00:00:00 --- log: started forth/02.08.02 00:19:43 Very 00:20:49 how fast compared to a pentium? 00:20:57 (and the right software i suppose) 00:21:44 Well, perhaps I shouldn't say faster. 00:21:48 Well, yes, it actually would b.e 00:21:57 For straight-ahead execution, it'd be just as fast. 00:22:08 But when taking a branch, however, it'd rock the house on the Pentium. 00:22:55 The Pentium has 25% of its transistors devoted to branch prediction just so it can try to keep the pipeline filled at all times. 00:24:37 And if the branch prediction turns out to be false, then it has to flush not only the pipeline, but the scoreboard which keeps track of renamed registers, the instruction queue, and abort post-humously any memory transactions it started on behalf of the code it thought it was going to execute. 00:24:52 In short, a failed branch is unjustifiably expensive. 00:25:33 This is why C programmers don't like short functions -- they're too expensive. This is also why, if you do have short functions, C compilers will sometimes try to compile them in as inline functions. 00:26:09 All this, because of a single CPU engineering concept: the pipeline. 00:28:27 --- join: Kitanin (~clark@h-209-91-66-234.gen.cadvision.com) joined #forth 00:31:54 hmm 00:32:00 heh 00:32:16 pentium is a mess eh? 00:32:28 what about forth with all its short functions running on a pentium? 00:32:42 must be slower because of the branch stuff? 00:35:04 Fortunately, most of it is usually predicted pretty well. But I'm willing to bet you that without a pipeline, a 266MHz processor will be noticably more "agile" than the Pentium. 00:35:59 I mean, the latest Athlons have more than 18 pipeline stages. At a minimum, it'll take 18 clock cycles to refill those stages, not counting the memory accesses required to get the requisite data (and there you introduce wild concept of cache miss overhead, etc) 00:43:37 ouch 00:43:37 heh 00:43:51 so transmeta went out of business?? 00:44:23 are there any cheap stack machines out there? 00:44:41 where stack machine != forthchip.. 00:48:33 Nope. 00:48:49 Every stack machine I'm aware of today is a Forth-inspired chip. 00:53:00 hmm 00:53:26 so you were thinking of building your own stack machine or something? 00:54:06 you said something about building your own something with discrete units instead of messing with those 68???? chips you have 00:54:53 One of my many hobbies is building my own CPUs from discrete TTL components. 00:54:58 65xxx by the way. 00:55:02 65816 to be specific. 00:55:13 TTL? 00:55:21 Never mind. It's just a type of digital logic. 00:55:29 heh 00:55:49 is it easy to put together? 00:55:57 or do you have to solder? 00:56:11 Depends on how you build it and your choice of components. 00:56:32 * futhin is thinking lego pieces, put them together and program them ;) 00:56:38 If you use fairly old technology (DIP packages) with sockets, you can get by almost exclusively with wire-wrapping. 00:57:24 Some soldering will be required for connecting things like decoupling capacitors, but those are generally permanent fixtures anyway. 01:00:00 my dad apparently built his own computer a long time ago, and he had the computer with him until one day he decided to throw it out.. boy was i pissed about it because i was like 2 years into my interest in computers & electronics and i never knew about the homemade computer 01:00:17 never got a chance to look at it :( 01:00:26 --- join: MrReach (~mrreach@209.181.43.190) joined #forth 01:00:33 hihi mrreach 01:00:38 hihi 01:02:11 what's new? 01:02:12 Hehe -- I can imagine. 01:02:19 * kc5tja can still remember the first 4-bit CPU he made. 01:02:26 It wasn't terribly powerful or useful. 01:02:29 re MrReach 01:02:38 blech! 4004 01:02:53 intel? 01:02:59 hm, i'm interested in building a simple cpu.. or programming one.. what would be the easiest and cheapest way to go about it ? 01:03:12 go buy a Lego Mindstorm 01:03:14 i would like to make a forthchip/stackmachine i guess 01:05:59 Note my phraseology: 4-bit CPU I made. 01:06:10 oh! ok 01:06:24 It was not an Intel chip -- it was about four Radio Shack white breadboards chocked full with discrete TTL MSI and LSI components. 01:06:38 It had a whopping 8 instructions too. 01:06:49 * MrReach laughs 01:07:03 Load LED register, Increment LED register, Decrement LED register (LED register was itself 4 bits wide) 01:07:30 Jump, a 1-deep CALL facility (1-deep return stack), return 01:07:48 --- join: Serg_penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 01:07:59 Wait (for a signal to go true), and something else that I can't recall. 01:08:12 hi 01:08:14 In short, there was no way in heaven or hell this thing would solve world hunger, let alone boot an OS. 01:08:24 And it could address 128 bytes (well, 256 nybbles) of RAM. 01:09:21 heh 01:09:22 cool 01:10:07 look at my sokoban 01:10:24 http://cryptomancer.chat.ru 01:11:34 And frankly, the 4004 wasn't entirely useless -- it was every bit as powerful as an 8008, just slower. 01:11:49 (The 8008, in fact, is just an 8-bit version of the 4004) 01:11:55 now i write bomb game, like scorch or worms 01:12:06 serg: or bomber man ;) 01:12:23 asteroids rocked 01:12:29 so to speak 01:13:31 * Serg_penguin doesnt like "spinal" games 01:15:29 "Spinal?" 01:15:35 Scorch is a *fun* game. 01:15:38 the greatest game - chess - is turn-based :) 01:15:58 Serg_penguin: Chess is a joke compared to a game called Go. 01:16:16 spinal = need only spinal brain, no head brain 01:16:37 An interesting description, but it makes sense. I'll have to remember that. 01:16:46 need only "brain" in backbone ;) 01:19:17 -> kc5tja Go is not real-time too :) 01:20:22 and Scorch too - maybe it has r/t mode, but i used to play it turn right way :) 01:21:20 Chess is definitely not real-time either, so I'm confused by your observation on Go. 01:22:14 ??? n/mind 01:24:17 * Serg_penguin liked air combat over net, be4re joysticks and complex flite models 01:25:32 --- quit: Serg_penguin () 01:27:27 --- join: Serg_penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 01:28:54 but in modern IL2-sturmovik i fite w/ complex flight model, let alone any enemy 01:29:29 and in MS combat sim, one can't even fly straight w/o stick ;( 01:31:56 --- quit: Kitanin (bear.openprojects.net irc.openprojects.net) 01:31:57 --- quit: Serg_penguin (bear.openprojects.net irc.openprojects.net) 01:31:57 --- quit: Soap` (bear.openprojects.net irc.openprojects.net) 01:31:57 --- quit: kc5tja (bear.openprojects.net irc.openprojects.net) 01:31:57 --- quit: MrReach (bear.openprojects.net irc.openprojects.net) 01:31:57 --- quit: futhin (bear.openprojects.net irc.openprojects.net) 01:31:57 --- quit: goshawk` (bear.openprojects.net irc.openprojects.net) 01:31:57 --- quit: ChanServ (bear.openprojects.net irc.openprojects.net) 01:32:11 --- join: ChanServ (ChanServ@services.) joined #forth 01:32:11 --- join: Serg_penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 01:32:11 --- join: MrReach (~mrreach@209.181.43.190) joined #forth 01:32:11 --- join: Kitanin (~clark@h-209-91-66-234.gen.cadvision.com) joined #forth 01:32:11 --- join: futhin (~thin@24.64.175.61) joined #forth 01:32:11 --- join: kc5tja (~kc5tja@ip68-8-206-226.sd.sd.cox.net) joined #forth 01:32:11 --- join: Soap` (~flop@202-0-42-22.cable.paradise.net.nz) joined #forth 01:32:11 --- join: goshawk` (goshawk@panix1.panix.com) joined #forth 01:32:11 --- mode: bear.openprojects.net set +o ChanServ 01:32:12 gfx gos kewl, AI goes stupid.... 01:32:37 AFK 01:41:15 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 01:42:12 --- quit: Serg_penguin () 01:47:57 --- part: MrReach left #forth 02:05:42 --- quit: futhin ("sleep") 02:05:42 --- quit: Kitanin (Read error: 104 (Connection reset by peer)) 02:08:13 --- join: Serg_penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 02:09:26 --- quit: Serg_penguin (Client Quit) 02:44:07 --- join: Stepan (~stepan@Charybdis.suse.de) joined #forth 02:55:13 hey :-) 03:48:02 --- join: Kitanin (~clark@209.91.66.234) joined #forth 05:48:03 --- quit: Kitanin (Read error: 104 (Connection reset by peer)) 05:48:10 --- join: Kitanin (~clark@h-209-91-66-234.gen.cadvision.com) joined #forth 05:53:44 --- join: Serg_penguin (~Z@nat-ch1.nat.comex.ru) joined #forth 05:54:05 --- part: Serg_penguin left #forth 06:32:18 --- join: tathi (~josh@wsip68-15-54-54.ri.ri.cox.net) joined #forth 06:32:18 --- quit: Kitanin (Read error: 104 (Connection reset by peer)) 07:37:30 --- join: Serg_penguin (~Z@217.10.35.226) joined #forth 07:38:00 --- quit: Serg_penguin (Client Quit) 07:54:09 --- join: Kitanin (~clark@h-209-91-66-234.gen.cadvision.com) joined #forth 08:27:04 --- join: proteusguy (~irc@24-197-147-197.charterga.net) joined #forth 08:27:04 --- quit: Kitanin (Read error: 104 (Connection reset by peer)) 08:29:21 --- quit: proteusguy (Client Quit) 08:30:49 --- join: proteusguy (~irc@24-197-147-197.charterga.net) joined #forth 08:53:47 --- join: futhin (thin@h24-64-175-61.cg.shawcable.net) joined #forth 09:02:07 --- join: kc5tja (~kc5tja@ip68-8-206-226.sd.sd.cox.net) joined #forth 09:02:18 good morning kc5tja :) 09:02:23 uug 09:02:31 get a good 5 hours of sleep? 09:02:37 or 6.. :P 09:02:39 7 actually 09:02:42 wow 09:02:55 actually i got 6 1/2 09:03:19 so not that big of a diff 09:03:31 tired? 09:07:19 I'm always somewhat tired after first waking up. 09:09:21 ah 09:09:35 well i stayed in bed for a half an hour then i showered 09:09:44 so i wasn't so tired when i got on the comp 09:10:21 --- quit: proteusguy (Read error: 110 (Connection timed out)) 09:11:24 Well, I'm also getting ready for work. 09:14:39 --- quit: Stepan ("Do you think it is air you are breathing? Hmm?") 09:28:14 --- quit: futhin ("laters") 09:46:40 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 09:54:14 --- join: Kitanin (~clark@h-209-91-66-234.gen.cadvision.com) joined #forth 10:15:09 --- join: geakazoid (~JB@adsl-63-202-178-239.dsl.snfc21.pacbell.net) joined #forth 10:15:10 --- quit: Kitanin (Read error: 104 (Connection reset by peer)) 10:17:04 * geakazoid is gone, autoaway/10m (l!on) 10:17:55 hello 10:23:58 --- quit: geakazoid ("Leaving... ") 11:54:19 --- join: Kitanin (~clark@h-209-91-66-234.gen.cadvision.com) joined #forth 12:22:46 --- join: kc5tja (~kc5tja@user-24-214-86-42.knology.net) joined #forth 12:22:46 --- quit: Kitanin (Read error: 104 (Connection reset by peer)) 12:23:12 --- join: proteusguy (~scherrey@24-197-147-197.charterga.net) joined #forth 12:30:37 brb 12:30:41 --- nick: kc5tja -> kc-meeting 12:51:46 --- nick: kc-meeting -> kc5tja 13:08:35 --- join: proteus (~scherrey@24-197-147-197.charterga.net) joined #forth 13:08:55 --- quit: proteusguy (Read error: 110 (Connection timed out)) 13:12:22 --- join: MrAway (~mrreach@209.181.43.190) joined #forth 13:15:42 --- quit: tathi (Read error: 104 (Connection reset by peer)) 13:16:58 --- join: tathi (~josh@wsip68-15-54-54.ri.ri.cox.net) joined #forth 13:39:36 --- quit: kc5tja (Remote closed the connection) 13:39:51 --- join: tcn (tcn@tc4-login40.megatrondata.com) joined #forth 13:48:29 --- join: kc5tja (~kc5tja@user-24-214-86-42.knology.net) joined #forth 13:54:24 --- join: Kitanin (~clark@h-209-91-66-234.gen.cadvision.com) joined #forth 13:55:52 Hmm...more impetus to complete my DOS version of FS/Forth. A friend of mine wrote me to ask if I could complete it for the task of controlling a CNC milling machine. That'll be a fun project... :) 13:56:07 agreed 13:56:48 Dag Nabbit, you got all the cool stuff... _My_ friends just want me to sit in front of the computer playing Neverwinter Nights... 13:57:06 get different friends 13:57:14 :-) 13:57:27 there is a difference between people who think and explore and people who just live 13:57:27 kc: cool :) 13:57:39 Hey, I play NWN all the time... :) 13:58:07 i mean the milling machine :) 13:58:31 tcn: Hehe :) Apparently they want to use FS/Forth because they know it'll be maintained by me in the future, while PygmyForth is a dead-end product. 13:59:04 --- quit: tathi ("leaving") 14:00:47 but that means it's not a moving target :) 14:09:01 Well, while true, it also means that the author generally won't offer support for the product either. 14:10:34 --- quit: proteus (Read error: 110 (Connection timed out)) 14:10:58 --- join: proteus (~scherrey@24-197-147-197.charterga.net) joined #forth 14:11:55 Unless you impress them by asking about it fourteen years later, making him go digging through his files to see if _he_ has a copy... Yeah, that was fun. 14:12:20 Heheh :) 14:12:51 Of course, then I had to find a machine I could _install_ my Turbo Pascal compiler on... 14:46:49 --- part: proteus left #forth 14:46:49 --- quit: Kitanin (Read error: 104 (Connection reset by peer)) 14:47:52 --- join: tcn2 (tcn@tc1-login32.megatrondata.com) joined #forth 14:59:01 Heheh :) 14:59:13 --- join: proteusguy (~scherrey@24-197-147-197.charterga.net) joined #forth 15:00:01 --- quit: proteusguy (Remote closed the connection) 15:02:14 --- quit: tcn (Read error: 110 (Connection timed out)) 15:27:35 --- nick: tcn2 -> tcn 15:42:49 --- quit: tcn ("Leaving") 15:56:04 --- join: Kitanin (~clark@209.91.66.234) joined #forth 15:56:25 Nevermind... 15:57:51 ? 15:59:14 Oh, my connection to opn has been dropping every two hours for two days now. I commented that it hadn't yet, and it dropped. 15:59:42 [22:55] Odd... My IRC server's been crashing every two hours like clockwork for two days now. Until now. Why hasn't it crashed yet? 15:59:42 [22:55] * Disconnected 16:00:57 Heheh :D 16:01:23 Yeah, I experienced something I've *never* experienced before today -- BitchX segfaulted. For no apparent reason. 16:01:50 Well, you see, the crash was a minute late, and that confused me. :-D 16:04:20 Heheh :D 16:39:55 --- join: CrowKiller (Vapo_Rulez@cnq5-233.cablevision.qc.ca) joined #forth 16:39:56 --- quit: Kitanin (Read error: 104 (Connection reset by peer)) 17:04:42 --- join: Kitanin (~clark@h-209-91-66-234.gen.cadvision.com) joined #forth 17:04:44 --- part: Kitanin left #forth 17:24:32 --- quit: kc5tja (Remote closed the connection) 17:34:30 --- join: geakazoid (~JB@adsl-63-206-89-171.dsl.snfc21.pacbell.net) joined #forth 17:44:35 * geakazoid is gone, autoaway/10m (l!on) 17:50:49 --- join: TheBlueWizard (TheBlueWiz@216.25.205.169) joined #forth 17:50:56 hiya all 17:53:00 --- quit: Soap` (Read error: 104 (Connection reset by peer)) 17:53:50 hello 17:59:17 hiya geakazoid :) 18:07:37 --- join: kc5tja (~kc5tja@24.214.86.42) joined #forth 18:08:01 Just one more reason to complete FS/Forth and get it ported to Linux -- so I won't have to put up with segfaulting IRC clients. 18:10:00 hmm? 18:10:17 This is the third time that my IRC client has spontaneously segfaulted on me. 18:10:42 Something these IRC servers are broadcasting is causing my version of BitchX to die. 18:10:54 And it's never happened before. 18:10:59 --- join: Mongrel (~ant@ool-435249eb.dyn.optonline.net) joined #forth 18:11:22 I was trying to /ctcp version you, but it is completely blank.... 18:11:55 * TheBlueWizard has used BitchX; never experienced segfaulting 18:12:09 Neither have I, until today. 18:12:27 hmm....the plot thickens.... 18:15:04 --- join: Soap` (~flop@202.0.42.22) joined #forth 18:15:14 hiya Soap` 18:15:39 * kc5tja is writing the regression test software for his assembler, and so far everything checks out (of course; for as soon as I find an error, I correct it. :) ) 18:16:33 --- join: tathi (~josh@ip68-9-68-213.ri.ri.cox.net) joined #forth 18:16:34 :) 18:16:39 hiya tathi 18:16:46 hey hey 18:16:52 * CrowKiller found an interesting article: http://x86.ddj.com/productivity/nulldescriptor.htm 18:17:09 already read that years ago :) 18:17:36 'ello 18:19:34 hmm 18:19:43 x86 architecture is weird... 18:19:46 TheBlueWizard hi BTW, I am off configuring my ports so I can set our local http server 18:20:04 geakazoid: ok :) 18:21:40 It's not weird. 18:21:45 It's just plain backwards. 18:21:59 Segment tables are nothing new to microprocessors. 18:23:50 I just meant it seemed silly to have an entry in the descriptor table that the processor doesn't use... 18:24:43 Oh. 18:24:52 I thought that when I first encountered tha tmyself. 18:25:05 You kind of "grow used to it" after a while. :) 18:25:22 BTW, the author doesn't mention this, but the zeroth LDT entry is also unused. 18:25:30 yeah, I remember that 18:26:03 Unfortunately, that's a bigger waste yet, since you can't really use it for anything, except maybe for per-process variables. 18:26:05 when I first started doing PPC it seemed really different, since all I knew was x86 18:26:28 Man, I *wish* I could work with PowerPC and actually be able to justify it. 18:26:53 I still find MIPS to be my favorite RISC architecture though. 18:27:11 don't know anything about MIPS 18:27:32 kc5tja: if Pios ONE were delivered two years or so ago, I'd bet you'd buy that :) 18:27:58 No. 18:28:05 * TheBlueWizard doesn't know much about MIPS 18:28:09 By that time, the hard lessons of economics had already been learnt. 18:28:17 I see 18:28:37 TheBlueWizard: It's *very* much like PowerPC (or, more accurately, PowerPC is very much like MIPS). It's a brutally simple processor architecture that is very highly optimized to run C code. 18:28:55 I would *love* to support the PIOS ONE, but I just can't justify it to myself. 18:28:55 so you initially were on TeamONE for that machine, but quickly lost interest in that 18:29:03 Very quickly. 18:29:06 yeah, unfortunately PPC isn't terribly well suited to forth :) 18:29:25 tathi: why is that? 18:29:33 tathi: There are techniques one can use to map stack manipulations into CPU registers. 18:29:43 * kc5tja has done it in the past on the x86 and it works astonishingly well. 18:29:45 just more job at edit time 18:30:05 or compile time, run time can be compensated by correct "magical" register adressing schemes 18:31:10 CrowKiller: The PowerPC architecture is very heavily optimized to run applicative, imperative languages, not concatenative languages. As a result, without performing register allocation and scheduling like a C compiler would, you're not going to get good Forth performance. 18:31:31 is dice.com decent enough for job hunting? 18:31:40 * kc5tja doesn't know. 18:31:43 I would like to try coding for the ARM arch 18:31:47 I've seen commercials for it on TV though. 18:32:10 ARM is a nice architecture. I've looked at some of its assembly language, and man o' man, it's very clearly 6502 influenced. :) 18:32:21 kc5: I was thinking about trying to map stack stuff into registers, but decided to leave that for farther down the road though... 18:32:45 IIRC some of ARM creators are also 6502 designers 18:32:50 tathi: Likewise. I did a proof of concept implementation and it worked rather well. But I'm going to add it only when I really need it. 18:33:25 TheBlueWizard: Not sure about that, but I do know that ARM was designed largely by the folks at Acorn Computers in Europe to supercede the 6502. That's the story I've heard, at least. 18:33:27 well, that's nice to know 18:34:38 i know theres an asynchronous version of ARM platform named AMULET that uses even less power is is faster in some cases 18:34:47 Anton Ertl did a nice write-up on dynamic stack caching and static stack caching, and it turns out to have rather profound performance increases. :) However, it can be sticky too -- depending on how aggressively you perform this optimization, you could start to incur subroutine call overheads that rival those of C. 18:35:02 kc5tja: well, the ex-6502 engineers angle was what I heard... 18:35:48 I know Bill Mensch is responsible for the 6502, and runs Western Design Center in Arizona (still making 6502 and 65816 chips!). 18:36:13 However, it's possible some engineers who used to work under him went to work for Acorn later. 18:36:30 I don't know. It's all shrouded in mystery. 18:38:10 * TheBlueWizard knows WDC *still* makes 6502 and 65816 chips, and knows that a certain company used to make SuperCPU for C-64...a 20 MHz 65816 "toy" :) 18:38:21 Yup. 18:38:23 CMD 18:38:36 Are they still making those CPU cards for the C64/128? 18:38:39 * TheBlueWizard nods re: the fuzziness of details...it's same for him 18:39:46 hmm...I think it went out of business not long ago....y'know, C-64s get crapped out once in a while, and there are finite supply of those 18:40:32 * TheBlueWizard is googling for supercpu 18:43:13 Heheh Yeah. 18:43:32 I think it'd be neat to get a Commodore 128, with SuperCPU attachment, and make a *real* Forth-type OS for it. 18:43:45 Even go so far as to compete head-on with GEOS, complete with GUI. 18:44:14 hmm...it looks like it's still around after all! see http://www.cmdweb.de/scpu.htm (note that the domain is .de, not .com...interesting...the latter don't seem to sell SuperCPU at all) 18:45:47 I think it'd be even neater to build a variant of VICE with some flavor of "Forth CPU" in it, and replace the ROM to run a homebrew OS...but that's a lot of work, and I have other more important things to do...like talk ;) 18:46:45 I once had a fantasy dream 8-bit machine basically looking like C-64 but with 6809 CPU running inside :) 18:46:47 Heheh :) 18:47:10 * kc5tja wouldn't touch the 6809 with a 100-foot pole until Motorola drastically improved its bus timing characteristics. 18:47:44 Besides, the CPU as a whole just seemed way too complex for its britches. I mean, for an 8-bit CPU, it's *almost* as powerful as the 68000. And that's just not right. :) 18:47:48 hmm...maybe that's why Mot didn't sell 6809 so well, despite its neat architecture 18:48:29 Bill designed the original 6800, and noted that it was a complex chip. 18:48:54 He originally wanted to produce a 6500 with Motorola, but Motorola said "Go take a shit," though in not so many words. 18:49:05 I agree 6809 is pretty powerful for a 8-bit....ironically, its power almost begs to get more powerful (e.g. go 16-bit or whatever) 18:49:10 So he left, and went to work for MOS technologies (yes, the same company that Commodore owned) 18:49:22 That's when he took the design and radically simplified it, and released it as a 6500. 18:49:35 It was even pin-compatible with the 6800, and Motorola sued over that. 18:49:48 So they changed the pin out, and the 6500 became, yes, our beloved 6502. :-) 18:50:02 * TheBlueWizard nods, knew the tale...and knew that Mot sued the startup, and eventually lost :) 18:50:31 Yup. And Commodore made a mint because they reaped the profits from MOSTek while Apple and Atari jumped on the 6502 bandwagon. 18:51:37 I remember the story of Steve Woz keeping the early Apple II mobo design "generic" enough to accept either 6800 or 6502, hedging on the outcome of the trial :) 18:52:43 Commodore aggressively verticalize the computer manufacturing, which was simply fscking brilliant....and Commodore made a lot of money as a result 18:53:16 * kc5tja nods 18:53:32 Commodore was a force to be reckoned with, all the way up until the end. 18:54:06 It's stupid that they went bankrupt because Ali didn't want the Amiga to be popular. Fuck him and the goat he sleeps with. If I ever meet that bastard, I'll fuckin' shoot him. 18:54:27 of course in the later years (especially after key people like oh-so-controversial Jack Tramiel left Commodore), the company basically lost its way...and the CEO even drove it into the ground 18:54:32 pathetic 18:55:07 Man, I so desparately want to purchase the Commodore name, logo, and rights to the 8-bit machines. 18:55:22 I would love to get it going again, with more modern architectures. 18:55:56 yeah...Commodore had great engineers, and perhaps very good technical managers as well...but the higher ups like Irving Gould (CEO) and Ali Mehdi...whoa...they're so idiotic! 18:56:31 lol....but it seems that the name Commodore is now something of a curse 18:56:58 The only thing is, I don't want to hire anyone like Dave Haynie because I fear he'd pollute my designs (I know full well he's vehemently against Forth and stack architectures, for example). 18:57:45 And since competing with the PC is futile, that leaves only Apple to compete with. And if they can release a box and bundled OS together, I should be able to as well. 18:57:54 And unlike Apple, I'd treat the developers with respect. 18:58:18 In fact, like the Commodore 64, I'd publish **EVERYTHING** about the platform, and sell the programmers reference books in places like WaldenBooks and Borders. 18:58:25 (as well as the usual avenues) 18:58:29 While Dave certainly is a fantastic engineer, but he seems to exhibit a control freak over the design...guess that comes in a package :) 18:58:41 Yeah. :) 18:59:00 With him (control freak in favor of register designs) and me (control freak in favor of stack designs) working together, problems are bound to arise. 18:59:33 heh 18:59:45 the proof has been made in Stack Computers: the new wave 18:59:56 GOOD IDEA re: publishing everything...I really love that! Apple in its early days did that, Commodore did that, etc, and they flourished then....now with closed machines, it is hard to hack :( 19:00:00 CrowKiller: That's not a proof, that's a description. 19:00:02 stacks are better by design 19:00:12 The proof is in the P21/F21, etc. But those don't have *ANY* market share of any kind. 19:00:22 CrowKiller: Chuck will actually disagree with you. 19:00:39 CrowKiller: A processor should have both stacks AND some compliment of registers. 19:01:06 Yes, the stack is a vitally important design characteristic, but there are times when a register just plain is superior. 19:01:45 Otherwise, he would never have added the A register to his x21 designs, or the B register to the C18. 19:01:56 a register is a buiding block, stacks come after registers becuase they are built with them 19:02:17 but 2 stacks are clearly an advantage 19:02:20 C'mon -- we're talking about programmer visible first-class objects here. 19:02:30 no operand instructions are way better 19:02:39 Crow, are we arguing that? No. 19:02:47 even the best engineer would have to reckon that 19:02:50 im only saying 19:03:04 that you'd win 19:03:11 The best processor design will be a stack processor, but it'll also have registers too. 19:03:16 if the two of you were on teh same design team ;p 19:04:15 No, I wouldn't ever be on the same development team with him to begin with. 19:04:24 addressable regs for specific, optimized purposes, like counter for loop and address generation 19:04:43 Early binding -- know your faults and know the faults of your partners. If you know you won't work well together to begin with, DON'T BOTHER TRYING. :) 19:05:06 Even general purpose registers, for things like local variables and such. 19:05:18 Those can be very handy 19:05:29 Granted, they're not NEEDED.... 19:05:38 But if present, I'd certainly make use of them. 19:05:39 if it were possible to throw in AltiVec like stuff into Forth chip that goes into kc5tja's dream computer, it might have some *interesting* applications! but AltiVec is quite register oriented.... 19:06:05 TheBlueWizard: Vector operations are definitely quite possible with stack CPUs. 19:06:25 [instruction] 19:06:32 hmm...haven't seen that discussed 19:06:59 A vector math unit is just a coprocessor with a DMA engine -- like the Amiga's blitter. 19:07:13 (The Amiga's blitter was a logical vector processor, but it was a vector processor nonetheless) 19:07:53 * TheBlueWizard nods, interpreting "logical" to mean bitwise logical math 19:07:58 Yes. 19:08:18 there is another type of logic :) 19:08:48 Forth doesn't distinguish between boolean and bitwise, so I figure I don't have to either. 19:09:05 In fact, for the purposes of comparisons, neither does C. :D 19:09:09 brb 19:09:32 I am alluding to conditional logic 19:12:28 back 19:12:45 Well, that's basically a form of boolean logic, but short-circuited. 19:14:04 in a way, yeah 19:16:03 * CrowKiller never heard of "conditional logic" 19:16:18 Well, I haven't finished testing all the stuff in the assembler, but it looks like everything is being produced correctly. 19:16:33 cool 19:16:35 This makes me a very happy person, because it's good enough to start writing FS/Forth itself now. 19:16:49 yah 19:17:15 CrowKiller: IF you've never heard of logic THEN IF you've never heard of conditionals THEN ... ENDIF ENDIF 19:17:34 That's equivalent to IF ( you've never heard of logic ) && ( you've never heard of conditionals ) THEN ... ENDIF 19:17:41 i thought about "conditional lofic" in circuit manufacturing 19:17:50 is CrowKiller very C oriented? 19:17:55 No. :) 19:18:04 ah 19:18:12 * CrowKiller don't know C at all, barealy able to read it 19:18:47 * CrowKiller is trying to become a skilled asm programmer though ;p 19:18:52 * kc5tja considers writing a set of books: The Forth Apprentice, The Forth Journeyman, and The Forth Master. 19:19:18 The Forth Apprentice teaches ANSI standard Forth, but while the reader is going through it, he's making his own Forth implementation. 19:19:53 The Forth Journeyman deals with more complex issues, like proving that portability is a myth, and that well factored software is the *only* solution to dealing with portability issues. 19:20:13 Of course, it uses the Forth written in Forth Apprentice to do this. 19:20:13 :) 19:20:39 um...I think that book should have some philosophical exposition of why Forth is structured thusly 19:21:09 That won't work -- the reader will want proof. 19:21:26 They want hard examples. 19:21:35 even proofs eventually rests upon axioms :) 19:21:41 making his own forth system is quite a hard example ;ppp 19:21:42 The philosophy is of course important, but that's contained ithroughout the whole of the book. 19:21:49 in every sens of the word ;p 19:21:51 Axioms != philosophy 19:21:52 sense* 19:22:29 philosophy is in many ways a lot like using axioms...you assume one way, and you go from there 19:22:33 CrowKiller: I disagree with that. 19:23:13 in assembly trying to come up with the single best way of doing things require a lot of thought power for a complete newbie to programming in assembly like me 4 months ago lol 19:23:18 CrowKiller: Making your own simple Forth environment is actually very easy to do. 19:23:44 CrowKiller: Then you're biting off more than you can chew, which is something I said 4 months ago, if you recall. 19:23:49 Forget the optimization. 19:23:56 Just get something *working*. Optimize later. 19:23:59 once you know what you want, yes its quite easy, the more you know it the easiest it is 19:24:03 And if necessary, discard and start over. 19:24:44 i actually discard a lot of ideas as they come 19:24:46 I respectfully disagree with your position. The things you're telling me are obvious, and aren't even worth mentioning. 19:25:08 No, you don't necessarily discard ideas "as they come." You sometimes have to discard the whole product after it's finished. 19:25:35 i'm saying that i dont code something right the moment I think of it, i try to weight if its woth it 19:25:38 worth* 19:25:54 Of course!!!!! 19:25:57 Who wouldn't?! 19:26:04 Again, that's obvious. 19:26:08 But you can't predict everything. 19:26:32 Only hard measurements (either in actual use or profiling, etc) can ultimately decide whether your software meets your requirements. 19:26:55 That's why my Forth is going to be dog slow to start with. *DOG* slow. 19:26:59 I know that, going into it. 19:27:12 Why? Because I want to have a baseline on which I can compare alternative structures with. 19:27:29 Does a skiplist dictionary outperform a ternary tree, and if so, by how much? 19:27:49 Is it even *significant* compared to the simple implementation of a linked list? 19:28:09 I can't make these determinations without actually building multiple versions of FS/Forth and actually putting them to a stopwatch. 19:28:40 There's no point in trying to achieve perfection the first try. NONE! You'll only end up with the worst possible product if you do. 19:29:02 hear hear 19:29:46 Another example: VIBE. It's a VI-like block editor. Why'd I write it? Because it provides a superior user experience for me. Pygmy's block editor is woefully inadequate for my needs. It's not expandable easily, and it's actually pretty hard to use. 19:30:01 I came to this conclusion after using it since 1988. 19:30:17 That's a different type of "measurement," but is just as valid as a stopwatch. 19:30:30 "How many more lines of code per minute can I write using VI, versus Pygmy?" 19:30:52 And funny thing is, because of VIBE's unique interaction with the surrounding Forth environment, it's smaller! 19:31:05 Again, simplicity rules over the complex. 19:31:45 And I made absolutely no attempt to optimize that either. No need to -- the time it takes from keypress to computer response is imperceptible to me. That's "good enough" no matter what. 19:32:38 I have to take a leave of absence... see ya'll later... maybe at the tomorrow's UK Forth IRC chat 19:32:47 bye geakazoid 19:32:51 Laters geakazoid :) 19:32:57 --- quit: geakazoid ("Leaving... ") 19:33:10 * kc5tja should probably attend one of those UK Forth IRC chats. 19:33:15 Anyone know if forthOS (forthos.org) is dead? 19:34:53 Well, the webpage is still up. 19:35:03 The latest version appears to be ForthOS07. 19:35:12 Not sure if that means anything to you... :/ 19:35:37 Webpage was last modified in febuary :/ 19:35:45 What year? 19:35:58 This year. 19:36:17 I'd think it's still live then. 19:36:20 alive even 19:36:33 I remember seeing that existance of that page at least for several years now. 19:36:52 Maybe not necessarily as active as it once was, but... :) 19:37:56 Maybe I'm being overly paranoid 19:38:00 None of the standalone forth-os projects ever seem to live long enough to make it to the stage where it's easy for a newbie like me to extend them :| 19:39:53 Well, extension requires that you know your way around the system. 19:39:58 That means it needs documentation. 19:40:09 And especially for the inner-most parts of the system, lots of it. :) 19:40:49 * TheBlueWizard nods 19:41:04 I can't promise I'll provide a lot of useful documentation for my FS/Forth system when it becomes available on bare metal, but when it does, I will have at least a Wiki set up to enable people to ask questions and I can provide answers. 19:41:18 The Wiki will help ensure I have a pseudo-permanent repository of information thus accrued. 19:42:16 It's sad that it takes so much effort to properly document a system. Even a basic Forth system can be quite complex to document if the user of the system doesn't understand how to read the Forth sources. 19:42:29 That's why part of my documentation will be devoted to my coding conventions. 19:42:32 The systems I've tried aren't as interactive as a forth system should be... if I tweak something in forthos, I have to recompile it in linux, reboot, test, etc. 19:42:44 I just want something thats... I think the phrase is "self-hosting" 19:42:51 Ah...that is where my Forth absolutely will be different. 19:43:00 It will be self-hosting. 19:43:06 Coo' 19:43:31 The DOS version of FS/Forth was originally designed to boot-strap the 32-bit version of FS/Forth on my laptop. 19:43:47 (But, now, it's also commissioned for use in controlling CNC milling equipment. Cool!) 19:45:47 Well, I think I'm going to take this opportunity to leave work and head for the hills ... specifically, my home. 19:46:03 bye kc5tja 19:46:11 see you 19:46:17 bye 19:46:53 I'm off too...to bed, so I can get an early start for Maine in the morning 19:47:24 bye tathi 19:47:29 good night 19:47:51 --- quit: tathi ("laters all...") 19:47:54 I'll be back online in about 30 to 45 minutes, traffic permitting. 19:48:01 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 20:44:15 --- join: kc5tja (~kc5tja@ip68-8-206-226.sd.sd.cox.net) joined #forth 20:44:42 re 20:44:54 Howdy -- just got in. 20:45:06 Well this sucks. It looks like the floppy drive in my computer is shot. :( 20:45:12 almost an hour offline 20:45:20 hiya kc5tja again 20:45:27 yeah me too its completely mislaligned and output false data 20:45:53 But I never *use* the floppy except in cases when I need to transfer data between my laptop and this box. 20:45:56 * kc5tja sighs 20:46:29 I think it's about time I create the optical link between these two machines like I've always said I would... 20:46:39 all my floppy drives here are fine.... 20:47:00 tho' I should start building networking system :) 20:47:18 does your laptop have a IR port? 20:47:20 No 20:48:32 * kc5tja would have to create his own optical adapters using the parallel port. 20:49:28 But that's OK, because I don't like any existing protocol in existance for optical links. 20:50:05 good luck with that project...maybe hifn would branch out into selling your oplink scheme :) 20:50:45 well, gotta go...bye all 20:51:03 bye TBW 20:51:07 bye CrowKiller 20:51:21 --- part: TheBlueWizard left #forth 20:51:37 Hifn wouldn't care less. 21:42:00 --- join: sbk_ (~kbs@dsl-65-184-98-221.telocity.com) joined #forth 21:57:59 --- quit: sbk_ ("Leaving") 22:04:23 --- quit: CrowKiller ("User pushed the X - because it's Xtra, baby") 22:49:43 --- join: XeF4 (~Kayaman@dyn030.wlan.ASMPARTY.net) joined #forth 22:53:29 --- quit: XeF4 (Client Quit) 23:43:13 --- quit: kc5tja ("THX QSO ES 73 DE KC5TJA/6 CL ES QRT AR SK") 23:59:59 --- log: ended forth/02.08.02