00:00:00 --- log: started forth/04.10.13 00:01:29 --- quit: ez4 (Remote closed the connection) 00:28:08 --- join: fridge (~Jim@CommSecureAustPtyLtd.sb1.optus.net.au) joined #forth 00:31:10 hi 00:31:42 hi 00:31:53 main server HDD glitch ;(( 00:31:58 --- part: Serg_Penguin left #forth 00:34:22 our balance loader had a hdd glitch 00:34:42 and client is having an unwarranted sook 00:34:52 it would've been better if we hadn't said anything 01:00:43 --- quit: mur_ (Remote closed the connection) 01:02:26 --- join: mur (~mur@smtp.uiah.fi) joined #forth 01:20:29 --- join: colorg (r@core-dc-1-55.dynamic-dialup.coretel.net) joined #forth 01:20:42 anybody up? 01:24:20 --- quit: skylan (Remote closed the connection) 01:24:24 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 01:41:40 --- quit: skylan (Remote closed the connection) 01:41:43 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 01:47:50 --- quit: skylan (Remote closed the connection) 01:47:53 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 01:52:02 --- join: oyd (~Miranda@80.178.218.205.forward.012.net.il) joined #forth 01:57:45 --- join: ez4 (~ez4@pcp01518726pcs.reding01.pa.comcast.net) joined #forth 01:57:56 anybody up? 01:58:12 i am 01:58:34 Good morning. 01:58:41 morning 02:00:04 --- quit: skylan (Remote closed the connection) 02:00:07 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 02:00:24 ez4 sounds like a Forth 02:01:04 actually it's my initials 02:02:49 i'm debugging eforth in osimplay. I posted it to clf yeserday 02:05:29 I seem to bomb in SP! 02:05:30 clf? 02:05:36 comp.lang.forth 02:05:57 what do you mean debugging in osimplay? 02:07:00 osimplay is an all-shell-script assembler. I'm porting eforth from Linux gas, and only a few words work. It doesn't initialize yet. 02:08:08 looks interesting 02:08:13 :o) 02:10:56 i'm on powerpc 02:11:24 What OS? 02:11:41 osx and linux 02:11:50 What 02:11:57 's the shell in SOX? 02:12:02 OSX? 02:12:02 mac os x 02:12:21 bash 02:12:27 MacOS over X, or a shell? 02:12:29 ah 02:12:53 ? 02:12:58 i run mac os x native 02:13:10 and linux native 02:13:19 No X? 02:13:40 or no MacOS GUI but with X? 02:14:00 what? 02:14:08 mac os x has a proprietary display server 02:14:33 it can also run X windows 02:15:06 does the in-house stuff use bash, or is it sans CLI like MacOS? 02:16:37 what in house stuff? mac os x is a mostly posix compatible unix. bash is the default shell. 02:17:17 that answers my question 02:17:39 It's not MacOS even at the UI 02:18:16 you mean mac os classic, so to speak? 02:18:22 yeah. 02:18:39 classic is pretty much dead 02:19:10 they ported the old APIs though 02:19:15 Chatting with you seems beneficial. I'm definitely bombing in SP!, which knowledge is progress 02:19:20 but everything else is different 02:20:33 I suspect there's a useful subset of x86|PPC that would result in a "portable" osimplay. 02:20:57 x86 addressing modes are basically ridiculous 02:21:14 and it has less registers than most comparable chips 02:22:01 i know nothing about x86. i only know risc chips 02:22:09 So with fewer addressing modes and the same register names you have portable assembly. 02:22:41 wow, a language with the amount of addressing modes of a risc chip with the register set of x86.. worst of all worlds, huh? 02:22:44 because it's the least common denominator... 02:22:50 x86 will do copy literal_offest + register + register to register 02:23:11 oxygene: kinda like C without the ALGOL 02:23:18 :o) 02:23:43 oxygene? 02:23:44 and the recursive toolchain dependancies 02:26:18 osimplay took about 2 months to write. No flats, no MMX 02:26:24 flats/floats 02:27:26 It's an easy port. The machine parts are well-factored. 02:28:44 i'm trying to learn isforth right now because i don't know forth at all 02:29:28 Forth is a wonderful example of technical elegance. 02:30:02 osimplay is heavily Forth-inspired. 02:30:25 Learning Forth is time well spent. 02:35:20 copy (INTeL mov) is = in osimplay because ! conflicts with the shell :o) 02:36:39 how is osimplay forth inspired? 02:37:50 I'm still trying to figure out what to do wrt assembler and portability in openbios/beginagain 02:37:57 The use of the cell concept ala the Forth standard and BCPL is huge on x86 02:38:27 oxygene: macros. 02:39:30 colorg: hm? 02:39:32 the x86 is a 16 bit fake machine and a 32 bit real machine, and in osimplay you switch between them by setting $cell to 2 or 4. Finito. 02:40:15 oxygene: begin is a label, again is a loop macro, non? 02:41:33 i see 02:41:38 colorg: beginagain is the name of the forth system we built for openbios - it's the attempt #\infty - 1 to get a portable forth system suitable for firmware up and running (with the difference to the attempts before that this one worked) 02:42:55 oh. It's in C? 02:43:01 yes 02:43:35 powerpc is a 64 bit machine with a 32 bit mode. i think it would be really cool to have a forth for 64 bit powerpc. it would simplify the math verbs so much not to have any numbers that take up two cells 02:43:37 6kb object code on x86 for the minimum configuration (without the grub filesystem drivers etc) 02:43:47 the FCode evaluator and device tree implementation is all forth on top of that 02:43:49 Forth is gobs more performant, and IMO easier to write, in asm. It's the wrongest thing there is to write in C. Voice of experience. 02:44:52 colorg: just that beginagain runs on x86, amd64, ppc now, can use foreign code (the grub drivers) on all platforms and is minimum hassle - firmware doesn't need best performance possible 02:44:55 FCode. Open Boot influence. That suggests to me it's not truly extensible. 02:45:36 what do you mean with that? 02:46:01 every word in eforth, a FIGFORTH descendant I believe, has a code entry. 02:46:31 I suspect your primitives aren't native code. 02:47:27 A real Forth inner interpreter has non-token primitives. 02:47:40 the fcode token table is built _in_ forth 02:47:56 --- quit: skylan (Remote closed the connection) 02:47:59 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 02:48:24 the fcode parser is written in forth and compiles fcode to beginagain's vocabulary 02:48:44 how does @ work? 02:48:47 the C code has no clue that there is a thing called fcode 02:49:19 This is @ in eforth.... 02:49:27 --- nick: colorg -> colorg_ 02:49:29 the CFA points to a dispatch table which points to a C function - by using that dispatch table the dictionary can be used on any machine with same endianess/cellsize configuration 02:49:33 --- nick: colorg_ -> \\ 02:50:20 --- nick: \\ -> colorg 02:50:58 __CODE 1 "@" AT 02:50:59 hike_xtp 02:50:59 = @ B to B 02:50:59 next_ 02:51:18 one machine insn 02:51:27 oh, three actually. 02:52:06 so RISC addressing and x86 registers might not be so bad ;o) 02:52:13 they're quite some more for us - we might try to convince the compiler to get rid of all that stack building crap, but it's not a priority 02:53:11 and = @ B to B is RISC/x86 portable I believe, except for maybe pure RISC like maybe SPARC 02:53:37 in which case it's a macro for two insns 02:56:02 ftp://linux01.gwdg.de/pub/cLIeNUX/interim/Forreal,tgz is a 1200 bytes osimplay thing that boots on the PC and sets up pmode and twiddles some VGA charcells as interrupts come in 02:56:26 er, Forreal.tgz, not a comma 03:02:56 switching to pmode and various other early init things (that you very likely won't do) is done by freebios2 in our case 03:05:19 Forreal.tgz has a lot of stuff in it about calling BIOS ROM from various modes. 03:05:55 Forreal Mode is mostly-32-bit unprotected. 16 bit interrupts. 03:06:53 if I recall correctly. I discovered it, but didn't do much with it. Could be fun as a BIOS. 03:10:58 and that kind of work was made possible by osimplay :o) 03:12:12 won't be of much use for us, I think 03:13:53 I guess. What it would really be good for would be hardcore realtime maybe. Or Freedos perhaps. 03:14:22 Forreal Mode I mean. I quite prefer osimplay to C. 03:16:01 and I prefer to only have to write drivers for a port and rely on the work of others for the other issues :) 03:17:22 what is a deferred word? 03:19:03 A word that compiles when the word it's in runs I think 03:19:15 a word that points to another - that way you can define words as existant, use them and define their behaviour later 03:19:40 oh yeah, I was waay wrong 03:21:52 What's the default stack size for a process on Linux these days? About 8 meg or so isn't it? 03:21:56 thanks 03:22:22 ez4: I was thinking of a postponed word. 03:23:32 I'm having trouble with eforth because it hand-manages the machine stack, which I don't want it to do on Linux 03:24:17 here's a really easy one. what does "if" do? 03:25:33 __COLON $((immedbit+2)) "if" IFF 03:25:33 aq COMPI ZBRANCH HERE 03:25:33 aq DOLIT 0 COMMA EXIT 03:25:36 :o) 03:26:03 immediate word, stores, at compile time, the current location on the stack, compiles a conditional branch.. else and then later fill out the offset of that branch 03:26:40 ez4: branch if TOS non-zero 03:28:20 so it looks like "if then "? 03:29:27 --- join: mur_ (~mur@mgw2.uiah.fi) joined #forth 03:29:34 if else then I think 03:29:48 my forth just had goto and labels. 03:30:11 the construct is done at THEN 03:30:11 oh, i see 03:30:28 --- quit: skylan (Remote closed the connection) 03:30:30 THEN should be ANYWAAY :o) 03:30:32 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 03:31:01 yeah, it really messes up people coming from other languages 03:31:03 : ANYWAY THEN ; \\ ;o) 03:31:32 ez4: that's why I stick to goto and label. It's unambiguous. 03:31:55 Good factoring is what structured programming is trying to be. 03:33:00 That's why I don't do flow control macros in osimplay, except for execution arrays, which are a bitch by hand. 03:33:36 Like an array of functions in C. VERY snappy for some things. 03:34:15 cells funcarray + execute? 03:35:28 An array of code addresses and an INDEXED jump or call into the array. 03:35:44 Real handy for parsing ASCII, for example. 03:36:18 The osimplay words to set one up are xray, beam and yarx. 03:36:42 It's on about page 223 in K&R2 :o) 03:36:57 Julian Noble likes them in Forth. 03:37:46 The original C Forth by Mitch Bradley, the guy that wrote Open Boot, was one big function array. 03:39:57 --- quit: mur (Read error: 110 (Connection timed out)) 03:42:57 There's also xjump in osimplay which is the requisite indexed jump 03:44:28 --- quit: skylan (Remote closed the connection) 03:44:32 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 03:50:19 The PDP-8 had about as many branch types as the x86. Like 16 or so. Ken Thomson used most of them in the original hand-coded ed. Maybe the scariest code ever. 03:53:25 I'm spent. The FORTH be with yooooou 03:53:28 --- quit: colorg ("ircII EPIC4pre2 cLIeNUX. Can you say that?") 03:56:31 --- join: crc (crc@75-pool1.ras11.nynyc-t.alerondial.net) joined #forth 04:29:42 --- quit: ez4 (Remote closed the connection) 04:41:21 --- quit: mur_ (Remote closed the connection) 04:43:03 --- join: mur (~mur@mgw2.uiah.fi) joined #forth 04:43:28 --- quit: onetom (zelazny.freenode.net irc.freenode.net) 04:44:29 --- join: onetom (~tom@novtan.bio.u-szeged.hu) joined #forth 04:52:08 --- quit: oyd (Read error: 110 (Connection timed out)) 05:13:39 --- quit: Raystm2 (Read error: 60 (Operation timed out)) 05:25:09 --- quit: skylan (Remote closed the connection) 05:25:13 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 05:39:29 --- quit: skylan (Remote closed the connection) 05:39:32 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 05:45:24 --- join: oyd (~Miranda@80.178.218.205.forward.012.net.il) joined #forth 05:47:40 --- join: Raystm2 (Rastm2@AC8EC824.ipt.aol.com) joined #forth 05:49:41 --- quit: skylan (Remote closed the connection) 05:49:45 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 05:57:43 --- quit: skylan (Remote closed the connection) 05:57:46 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 05:58:28 --- join: Herkamire (~jason@h000094d30ba2.ne.client2.attbi.com) joined #forth 05:58:28 --- mode: ChanServ set +o Herkamire 06:03:40 --- join: mur_ (~mur@kyberias.uiah.fi) joined #forth 06:14:06 --- part: Raystm2 left #forth 06:14:52 --- quit: mur (Read error: 110 (Connection timed out)) 06:23:30 --- join: Raystm2 (Rastm2@AC875A97.ipt.aol.com) joined #forth 06:32:27 --- quit: fridge (Remote closed the connection) 06:40:32 --- join: Serg_penguin (~Miranda@212.34.52.140) joined #forth 06:49:07 --- quit: mur_ (Remote closed the connection) 06:49:43 --- join: mur (~mur@kyberias.uiah.fi) joined #forth 06:51:10 --- join: Robert (~pink@c-bf5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 06:51:40 --- quit: skylan (Remote closed the connection) 06:51:47 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 07:02:37 --- quit: crc (Client Quit) 07:06:13 --- join: madwork__ (~madgarden@derby.metrics.com) joined #forth 07:09:04 --- quit: Serg_penguin (Read error: 104 (Connection reset by peer)) 07:29:32 --- join: qFox (C00K13S@82-169-140-229-mx.xdsl.tiscali.nl) joined #forth 07:44:09 --- join: Stepan (~Stepan@khepri.openbios.org) joined #forth 07:44:13 hi there 07:44:36 Hi Stepan, long time no see 07:46:04 Hey Robert :-) 07:46:05 indeed. 07:46:37 changed my job, moved to the other end of germany and working in the embedded area now 07:50:04 Cool :) 07:50:18 * Robert is still in the same room, using the same computer. 07:56:53 hey Robert how are ya. 07:57:38 Hi Ray 07:57:41 Well, I'm OK 07:57:47 Been playing with Perl lately 07:58:06 rj_cf is teaching me colorForth jumptables and arrays :) 07:58:23 in #c4th 07:59:01 what are ya coding in Perl ? 07:59:38 --- join: madgarden_ (~madgarden@Kitchener-HSE-ppp3576522.sympatico.ca) joined #forth 07:59:39 A program to analyze IRC logs. 08:00:21 what about a program to analize IRC loggers ? :) 08:01:04 That would be easy, I just need to sit down on a pole. 08:02:18 --- join: mur_ (~mur@smtp.uiah.fi) joined #forth 08:02:22 he he 'anal ize irc log' 08:03:13 --- quit: madgarden (Read error: 104 (Connection reset by peer)) 08:05:10 --- quit: mur (Read error: 60 (Operation timed out)) 08:15:09 --- quit: skylan (Remote closed the connection) 08:15:13 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 08:19:17 --- quit: mur_ (Remote closed the connection) 08:20:19 --- join: mur (~mur@smtp.uiah.fi) joined #forth 08:21:19 --- quit: skylan (Remote closed the connection) 08:21:22 --- join: skylan (~sjh@vickesh01-4738.tbaytel.net) joined #forth 08:37:16 --- join: das (das@adsl-64-219-100-35.dsl.lgvwtx.swbell.net) joined #forth 08:37:59 --- quit: das (Client Quit) 08:46:54 --- quit: skylan (Read error: 101 (Network is unreachable)) 08:48:07 --- quit: tgunr (Read error: 104 (Connection reset by peer)) 08:57:01 --- join: tgunr (~davec@vsat-148-63-4-108.c001.g4.mrt.starband.net) joined #forth 09:09:33 --- join: wossname (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 09:12:03 --- join: arke (apache@11.198.216.81.dre.siw.siwnet.net) joined #forth 09:44:44 --- quit: Robert ("BAI") 09:50:14 --- quit: tgunr (Read error: 104 (Connection reset by peer)) 09:52:15 --- join: tgunr (~davec@vsat-148-63-4-108.c001.g4.mrt.starband.net) joined #forth 09:57:03 --- join: warp0x00 (~warpzero@mi238.dn185.umontana.edu) joined #forth 10:05:32 --- quit: arke ("CGI:IRC (EOF)") 10:07:09 --- quit: tgunr (Read error: 104 (Connection reset by peer)) 10:07:24 --- join: mur_ (~mur@mgw2.uiah.fi) joined #forth 10:08:15 --- join: tgunr (~davec@vsat-148-63-4-108.c001.g4.mrt.starband.net) joined #forth 10:18:54 --- quit: mur (Read error: 110 (Connection timed out)) 10:31:15 --- quit: tgunr (Read error: 54 (Connection reset by peer)) 10:32:43 --- join: tgunr (~davec@vsat-148-63-4-108.c001.g4.mrt.starband.net) joined #forth 10:49:25 --- join: jrogers (~jrogers@h228.40.40.69.ip.alltel.net) joined #forth 10:53:04 --- quit: jrogers () 10:53:25 --- join: jrogers (~Olorin@h228.40.40.69.ip.alltel.net) joined #forth 10:54:11 --- nick: jrogers -> Olorin 10:54:46 --- nick: Olorin -> zaphod 10:55:14 --- nick: zaphod -> wandelf 10:57:36 --- nick: wandelf -> wandelf_ 10:58:42 --- nick: wandelf_ -> wandelf 11:00:03 --- quit: wandelf () 11:02:18 --- join: wandelf (~Olorin@h228.40.40.69.ip.alltel.net) joined #forth 11:03:08 --- join: Robert (~pink@c-bf5a71d5.17-1-64736c10.cust.bredbandsbolaget.se) joined #forth 11:03:25 everyone seems to be coming or going. 11:04:25 If you don't like it I guess I'll talk! 11:06:35 --- quit: wandelf () 11:07:18 hehe 11:12:00 --- quit: tgunr (Read error: 104 (Connection reset by peer)) 11:13:10 --- join: wandelf (~Olorin@h228.40.40.69.ip.alltel.net) joined #forth 11:13:13 --- join: tgunr (~davec@host-66-81-158-57.rev.o1.com) joined #forth 11:16:38 hrm. just had a fire drill. 11:16:54 Heh, OK. Are you at work? 11:16:57 yea 11:17:01 i'm always at work. :P 11:17:19 : 11:17:21 :) 11:17:34 i've been thinking of doing a forth assmebler in gforth and then using that to metacompile my own crappy forth. 11:18:16 Is it a Linux Forth? 11:18:28 i'd rather use ppc or arm. but x86 is easier to get to. Unless I put my gumstix(ARM) on a network so I can access it from work during my lunch break. 11:18:48 yea. i think i'll do it on linux first. self hosting would be loads of fun though. 11:18:59 What is that "gumstix"? 11:19:13 like I could put it on one of those Mini-ITX systems with a compactflash card and have a really sweet little forth workstation. :) 11:19:45 http://www.gumstix.com/ .. it's a 400Mhz Xscale/StrongARM on a tiny board with two serial ports. USB slave. 64Mb RAM, 4Mb flash. and MMC slot (i have a 256Mb MMC card in it) 11:20:14 $110 to $350 depending on what model and features you want with the gumstix. 11:20:31 it has bluetooth too. but the BT is not reliable in linux at this time. 11:20:54 Cool. 11:21:06 I'm writing a Forth for another one-board computer. 11:21:12 oh? what one? 11:22:00 AVR experiment board. ;) Featuring 3.7MHz 8-bit AVR microcontoller with 1/2 kB of RAM, 4k program memory words, 256kB flash chip, serial port... 11:22:04 i haven't writen a good forth yet. i've written really broken ones for VMs that ran in C. but it's looking easier to just write one in asm or a metacompiler so I'm going to do that instead and hopefully get my first real forth 11:22:09 So yours is, uhm, slightly better! 11:22:19 mine sucks 250mA :( 11:22:28 From 12V? 11:22:38 from 3.6V - 5V 11:22:44 Ah. 11:22:54 You're speaking of the processor now? 11:22:59 I think the AVR uses ~3mA. 11:23:10 I run it on 5V, but I think it will survive on 3V. 11:23:14 the whole board. sdram+cpu+etc 11:23:19 Ah, I see. 11:23:24 yea. AVR is really low power. 11:23:34 I was playing with MSP430 a while back. that's super low power. 11:23:46 Oh, like what? 11:23:47 and it can directly drive LCD segments. 11:24:01 like 12uA in a useful semi-sleeping mode :) 11:25:34 basically it has a mode where on every interrupt it runs a few instructions and goes to sleep when your routine is done. usually you put a 32.768kHz crystal used in watches to do this. when the system is "on" it can run at 8Mhz 16-bit or faster. but it's only on for a tiny amount of time. 11:26:08 i still have one running off a lithium battery driving an LCD. it should last about 3 or 4 years on one of those batteries. 11:26:28 it's also sampling a thermistor displaying the temprature and polling a few buttons. 11:27:09 it's doing multiplication endlessly for 3 years to scale the thermistor to celcius :) 11:27:40 Heh, cool! 11:28:12 msp430 is pdp-11 ish instruction set. kicks butt for both C and forth. (since all the general purpose registers can be used as stack pointers in a single instruction) 11:28:49 i haven't got to program it though. the clock/therometer was what came preinstalled on it. 11:29:06 I built a brainfuck computer from a downclocked PIC, it should use ~10uA. Of course the external pulldowns and LEDs will use some more, though :) 11:29:19 i wish it had an external memory bus so i could interface it to a NTSC ram dac or something. 11:29:29 hehe. neat 11:30:11 you should put buttons on it so you can input brainfuck on the fly. you could have a very tiny and cheap calculator that is extremely powerful and impossible to use. 11:30:38 Heh, I admit that would be neat. 11:30:40 you could calculate sin in brainfuck, but ouch! 11:30:45 I think I'd need a better display though. 11:30:51 A few LED segments or so. 11:30:53 those 16x1 displays are really cheap. 11:30:55 Maybe an LCD? 11:30:58 How much? 11:31:08 like $7 tops at the surplus places. 11:31:23 I paid $6 at halted for a 40x2 .. that's a little big though. 11:31:36 Wow. :) 11:31:47 That could be the main screen for a Forth computer. ;) 11:32:15 super tiny graphic displays for $40 or so. http://www.crystalfontz.com/products/index-grph.html 11:32:27 Ouch, $40 11:32:31 yea. expensive. 11:32:47 but the display is like a little bigger than a quarter for 128x64 11:33:11 Heh. 11:33:32 the 16x1, 20x2, 20x4, 40x2 character displays are cheap and only use 6 to 12 wires 11:33:41 How do you interface those? 11:33:55 * Robert has never used an LCD. 11:34:27 they have a 4 or 8 bit interface. you just set your data pins. hit a latch and that loads a byte or nybble into it and you just send simple commands to it. 11:35:04 you can actually play with one if you get a couple of debounced switched tied to the pins. the protocol is super simple. 11:36:06 http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=26206&item=3842392227&tc=photo .. $4 for a surplus 16x1 :) 11:37:35 you'll want anything that has a Hitachi HD44780 on it. here is a short guide to wiring and protocol. http://www.doc.ic.ac.uk/~ih/doc/lcd/operatio.html 11:39:01 man displays to choose from. you'll probably spend more on shipping than the stupid display costs. :) http://www.allelectronics.com/cgi-bin/category.cgi?category=365&type=store 11:39:37 woo.. they have naked LCDs there for $0.20 .. the MSP430 can interface those directly :) 11:39:51 that's a lot of pins to wire up though. :( 11:41:14 --- join: Baughn (~svein@cloud.brage.info) joined #forth 11:41:30 Thanks. 11:42:03 hrm. allelectronics has specs for all these. that's handy. i might order that 640x480 graphic display for my gumstix. 11:43:25 btw, what kind of power consumption do these small character LCDs have? 11:44:29 it's mediocre 11:44:43 especially if you use the backlight 11:44:53 Hehe, but without? 11:45:10 i don't have the numbers. but it's not that great. 11:45:45 you could run one on a set of AAs for a very long time. especially if you have it turn off the display when it's not in use. 11:46:00 but you won't be running it off a solar cell or anything cool like that. 11:46:10 (the msp430 kit I have can be ran off two potatos) 11:46:57 Hehe. :) 11:47:58 --- quit: wossname (Read error: 60 (Operation timed out)) 11:48:40 --- join: wossname (~wossname@rn-v1w5a06.uwaterloo.ca) joined #forth 11:51:06 wow. contiki runs on the msp430. http://www.sics.se/~adam/contiki/ports/ .. very tiny desktop because the msp430 doesn't have much ram. :P 11:53:37 Heh, incredible 11:53:51 AVR too. 11:54:16 C64..Atari... 11:55:08 *Anything* can be run off two sufficiently big potatoes. 11:55:38 Heh. 11:55:40 I have experience - I once ran a TV off a bunch of potatoes. 11:56:08 That's why you ended up in the #1 IRC nuthouse. 11:56:36 Actually, #nuthouse is empty at the moment. :{ 11:57:06 how big a potato would you need to run a desktop PC? 11:57:32 Bigger potatoes last longer, but you probably need more of them to supply sufficient amperage. 11:57:37 OrngeTide: could you easily hook one of these to a gumstix? http://www.allelectronics.com/cgi-bin/category.cgi?category=365&item=LCD-104&type=store 11:58:12 Given that I ran a 50W TV off fifty... three hundred? 11:58:21 The potato-per-watt model usually works. 11:58:45 how long did those fifty last? 11:58:55 ~4 minutes 11:59:07 heh 11:59:16 I'd doped the potatoes, though; don't know how much that helped/hurt. 11:59:34 LOL 12:00:05 hrm.. $50 ARM7 on a dip would be a pretty roomy single chip computer for forth. http://olimex.com/dev/lpc-h40.html 12:00:06 * Baughn hands Herkamire a handy roof-attached rope. 12:00:34 Herkamire, yea. that LCD should work. 12:00:57 OrngeTide: and how much are the gumstix? 12:01:10 Herkamire, the intel PXA255 is pretty flexible with it's onboard LCD controller. it's mostly a matter of how weird the memory layout is going to be to the Linux framebuffer :) 12:01:31 Herkamire, $140 gets you a 200Mhz gumstix with a case. 12:01:49 you have to get this itty bitty 60 pin connector to hook an LCD up to it though. have fun soldering that. :) 12:02:07 oh boy 12:02:15 gumstix currently doesn't have ethernet. everyone is using rs232 or usb slave. 12:02:40 maaan 12:02:44 I really want ethernet 12:02:45 Wait. You have USB, but no ethernet? 12:02:49 * Baughn boggles. 12:02:56 Baughn, it's usb slave. so you can only plug it into a PC 12:03:07 I've *read* the USB specs. 12:03:11 they will be doing a compactflash version very very soon. 12:03:33 --- join: futhin (thin@bespin.org) joined #forth 12:03:33 --- mode: ChanServ set +o futhin 12:03:33 OrngeTide: it doesn't have flash? 12:03:38 what does it do for storage? 12:03:40 Baughn, some people mistake what i mean and think they can plug a usb ethernet nic in it or something crazy. :P 12:03:56 Herkamire, 64Mb SDRAM, 4Mb flash, and MMC/SD slot (I have 256Mb MMC on mine) 12:04:07 :) 12:04:14 mostly i live out of the 4Mb flash because it's just plain faster than MMC or single channel SD. 12:04:16 sounds great 12:04:22 No mistakes here. It's just that USB looks far more complex to support than, say, 10(0?)Mbit ethernet. 12:04:45 http://gumstix.com/ .. see the picture of with the little white connector. that's the crazy 60pin hirose connector. they sell them in 3 packs on their site at least. 12:04:51 OrngeTide: do you have a keyboard hooked up to yours? 12:05:16 Baughn, well it runs linux. so if you have uhci or ohci and point the drivers at it, it should just work. :) 12:05:47 when the CF version comes out you'll be able to do usb host, 802.11, ethernet, gps, whatever. 12:06:07 Herkamire, i had my sun keyboard plugged into it for a while. that was amusing. 12:06:07 Ah... whatever, so long as it's cheap. 12:06:22 it's cheap until you try and put ethernet on it. :) 12:06:41 grrrr 12:06:42 if you must have ethernet then there are other boards that are a better deal 12:06:48 I don't need ethernet, really, but my chips have a nasty tendency of getting holed. ^_^ 12:06:49 kinda useless to me without ethernet 12:06:50 CF as in colorforth? 12:06:51 gumstix is just super small :) 12:06:55 futhin, compact flash 12:07:57 I mean, I cover them up as well as I possibly can, but the shrapnel finds a way in anyway... 12:09:17 Herkamire, there is a bluetooth version of gumstix. and there are bluetooth->ethernet bridges available. 12:09:36 although currently the BT linux drivers are having problems. 12:10:33 if you want a cheap ethernet appliance get a Mini-ITX. 17cm square and under $90 now. 12:11:37 nice :) does it have a RISC chip? 12:11:50 You of the cheap computers... 12:12:02 Mini-ITX starts at around 250$ over here. 12:12:03 Herkamire, well it's superscale x86. which is risc-like internally :) 12:12:09 Baughn, yikes! 12:12:15 I'm not going to touch x86 12:12:16 never 12:12:23 eheh. it's a VIA C3 though. 12:12:30 it's the least offensive x86 there is. 12:12:48 when I program in asm, I want it to be a low level language 12:12:53 x86 is rediculous 12:13:00 but to be honest, the only reason I got gumstix was because it was cheap, runs linux and NOT an x86. 12:13:29 Ok, *why* do you want low-level assembler? 12:13:36 I don't really need anything now, but in a few years I want to build a nice little computer for $200 12:13:44 preferably so I can sell it for $200 12:13:48 baughn: order your mini-ITX thru the net then 12:14:12 I thought that was the point of programming in assembly. 12:14:39 futhin: That would imply a 100$ tolling charge, and probably ~20$ actual toll, as well as some 30$ transportation on top of what it would be internally. 12:14:48 toll charge? 12:15:08 i prefer pdp-11 and m68k asm over RISC. 12:15:20 i prefer machineforth ;) 12:15:37 futhin: Our toll dept. charges us for looking at tollable packages, regardless of whether there's any actual toll involved. 12:15:47 x86 stinks. too much black magic to optimize L1 and L2 cache access. 12:16:01 futhin, i don't like machineforth at all. 12:16:36 i only accept addressable stacks. because i don't like long context switches. 12:16:51 OrngeTide: Agreed. I wound up doing Bad Things to %rsp just so I could get enough registers, and I *still* didn't beat icc. 12:17:03 OrngeTide: My code was marginally readable, though. :p 12:18:01 my job here is to do performance testing and optimization for Opteron. it's a nightmare :P 12:18:52 luckily linux does some very generic and unoptimized things in many places. so I can collect all the "low hanging fruit" and feel like I am accomplishing something. 12:19:10 well i'm hungry. ttyl! 12:22:46 OrngeTide: addressable stacks? huh? x86 doesn't have addressable stacks 12:25:20 Hi futhin: thanx again for hosting the code! :) 12:26:12 Raystm2: no problem 12:26:22 Raystm2: got more to host? ;) 12:26:55 working on changeing to useing arrays and jumptables later this week :) 12:27:15 make it more 'colorForth'y is that a word :) 12:27:35 sure it is ;) 12:34:23 I think I know how to get rid of the 'm' command needed each move if I stuff it into the direction-move controllers 12:53:24 --- join: Topaz (~top@sown-86.ecs.soton.ac.uk) joined #forth 12:55:49 --- quit: mur_ (Remote closed the connection) 12:56:22 --- join: mur (~mur@smtp.uiah.fi) joined #forth 13:13:08 --- quit: Topaz (Read error: 113 (No route to host)) 13:15:06 hmmm 13:15:10 gumstix.com rules 13:15:16 imagine colorforth on that 13:15:23 --- join: Topaz (~top@sown-86.ecs.soton.ac.uk) joined #forth 13:15:29 enough ram to last ya forever :) 13:17:02 --- quit: qFox ("this quit is sponsored by somebody!") 13:28:29 --- part: Raystm2 left #forth 13:32:19 --- part: wandelf left #forth 13:37:40 futhin, x86 has addressable stacks. what do you think %ebp is? 13:38:36 are you talking about system stack? 13:38:45 stack in general. 13:39:03 i don't know what stack in general means 13:39:11 then don't worry about it. 13:39:29 obviously forth data stacks and return stacks are completely different "stacks" than whatever "stack" is in x86 13:39:38 theres no stack in x86 13:39:43 so you must be refering to system stack 13:39:43 x86 can have an arbitrary number of stacks 13:39:44 hi 13:39:45 memory stack 13:39:57 futhin, ESP is usually return stack in x86 forths IIRC 13:40:15 slava: yup 13:40:27 slava not always 13:40:35 only in native forths 13:40:35 i wonder what leads to simpler code -- callstack on esp, or datastack on esp? 13:40:41 stack cpus that don't have a "stack pointer" can only has as many stacks as the hardware has built-in. 13:40:42 i think with dtc you'd want datastack on esp 13:40:46 I440r_ right? 13:41:30 which makes context switches really slow because on really simple stack cpus you have to pull a fixed number out of the stack (the max stack depth) and store it somewhere. then load up another copy of the stack. 13:41:33 native forths arent like DTC or ITC they use a call/ret mechanism 13:42:06 with your typical cpu you just keep a few different memory regions for different stacks of different contexts. switch context, just set a few registers 13:42:06 dct/ict usually use esp for the parameter stack and use epb as a return stack pointer into a software controlled stack 13:42:37 which makes more sense because esp is faster and one does handling of parameters far more often than call/return happens 13:43:07 OrngeTide: so you believe that a stack cpu should just have TOS and then keep the rest in memory? 13:43:40 futhin, only if it's fancy enough to cache memory intellegently. or that it's slow enough that it operates at the same speed at the SRAM it's connected to 13:44:40 if a stack cpu has internal stack. it should have a bunch of them and possibly keep track of the depth so you can copy only as much as you need when you do need to unload a context 13:45:48 I440r_: remember how you were complaining about the A register in forthchips? well chuck moore put that in because he's optimizing for moving bytes around as thats the primary activity of computers nowadays. and it pretty much removes any need for local variables (not that you would ever use local variables ;) 13:46:24 i like his non destructive IF, gets rid of any need for ?DUP too :) 13:46:43 would be better then if he added an and a a b and a c rgister - a = source, b = destination, c= count 13:46:57 local variables are a neat idea. given names to things so you can refer to them. but of course I often can't come up with a good name for a local variable so they are all named thigns like q, s, r, or i, j, k in my c programs. :) 13:47:01 he's got the @+ 13:47:06 fetches and increments the a register 13:47:06 but the fact that it uses a register makes it unforthlike imho 13:47:10 no need for a counter at all! 13:47:41 well you cant fetch from any memory except you have the address in the a register right ? 13:48:10 you can use the return stack to fetch from memory too 13:48:11 so if word a is scanning throiu a range of addresses and calling word b. and word b is processing that data and writing results into a buffer 13:48:19 they BOTH want to use the a register 13:48:30 I440r_, yea. moore and fox had a lot of problems with people who only did ANS forth. machineforth is very unforthlike. of course it's an assembly language. it's quite easy to write a forth for they say though. 13:48:39 that register causes MORE problems than it fixes 13:48:42