Copyright (c) 1989 GEnie Forth RoundTable. Contents may be freely distributed provided credit is given to the GEnie Forth RoundTable. Minutes of guest FIGGY RT Conference. with special guest Chuck Moore, owner Computer Cowboys and creator of Forth as a programming environment. Date: 08/24/89 Time: 22:29EDT Attendees: [[Gary] GARY-S] [[Wheels] VESTA] [[Chuck] PRESS33] <--- guest, Charles Moore [[Jerry] A2.JERRY] [[Dave] DHWEINSTEIN] [[Jim] JTITTSLER] [[Hadil] H.SABBAGH] [[Wil] W.BADEN1] [[Howard] H.HARKNESS] [[Dennis] D.RUFFER] [I.MONTANEZ] [[Len] NMORGENSTERN] [[Frank] F.SERGEANT] [DANMILLER] [[RubberDust] FIGCHAPTERS] [[Alan F.] ATFURMAN] [R.VANNORMAN4] [W.ROBERTSON2] [M.HALL] [[mitch] TA] o Special guest, Chuck Moore, discusses the future of Forth, his dismay at current computing concepts, his latest project, and some philosophy we can all profit by. o There was a short prelude file, GEnie file #1805, uploaded prior to the conference to set the tone and generate discussion. It is included here for convenience: Chuck Moore, the creator of Forth, is the special guest in conference on the GEnie Forth RoundTable for 24 August 1989. The topic of Chuck's conference is 'The future of Forth...' . The following is a brief prelude to the conference designed to set the tone and stimulate discussion. Chuck uploaded it prior to his conference. Gary Smith, host SysOp Prelude: The 'Future of ...' is a catchall for everything having to do with Forth. Its current place in the world is impossible to determine - and largely irrelevant. Forth is a valuable tool - and will remain so - regardless of the number using it. Recently I was obliged to use conventional CAD software. I am dismayed that it hasn't evolved from the 60's. Forth is the only hope for improved software, ignoring the ever-hopeful AI and neural nets. Computers are getting ever more complicated, in violation of the first principle of human activity: KEEP IT SIMPLE . In respect for this unique forum - 25 words or less - I offer the following statements to challange/guide question/comment. 1- I like classic Forth. 2- This includes BLOCKs - simpler, faster, better than files. 3- VOCABULARY has been mis-used by FIG-Forth. It is a poor substitute for fast compile. 4- Forth must evolve. Standards are very dangerous. 5- ANSI committee deserves thanks for 'above and beyond call of duty'. Theirs is the impossible dream. 6- Marvelous opportunity for non-ANSI Forths. 7- Forth architecture is superb for micro (macro) computers. Many variants should be explored. 8- Three keys are necessary and sufficient. QWERTY is a joke. 9- Marvelous opportunity for non-IBM PCs. 10- Work smart, not hard - Forthought. 11- A program that can do everything (ie, SPICE) can do nothing well, fast, easily. 12- PUSH and POP are better names for >R and R>. 13- Multiply is a much over-used arithmetic operation. Ie, FFT can be replaced by Walsh-Hadamard. 14- Floating-point is a bad joke. 15- Withold source code only when you're ashamed of it. Over to you, Gary Chuck Minutes: is here <[Gary] GARY-S> any last questions chuck ? I'm Jerry Kindall... A2 RoundTable Assistant is here. Sorry, guys, but Jax is being FigChapters tonite. <[Dave] DHWEINSTEIN> It's gonna be a nice crowded conference. Actually, I'm JaxBoss, Steve. <[Gary] GARY-S> conf in 4 minutes is here. <[Wil] W.BADEN1> is here. is here. is here. is here. is here. <[Gary] GARY-S> <[Gary] GARY-S> The GEnie Forth RoundTable is very pleased to welcome ... <[Gary] GARY-S> Charles 'Chuck' Moore, the creator of the Forth language. ... <[Gary] GARY-S> Chuck, who is doing business as Computer Cowboys, ... <[Gary] GARY-S> discovered Forth twenty years ago. .. .. <[Gary] GARY-S> He spent 15 years creating applications with FORTH, Inc. ... <[Gary] GARY-S> in areas such as control, analysis, database...and numerous others ... <[Gary] GARY-S> Mr. Moore designed the Novix NC4000, now the Harris RTX... <[Gary] GARY-S> the world's first, and not the last, Forth engine. ... <[Gary] GARY-S> He is 'Amazed the world is so timid it is... <[Gary] GARY-S> willing to let us (few in the Forth community) ... <[Gary] GARY-S> monopolize a good idea' ! .. .. <[Gary] GARY-S> Chuck, welcome to the GEnie Forth RoundTable .... <[Gary] GARY-S> and thanks very much for comming and sharing this evening. ... <[Gary] GARY-S> .. .. .. .. ga Chuck <[Chuck] PRESS33> Forth is the best computer language... <[Chuck] PRESS33> I'll be using it another 20 years... <[Chuck] PRESS33> With a few changes... <[Chuck] PRESS33> Too bad Veger's not Forth. <[Chuck] PRESS33> I'm not used to QWERTY... <[Chuck] PRESS33> but I'll try... <[Chuck] PRESS33> Questions/comments? <[Dave] DHWEINSTEIN> What new hardware/software projects are you working on? <[Chuck] PRESS33> Micro called SH-BOOM... <[Chuck] PRESS33> So far have designed version called SH-BOO... <[Chuck] PRESS33> Couple missing features... <[Chuck] PRESS33> Software to support - namely CAD is here. <[Hadil] H.SABBAGH> Could you explain the idea of "three keys are nec... <[Hadil] H.SABBAGH> and sufficient" ? ga <[Chuck] PRESS33> I have used a 7-key keypad for years... <[Chuck] PRESS33> I's hard to learn - offers great security... <[Chuck] PRESS33> Three keys select from 7 menu items... <[Chuck] PRESS33> Can be color coded... <[Chuck] PRESS33> And is easy to learn <[Dave] DHWEINSTEIN> I was wondering... <[Dave] DHWEINSTEIN> ...what the features of the SH-BOOM (or the SH-BOO) were. <[Chuck] PRESS33> I bet... <[Chuck] PRESS33> Can't say much, but... <[Chuck] PRESS33> 32-bit, 50 MHz, DRAM interface, $20 <[Gary] GARY-S> Chuck - is this the Oki project <[Chuck] PRESS33> yes... <[Chuck] PRESS33> I guess Oki is state of the art... <[Chuck] PRESS33> But Sun Valid and IBM software are all dreadful is here. is here. <[Frank] F.SERGEANT> do you have thoughts on an equiv. breakthru to the 3 key keybd ... <[Frank] F.SERGEANT> for use as an output device? ga <[Chuck] PRESS33> No... <[Chuck] PRESS33> I figure to stick with multi-sync monitors <[RubberDust] FIGCHAPTERS> In magazine articles .... <[RubberDust] FIGCHAPTERS> i have (pretentiously) referred to Forth as ... <[RubberDust] FIGCHAPTERS> " a re-thinking of the nature of serial computation" ... <[RubberDust] FIGCHAPTERS> we are not far from ... <[RubberDust] FIGCHAPTERS> parallel desktops ... <[RubberDust] FIGCHAPTERS> what is the future of Forth in parallel? <[Chuck] PRESS33> Doesn't so much rethink as follow the rules everyone agrees on... <[Chuck] PRESS33> Parallel sounds great, but costs much... <[Chuck] PRESS33> And I've yet to meet a parallel problem... <[Chuck] PRESS33> But I plan to connect SH-BOOMs... <[Chuck] PRESS33> and program them in Forth... <[Howard] H.HARKNESS> How about speech input? <[Chuck] PRESS33> Noone wants to hear elevators talk... <[Chuck] PRESS33> I doubt they want to talk to them... <[Chuck] PRESS33> Besides, speech is not private.... <[Alan F.] ATFURMAN> Now what would Brave New World be like... <[Alan F.] ATFURMAN> without talking elevators...regarding parallel Forth... <[Alan F.] ATFURMAN> MicroProcessor Engineering Ltd, Southampton, England... <[Alan F.] ATFURMAN> have [<-check out Brit grammar] just announced... <[Alan F.] ATFURMAN> Transputer Forth in beta test. <[Chuck] PRESS33> New Coke... <[Chuck] PRESS33> Great... <[Chuck] PRESS33> Forth, unlike Occam, does not require all problems... <[Chuck] PRESS33> to be solved in advance... <[Chuck] PRESS33> Size increases exponentially with scope of problem <[Dennis] D.RUFFER> you say you will hook sh-boom up in parrallel... <[Dennis] D.RUFFER> have you heard of anyone, yet doing that with... <[Dennis] D.RUFFER> novix or harris chips? <[Chuck] PRESS33> Only the classic Lockheed 8 serial Novix... <[Chuck] PRESS33> I disagree that we've reached any limit with serial computers. .. <[Chuck] PRESS33> I can see SH-BOOM going to 10 GHz. <[Wheels] VESTA> Regarding your provocation file ... <[Wheels] VESTA> you state that floating-point is a joke, and multiplication overused ... <[Wheels] VESTA> I agree that floating-point is often (maybe usually) ... <[Wheels] VESTA> unnecessary, but I would like you to elaborate some. <[Chuck] PRESS33> FP a great convenience for one-time calculations... <[Chuck] PRESS33> Drawbacks are amount of hardward/soft... <[Chuck] PRESS33> and uncontrollable round-off... <[Chuck] PRESS33> Numeric codes weather, hydro ... <[Chuck] PRESS33> should have carefully analyzed numeric representations... <[Chuck] PRESS33> which means scaled integer/fixed point <[RubberDust] FIGCHAPTERS> If you hook SHBOOM together ... are you going to publish app notes on such arrangement? ... <[RubberDust] FIGCHAPTERS> and will there be a cmForth for it too? <[Chuck] PRESS33> I gave away lots of stuff for Novix... <[Chuck] PRESS33> The deal with SH-BOOM is much more commercial... <[Chuck] PRESS33> The software I want to use is not cmForth... <[Chuck] PRESS33> Still, I'd like it available... <[Chuck] PRESS33> So, the jury is still out <[Howard] H.HARKNESS> I guess Chuck has never gotten onto an elevator with... <[Howard] H.HARKNESS> two bottles of wine and a bouquet of roses. We want to talk to them. ga <[Chuck] PRESS33> How about the depressed robot in Hitchhiker's... <[Chuck] PRESS33> Or arguing with your autopilot? <[Frank] F.SERGEANT> when you say Forth should be under 4K bytes, what do you include/exclude? ga <[Chuck] PRESS33> Assembler, compiler, interpreter, editor <[Wheels] VESTA> What applications do you see for desktop networks ... <[Wheels] VESTA> of Forth machines, not necessarily total parallel ... <[Wheels] VESTA> computers? ga <[Chuck] PRESS33> Shared database... <[Chuck] PRESS33> In the sense that each machine... <[Chuck] PRESS33> has a copy of the data it needs... <[Chuck] PRESS33> but no personal disk... <[Chuck] PRESS33> Archival storage is shared... <[Chuck] PRESS33> PolyFORTH does this very well <[Dave] DHWEINSTEIN> If you were doing Forth all over again (no body of existing code, or other Forth programmers), what, if anything, would you change or do differently? ga <[Chuck] PRESS33> That is the situation I face... <[Chuck] PRESS33> And have for years... <[Chuck] PRESS33> I don't seem to share much code... <[Chuck] PRESS33> What I plan to do... <[Chuck] PRESS33> is implement 3-keys very nicely... <[Chuck] PRESS33> in combination with de-compiler... <[Chuck] PRESS33> I see nothing better than present Forth syntax... <[Chuck] PRESS33> but don't think I'll need interpreter/compiler <[Len] NMORGENSTERN> You include assembler, compiler, interpreter, and editor <[Len] NMORGENSTERN> I assume you also include keyboard i/o in your 4k <[Len] NMORGENSTERN> You do not include disk management. <[Chuck] PRESS33> No, I include both in interpreter... <[Chuck] PRESS33> I'd leave out multi-programmer... <[Chuck] PRESS33> since computer/user is optimal now. <[Frank] F.SERGEANT> re 4k - even on regular micros? Even 8088? I'm at about 12K now so ... <[Frank] F.SERGEANT> I have a lot of work left. ga <[Chuck] PRESS33> ah... 86 is a nasty assembler... <[Chuck] PRESS33> but most instructions are useless... <[Chuck] PRESS33> Novix is certainly 4K... <[Chuck] PRESS33> 1802 was... <[Chuck] PRESS33> 6502 probably... <[Chuck] PRESS33> but not the new biggies, maybe <[Wheels] VESTA> Any plans for lower-cost 16-bit Forth chips? ... <[Wheels] VESTA> Company I work for makes boards for embedded control ... <[Wheels] VESTA> So 32 bits is overkill ... <[Wheels] VESTA> and we can't produce boards for our customers ... <[Wheels] VESTA> when quantity RTX price is more than our boards. ga <[Chuck] PRESS33> absolutely... <[Chuck] PRESS33> Could do 16-bit version of SH-BOOM if market there... <[Chuck] PRESS33> Markey for 16 10 times that for 32... <[Chuck] PRESS33> But lots of demand for overkill... <[Chuck] PRESS33> SH-BOOM parts cost $100 including 1Mbyte ram <[Jim] JTITTSLER> is here. <[Hadil] H.SABBAGH> Are you suggesting that numerical analysts... <[Hadil] H.SABBAGH> build their own hardware for every new program? ga <[Chuck] PRESS33> No... <[Chuck] PRESS33> I'm suggesting they analyze their numbers... <[Chuck] PRESS33> Every time someone counts Mflops seriously... <[Chuck] PRESS33> I consider they forgot the problem... <[Chuck] PRESS33> For example, SH-BOOM has 32 x 'n' multiply... <[Chuck] PRESS33> Much signal processing can be done with 4-bit coefficients. <[Dave] DHWEINSTEIN> What is your opinion of some of the concepts being experimented with in Forth systems today, such as multiple-cfa Forths, object- oriented Forth extensions, and the possibility of a "type" stack? <[Chuck] PRESS33> May a thousand flowers bloom... <[Chuck] PRESS33> But these concepts are aimed at expanding the... <[Chuck] PRESS33> scope of a program... <[Chuck] PRESS33> I'd rather have 10 small programs than one 110 times as large. .. <[Chuck] PRESS33> Can you see my 1111 repeat? <[RubberDust] FIGCHAPTERS> History lesson for relative newcomer, please! What is biggest iron you ever personally implemented FORTH on? (Mainframe? Cray?) <[Chuck] PRESS33> I guess IBM 360/50... <[Chuck] PRESS33> Or SDS 92 something (24-bits)... <[Chuck] PRESS33> No - 1108 <[Chuck] PRESS33> ... <[Chuck] PRESS33> I never got a chance at the big ones... <[Chuck] PRESS33> and now that I could, I don't want it any longer is here. <[Frank] F.SERGEANT> 2 questions: 1. Can you say how you are prototyping SH-BOOM? ... <[Frank] F.SERGEANT> 2. Re fig misuse of vocabs - what are your suggestions ... <[Frank] F.SERGEANT> Need more than just COMPILER & FORTH? ... <[Frank] F.SERGEANT> Menus replace vocabs? ga <[Chuck] PRESS33> Yes, as in cmFORTH COMPILER FORTH enough... <[Chuck] PRESS33> except perhaps EDITOR... <[Chuck] PRESS33> But, exactly, menus substitute nicely... <[Chuck] PRESS33> FIG tends to precompile everything into seperate VOCABULARYs.. <[Chuck] PRESS33> This exacerbates slow compile, and has horrendous interactions anyway... <[Chuck] PRESS33> Prototyping quite a problem... <[Chuck] PRESS33> We're doing a gate array (8000 gates)... <[Chuck] PRESS33> But this is almost as much work as full custom... <[Chuck] PRESS33> The turn-around and cost advantages don't obtain... <[Chuck] PRESS33> when you press the envelope. <[Gary] GARY-S> Jax you have the priviledge of the last question tonight - ga jax <[RubberDust] FIGCHAPTERS> GENie user commented recently that his (defense) project replaced autorouter in Symbolics LISP with Forth app on RTX. Comme nts on uses of FOrth in symbolic computing problems? <[Chuck] PRESS33> Forth ideal for symbolic computation... <[Chuck] PRESS33> and other applications with unique data representations... <[Chuck] PRESS33> But I wouldn't represent data in ASCII... <[Chuck] PRESS33> Some nicely compact binary-coded representation is much more efficient... <[Chuck] PRESS33> and easy to handle with custom words. is here. <[Gary] GARY-S> closing comments please, Chuck - ga <[Chuck] PRESS33> Thank you all for your interest and support... <[Chuck] PRESS33> I have been discouraged from time to time the last 20 years... <[Chuck] PRESS33> and am always reassured about the viability of Forth by the people who use it...Recently, I have been re-dismayed at the quality of commercial software... <[Chuck] PRESS33> There are not words strong enough to express my shock... <[Chuck] PRESS33> It might be better than 30 years ago, but not much... <[Chuck] PRESS33> The achievement seems to be: mainframe schlock on a micro... <[Chuck] PRESS33> And somehow, noone cares... <[Chuck] PRESS33> I think the future of computers depends upon... <[Chuck] PRESS33> the quality of software... <[Chuck] PRESS33> Aviation Week is a litany of bugs, delay and cost... <[Chuck] PRESS33> How can we present the solution to what are becoming... <[Chuck] PRESS33> major social problems?... <[Chuck] PRESS33> Thanks, and good night. <[Gary] GARY-S> G-o-o-o-o-o-D STUFF !!!! <[Gary] GARY-S> Chuck - on behalf of the GEnie Forth RoundTable .... <[Gary] GARY-S> 'Thank you' very much for a stimulating conference. .... <[Gary] GARY-S> All may stay and chat, but this conference has officially ended. <[Hadil] H.SABBAGH> Thank you <[Howard] H.HARKNESS> thankx bye === End of Steno notes. ===