -- Karl S. Edited: 18 Apr 2004, 3:40 p.m.

Re: HP-15S Message #11 Posted by Eric Smith on 20 Apr 2004, 2:03 a.m., in response to message #10 by Karl Schneider Quote: I wonder if KinHPo even has the microcode. I've been told by usually reliable sources that HP no longer has most of the old calculator ROM code in either source or object format. They do still have the 12C source code, though that was apparently nearly lost and required a significant effort to recover. Each time the calculator responsibility moved from one HP division to another (from Corvallis to Singapore to Australia to San Diego), huge amounts of records were either lost or deliberately destroyed. Quote: If not, maybe Eric Smith could provide them his 14-kword scan if requested...) My original report of a 14 Kword ROM size for the HP-15C was incorrect. I had not noticed a 2K gap in the self-test trace, so it's really 12 Kwords. Each of the R2D2 chips (1LE2 or 1LH1) has 6K of ROM aligned to an 8K boundary. One is from addresses 0000-17FF hex, and the other is from 2000-37FF hex. I doubt very much that HP would ask me for a copy. Even if they don't still have it themselves, I doubt that they'd want to do anything with it. However, I'd be happy to discuss it with them. They'd most likely require me to sign an NDA first, as they did a few years ago when they considered having me write a 12C simulator for a product that was under development (but got cancelled).

Re: HP-15S Message #12 Posted by Veli-Pekka Nousiainen on 20 Apr 2004, 2:46 a.m., in response to message #11 by Eric Smith I suggest that you immediately contact Cyrille de Brebisson at HP!!! [VPN]

Re: HP-15S Message #13 Posted by Eric Smith on 20 Apr 2004, 2:48 p.m., in response to message #12 by Veli-Pekka Nousiainen Quote: I suggest that you immediately contact Cyrille de Brebisson at HP!!! VPN, since I can't figure out your email address, I suggest that you immediately contact me. My email address here is not obfuscated.

Re: HP-15S Message #14 Posted by Veli-Pekka Nousiainen on 21 Apr 2004, 4:09 a.m., in response to message #13 by Eric Smith You could just DROP LETTER from the address. Sorry about the confusion - the idea is mine: Cyrille is totally unaware of these ideas. I still think that you should try to persuade him! [VPN]

Re: OpenRPN - A Dissenting View Message #15 Posted by Bill Wiese on 16 Apr 2004, 3:01 p.m., in response to message #1 by Steve S Hi all, Well, your dissenting view is fine, you seem to want a calculator to do what a computer readily does better. Many early prog calcs were designed for folks who needed some number crunching but couldn't get time on the minicomputer. That's moot now. When I need big number crunching, matrices, etc. I'll be near a computer - Excel, Matlab or quickie C programs all do their part. Keystroke programmability is important as that lets you essentially have custom functions for quick one-button solutions. (I'm gonna load up my HP41C w/keystroke external ballistics programs to deal w/custom cartrdige loads - though I hesitate to take calc to the rifle range.) But a calc is for playing with numbers on a smaller scale. Gotta be easy on the hand/fingers/eyes. That's where real RPN + a well-designed keyboard + form-factor help. So, yes, I don't mind staying in the past. Making a calc into a full-house computer that can plot, etc. doesn't do anything for me... My dream calc: - larger Woodstock form-factor (size of HP55); - Woodstock case, keyboard layout, keytops/legends; - backlit LCD text display; I can live w/one line, or 1 main line and 1 smaller line of upper-row labels like the 42S; - use 2x"AAA" cells for decent batt life w/backlite; - standard HP sci calc functions; - hex/bin/oct/dec conversions & arithmetic; - some financial functions (at least PV/FV/PMT/i etc) - keystroke programmability; - I/O connector for up/download of prgms/data; Might have to tweak the dead HP25C I have. Bill Wiese San Jose CA

Re: Proper quantum leaps Message #16 Posted by Andrés C. Rodríguez (Argentina) on 17 Apr 2004, 12:42 a.m., in response to message #1 by Steve S It should be noted that a "quantum" is the SMALLEST possible quantity for a energy exchange, step, charge, or whatever. While brisk because of being a discrete quantity (instead of continuous), the much famous "quantum leap" should be a very small one!

Re: Proper quantum leaps Message #17 Posted by Arnaud Amiel on 17 Apr 2004, 2:57 a.m., in response to message #16 by Andrés C. Rodríguez (Argentina) I got some problems with this one when I first arrived in England and quite offended some people when I told them I had heard their work didn't amount to much. Arnaud

Re: Proper quantum leaps Message #18 Posted by Wayne Stephens on 17 Apr 2004, 8:04 a.m., in response to message #17 by Arnaud Amiel That's funny. Thanks for keeping us "down to earth", guys. Take care. Wayne.

Re: Proper quantum leaps Message #19 Posted by Mark A. Ordal on 17 Apr 2004, 2:14 p.m., in response to message #16 by Andrés C. Rodríguez (Argentina) I just wanted to point out that the famous "quantum leap" has nothing to do with already being quantized and changing things a quantum at a time. Rather it refers to the HUGE conceptual leap from the non-quantum universe of Galileo, Newton, etc. to the quantum universe of Bohr, Heisenberg, Fermi, etc. In other words, "quantum leap" refers to the conceptual leap from "classical physics" to "quantum physics" (also called "modern physics"). You can't make such leaps on a routine basis---hence the reason for its notoriety, and the use of the term outside of physics. --Mark

Quantum mechanics is BIG, but quantums are small... Message #20 Posted by Andrés C Rodríguez (Argentina) on 18 Apr 2004, 11:11 a.m., in response to message #19 by Mark A. Ordal Then, why not "Bohrian advance" or "Planckian leaps", or "relativistic jump", or "Columbus journey" or "Armstrongian moonstep" ... all them are important stages of discovery and knowledge advance. (Penicilin, Rosetta Stone, transistors, celestial mechanics and many others also come to mind). I think that "quantum leap" is used more for it's discrete, discontinuous character (to oppose to just evolutionary, which seems rather continuous), than for the importance of modern physics (which I did studied and very much admire, indeed). I am afraid that most of the social communicators, journalists and politicians who regularly talk about "quantum leaps" had not been exposed enough to the Schrodinger wave function, Dirac works, permitted levels of energy, or other concepts, to so fully understand the "dramatic" change implied by the transition from classical to modern physics. I would suggest that we may, from now on, adopt the "HP35 dawn" as a manner to signify a radical change in human knowledge! Thank you all for a lively and stimulating discussion!!

Re: Quantum mechanics is BIG, but quantums are small... Message #21 Posted by Veli-Pekka Nousiainen on 18 Apr 2004, 11:31 a.m., in response to message #20 by Andrés C Rodríguez (Argentina) It's quantum leap in macro scale. [VPN]

Re: Quantum mechanics is BIG, but "quanta" are small... Message #22 Posted by Andrés C. Rodríguez (Argentina) on 18 Apr 2004, 2:45 p.m., in response to message #21 by Veli-Pekka Nousiainen So, the next step will be at least as big as it? Or, perhaps, a whole multiple of it? (fractions would not be allowed; whether macro or not) I'm correcting my previous error: many quantum"s" are "quanta" and not "quantums" as I wrongly wrote before Edited: 19 Apr 2004, 8:59 p.m.

[ Return to Index | Top of Index ]

Go back to the main exhibit hall