00:00:00 --- log: started forth/03.05.03 00:35:56 --- quit: njd ("praetorian") 00:37:04 --- quit: divgrad (Read error: 54 (Connection reset by peer)) 00:38:49 --- quit: Speuler (Remote closed the connection) 00:55:36 --- join: ramnull (~nicad@12-241-145-39.client.attbi.com) joined #forth 01:50:00 --- quit: ramnull ("This isn't Happy Hour!") 02:52:05 --- part: a7r left #forth 03:35:32 --- join: Speuler (~Speuler@mnch-d9ba449a.pool.mediaWays.net) joined #forth 05:52:43 --- join: ramnull (~nicad@12-241-145-39.client.attbi.com) joined #forth 05:56:08 --- quit: ramnull (Client Quit) 06:19:20 --- join: ramnull (~nicad@12-241-145-39.client.attbi.com) joined #forth 06:19:33 Everyone here lurking? 06:20:28 Yes 06:21:06 Apparently ALMOST everyone. Heh. 06:21:15 Just taking a break. 06:21:39 Got the Framebuffer running under IsForth. 06:22:23 Neat 06:23:04 Problem is, whenever I exit a draw procedure, I dumps my terminal into the deep dark. I gotta reboot in order to get the terminals working again. 06:23:37 i.e. I have to type shutdown -r now blind. 06:24:50 I just got an idea. brb 06:24:52 --- quit: ramnull ("This isn't Happy Hour!") 06:39:35 --- join: PoppaVic (~pfv@s61.waters.gtlakes.com) joined #forth 06:42:48 --- join: sifbot (~sifforth@h000ae66d7c31.ne.client2.attbi.com) joined #forth 06:42:49 Type sifbot: (or /msg sifbot to play in private) 07:04:41 --- join: ASau (~asau@158.250.48.197) joined #forth 07:04:50 Good evening! 07:07:10 --- quit: ASau (Client Quit) 07:13:55 Hello 07:29:24 Hi :) 07:30:07 lo' ;-) 07:53:55 --- join: gilbertdeb (~gilbert@fl-nked-ubr2-c3a-74.dad.adelphia.net) joined #forth 07:55:16 http://home.hccnet.nl/a.w.m.van.der.horst/forthlectureA.html 07:56:11 what on earth is the Stallman convention? 07:56:26 No idea at all - it sounded like bridge, to me 07:58:25 http://home.hccnet.nl/a.w.m.van.der.horst/forthlectureD.html (stallman convention) 07:58:38 ah I found it. 07:58:52 We'll using Stallman's convention for the comment. I have 07:58:52 adapted it to Forth a little bit. 07:58:52 Stallman (me too) is of the "I'm not your dog" type. 07:58:52 That means that only fully formed grammatical sentences are allowed 07:58:52 in the comment. 07:59:36 heh - is that all? 07:59:40 yep. 07:59:49 geezus - People still scare me.. 08:00:33 hehehe. 08:00:53 In other news, Douglas Adams learnt forth. 08:01:01 heheha 08:01:02 he didn't write anything in forth though. 08:01:30 I've been looking around for a couple days now - for a decent set of string words.. It's just amazing how BAD most shit is. 08:01:52 decent set of string words? 08:01:59 you mean a string dialect of sorts? 08:02:26 s/of sorts/kinda thing/ 08:02:27 :D 08:02:41 well, yeah kinda-sorta' 08:03:07 In some cases, I find good 'names' - but the code sucks - and then vice-versa. 08:05:25 ..I do think I'll try to adopt those regex things, and - of course - the bnf.fs file.. But, generally speaking, there just ain't a "good" set of string words - heh.. C has better funcs, actually ;-) 08:05:52 biggest selling point of C isn't it? 08:06:15 not for strings, no: everyone on the planet whines about 'em 08:06:28 whats best for strings then? SNOBOL? 08:06:47 I suspect a lot of folks would holler "Perl", oddly. 08:07:02 ah sed/awk :D 08:07:08 yeah 08:07:40 Well, for "autofoo" - regardless which way it grows - I simply MUST get together a decent string lexicon. 08:08:06 so you'll have to read the perl sources? 08:08:15 no way. 08:08:35 I'd almost prefer to relearn asm than suffer perl 08:09:14 Mostly, I'm thinking of what you'd want to do with a string, why, and then how. 08:09:54 in my usual universe, I can almost always get it all done with fprintf/sscanf and the occaional pointer-loop 08:10:14 eeek. pppoointers 08:10:21 But, of course, when dealing with dirs and files this gets a touch more complicated. 08:10:55 I wonder how snobol'll stack up against perl in string handling. 08:11:07 of course, I know neither language :D 08:11:13 no idea - I can only be happy it's going/gone the way of cobol 08:11:28 cobol is still quite alive and well I believe. 08:11:46 yeah - in very old places & code. 08:12:10 I think its the same sort of hope people express for fortran. neither are going anywhere. 08:12:47 Which is fine with me - as long as I never have to see it again. 08:14:27 --- join: mur (murr@baana-62-165-187-213.phnet.fi) joined #forth 08:14:49 terve mur 08:15:01 terve gilbertdeb! 08:15:21 what's up? 08:15:33 PoppaVic's string headaches. 08:15:46 :) 08:15:53 ;) 08:16:08 Not really "headache" as much as it's just really a noxious situation. 08:16:10 Terve 08:16:49 terve Robert! 08:25:54 --- join: tcn (~tcn@tc1-login18.megatrondata.com) joined #forth 08:25:55 --- quit: sifbot (Read error: 110 (Connection timed out)) 08:26:20 hey, anyone here? 08:27:10 yep 08:28:41 hmm.. you were talking about something interesting the other day 08:30:08 Hmm? Whazat? 08:31:52 good question :) 08:32:04 well, I just released Retro for Linux 08:32:30 I dunno' - I was back & forth into #c all day as well.. Only forth-stuff I have cooking is "autofoo" and such. 08:33:21 retro.. Url? 08:33:31 bespin.org/~tom 08:34:49 ok, lemme try that.. I'll have to use the phone in few, though - but Shall Return ;-) 08:35:03 I hate incomplete urls 08:35:19 heh 08:35:37 Xchat can seriously spoil yah ;-) 08:36:01 tinyirc won't :) 08:36:21 bah ;-) What? no docs visible for this? shame! 08:37:57 OK. Downloaded.. After I make a couple calls, shall return.. Much more cogitation to do on strings and such. 08:38:03 --- quit: PoppaVic ("brb - stay tuned") 09:02:38 hi 09:03:12 hey 09:03:27 gilbertdeb: you read douglas adams ? 09:03:45 * Speuler reading the log 09:09:55 t'is getting more difficult to connect to freenode 09:10:17 seems the ident timeout has been shortened 09:11:10 *** No identd (auth) response 09:11:10 --- Closing Link: (Connection Timed Out) 09:11:10 --- Disconnected (Remote host closed socket). 09:11:17 went on for on hour or so 09:11:43 * Speuler doesn't run identd 09:12:06 went to oftc then 09:12:20 got in straight away 09:14:33 --- join: PoppaVic (~pfv@s132.waters.gtlakes.com) joined #forth 09:14:53 Back in the saddle 09:15:25 tcn: you ever remember whatever it was? ;-) 09:19:14 heh.. i've been looking at the logs.. can't see anything that really jumps out 09:19:23 s'ok ;-) 09:20:24 hm.. forth/c interaction, maybe 09:20:39 ahh.. Yeah, that's always fun. 09:21:28 shit, maybe if I wasn't in such a hurry to this done with and get outside, I could remember what I was thinking :) 09:21:37 heh 09:21:57 just a sec.. gotta fire off a mail to i440r :) 09:21:58 S'ok, I'm taking a break to try working with qcad, myself. 09:23:48 --- join: wossname (wossname@HSE-QuebecCity-ppp80751.qc.sympatico.ca) joined #forth 09:24:51 howdy 09:25:10 allright, done with that email 09:25:22 heh 09:26:09 so you've been talking with i440r, eh? 09:26:28 me? Not much - I refuse to embrace asm 09:26:42 me neither.. but it beats gforth 09:27:19 hmm? GForth works - and has the bonus of sources and some docs - all fairly readable 09:27:39 you think i440r's convinced yet that C is sometimes a good thing? 09:27:49 no - he hates it ;-) 09:28:00 ..and I refuse to suckle-up to asm 09:28:03 heh.. damn 09:28:50 i've picked up some good stuff by working with isforth 09:29:08 oh? forth? asm? 09:29:33 hi Speuler. 09:29:36 well.. it helps w/ interfacing Retroforth to linux, for example.. helps w/ testing. 09:29:44 No, but I just got my first terry pratchett. 09:30:05 heh "kids" ;-) 09:30:11 hoorj 09:30:42 kids? which kids? 09:30:53 gil ;-) 09:31:05 terry pratchett is good. 09:31:28 I rarely read non-fiction... 09:31:41 I rarely read fiction. 09:31:55 its been a while so its my first trip back into fiction . 09:31:59 you rarely read/ 09:32:04 s/non-fiction/fiction/ 09:34:46 poppavic: well at least CRC is sensible about C.. heh.. maybe i440r will see the light when we have a 5k kernel, 5k forth and 20k C compiler, all playing together without Linux :) 09:35:08 bah ;-) I like linux 09:35:22 it could be better 09:35:37 Many things could be 09:35:45 say, was that you talking about namespaces vs. vocabs ? 09:35:58 prolly, I tend to like 'em 09:36:07 both?? 09:36:14 I can't recall - did I dcc a file to you yesterday? 09:36:18 nope 09:36:25 ok - twas another gent.. 09:36:35 btw, this client don't do DCC :) 09:36:39 hehe. another gent. 09:36:43 my stack.fs file uses a 'namespace' sort of lexicon. 09:36:51 it appears there is exactly one female 4thsmith on earth. 09:37:12 I never lift their skirts to find out. 09:37:13 'gent' is a safe assumption around here :) 09:37:38 heh.. "just makin' sure" 09:37:55 * gilbertdeb lifts up his skirt. 09:37:59 Ooops. 09:38:03 :D 09:38:11 But, yes.. I like "namespaces" as well as "vocabularies".. Not too hip to the latest shit, though - or the gforth setup of 'em. 09:38:35 hah! I got to see what was worn under yer kilt! (what was that short lil' thang? ;-) 09:38:55 hehehe. 09:39:12 ..see a doctor - get that boil removed! 09:39:21 namespaces, that'd be a strict hierarchy, right? like forth.stack.push or whatever 09:39:30 hehehe 09:39:56 yeah - it's much like the forth idea of a specialty "lexicon".. But not a lot of fun to implement. 09:40:41 well, i guess we can hang it off the filesystem if we need namespaces in Retro.. sort of like /proc 09:40:46 in C, I'm pretty much to the point I'm starting to define structures mostly containing func-vectors. 09:41:24 yeah, that's the power of C :) 09:41:26 ..writing a new/extra preprocessor would work, too.. 09:42:29 well, I'm not sure why, but Forth simply doesn't do much with 'em.. I really am sorta' disappointed at what has become 'standard' - after leaving for 2 decades. 09:42:31 procfs is kinda stupid though, when you have the kernel translating something to text and a program translates it right back to binary :) 09:43:01 I can't recall ever using the proc/ stuff - except interactive from console. 09:43:40 i'd use /proc like a dictionary, with the "CFA" in the inode slot.. 09:43:55 nahh.. Voodoo like that is just nasty 09:44:12 oh - you mean as forth words? 09:44:27 not exactly sure yet.. 09:44:48 I can't even really puzzle out how the dictionary stuff works anymore - everyone seems to have gone off in insane directions. 09:44:58 yeah 09:45:07 vocabularies and crap 09:45:32 FIG was ok, f-83 was ok, and F-pc worked great.. I leave, come back - and no one seems to want to make a sensible stand. 09:45:40 in Retro there's just 'forth' and 'macros'.. macros are always 'immediate' 09:46:04 heh - I added wil badens macro thinge to my files.. Never used it yet, though 09:46:28 i've looked at F83 a bit.. if only we had a standard or "model" like that for today.. 09:46:53 yeah, I know *sigh* I can't understand wtf the standards morons thought they were doing. 09:47:32 abstract specs instead of realistic models 09:48:21 Well, within limits.. 09:48:34 you know, /proc causes some problems.. if you try to 'cp -R' it'll hang up on /proc/mem or something :) 09:48:41 The nice thing about C is that generally, you can learn your limits.h & shit. 09:49:20 ..in forth, it appears ANS is half-assed, and everyones implementation is at least as half-assed.. It's sorta' depressing. 09:49:46 PoppaVic: too many synonyms I think. 09:50:00 gilbertdeb: how do you mean? 09:50:25 I think i'd put all this 'pseudo-filesystem' shit one level above the fs.. so you have /kernel /drivers /forth /dev and /fs at the top.. and /fs being '/' as far as open() is concerned. 09:50:55 I simply can't see forth-as-linux, tcn ;-) 09:51:04 PoppaVic: I mean people only hal-assedly write ans-forth's because the little they do contains enough of forth. 09:51:08 oh yeah, THAT's where i440r is smart.. he says, fuck ANS :) 09:51:51 I can agree with "Fuck ANS" - but not unless you have a new standard you propose that makes as much sense as fig or fpc 09:52:15 FIG-32 09:52:33 gilbertdeb: yeah, everytime I look at the ANS stuff my guts churn - but everones attempt to sorta'-maybe-kinda' comply is at least worse 09:53:01 they should be looking at F83 instead 09:53:07 PoppaVic: if _anyone_ ever agrees to a minimal forth environment from which all the other words can be generated... 09:53:15 perhaps the minimal env, CAN become the standard? 09:53:21 I should think it's really time to cobble up a new std & fuck the comittee's 09:53:52 yea. meet in here, set up a mailing list (or CC lis).. 09:53:59 gilbertdeb: yes, sorta'.. I've been looking at the ans-shit, and gforth, pfe, etc trying to see what NEED be, SHOULD be and CAN be.. 09:54:35 did you see how few words are in Retro? 09:54:47 I"ve never actually seen retro. 09:54:51 no - I grabbed it, but ain't been into the tgz yet. 09:54:56 but I've seen its haiku :D 09:56:08 I believe it may be - finally - time to admit that the datastack really can't contain floats and shit, unless we can identify such.. *sigh* It's really rather irksome. 09:56:18 i was thinking about some minimal forth too. only, i tend to call it "common forth" as it is not the intention to just offer a minimal word set, but a word set common to several standards 09:56:36 and some standard-specific extensions files 09:56:41 Speuler: exact same thing I think of in 'minimal'. 09:56:50 it's got a minimal set of asm words, just enough to compile everything else from forth.. stuff like * is done with machine code using C, :) 09:56:50 Well, you need to start with basic concepts first - I think that's where ANS blew it. 09:56:50 perhaps common is better :) 09:56:59 CommonForth, CommonLisp. why not? 09:57:24 But, I agree - the standard should be as minimal as 'sh' - with a clear set of minimal tools. 09:57:37 the name transports the meaning of "standard independance" quite well 09:58:21 I'd also suspect we'd want clear "extensions" rules - folks are just rehacking the same shit over & over - for 20+ years.. It's maddening. 09:58:38 hmm.. modularity? 09:58:45 would be lovely 09:58:46 forward references 09:58:49 heh 09:58:52 * Speuler ducks 09:58:56 hmm 09:59:04 there's a reason for those ... 09:59:09 no reason not to make them deferred words, sure. 09:59:13 binding to library words 09:59:27 include library at end of compilation 09:59:36 jsut resolve what has been referenced 09:59:59 I really had high hopes when I downloaded FICL, but.. *sigh* Man... it looks like a real pain to deal with. 10:00:22 Better standardize a tiny core Forth first.. i dunno if you'll ever get anyone to agree on the rest of it. 10:00:30 problem arises when using a compile-time defined word with forward references during compile time 10:00:47 but, i could imagine resolving against library several times 10:00:57 Well, I'd HOPE the goal would be a core, then extensions - but all such that a rommable-copy could be generated as well as interactive or compiled-native-code 10:01:28 for example. when a compile-time defined word with unresolved references is executed at compile time 10:01:48 Speuler: yes.. I've noticed all sortsa' state/mode voodoo - almost suggesting a third mode is waiting to be born. 10:01:56 haha 10:02:02 that's not really a third mode 10:02:14 no - but I believe there IS a 3rd.. 10:02:22 i've been experimenting with automatic forward references resolving a bit 10:02:34 interpret is easy.. compile is fine... meta? 10:02:42 most problematic thing was. typos 10:03:09 Speuler: example? Sorta' lost me there 10:03:25 What if you think of Forth as just a prototyping language? Just one tool in the toolbox.. Does that help? 10:03:40 : foo 10 0 do grr . loop ; 10:03:48 grr builds a fwd ref 10:04:02 : grrr 100 rnd ; 10:04:04 tcn: forth is a language for writing dialects. 10:04:12 grrr was meant to resolve grr 10:04:15 so there needn't be THAT much provided in the commonforth no? 10:04:18 but it doesn't 10:04:20 typo 10:04:47 gil: So it's no surprise there are a million different versions 10:04:52 i "solved" it by color-reversing the prompt if there were any unresolved refs left 10:05:13 tcn but they all really should have a common denominator, so that compatibility issues won't be _that_ severe. 10:05:17 fwd ref headers where placed in an own voc 10:05:29 thus, unresolved words showed me those 10:05:49 tcn: well, as I said - I'm working on "AutoFoo" - but that's supposed to be for bypassing the GNU autoshit learning-curve/verbosity. 10:05:51 gil: yeah, just so it sorta makes sense from one to the next. 10:05:57 i was helped by the unusual header structure i used 10:06:18 because i could easiily remove headers, regardless of order 10:06:25 No.. I'd rather see a deferred word than that forward. 10:06:44 PoppaVic: you wouldn't want to defer all words you use 10:06:45 speuler: makes sense.. I was gonna do forward refs that way in my assembler, in 1 pass.. 10:06:54 i think of resolving words against a library 10:07:12 so that just the required words are compiled into the program 10:07:20 OK, let's sidestep for a moment... How would we deal with namespaces, lexicons & vocabularies? I suspect that 20 years is long enough to suggest some thoughts. 10:07:54 i keep headers seperated, and no pointer from body to header 10:08:05 just a pointer from header to body 10:08:06 "word word word[ word...]" is all fine & dandy, but.. sheesh 10:08:32 that allows me to rearrange headers, including their storage format, however i like 10:08:45 i can put them into a btree, for example 10:08:48 just by loading a source file 10:08:58 hmm... well, a way to get to nfa from cfa is still nice - interactively/interpretive. 10:09:04 no need to patch up the bodies 10:09:15 but. i can 10:09:28 i just search word list until cfa i'm looking for is matched 10:09:34 that's slow, admitted 10:09:45 I certainly preferred the old f83/fpc sep of headers & bodies - it was to use segments & 640K mem, but.. applicable to ram/rom, too 10:09:47 but ref'ing headers doesn't need to be quick 10:09:51 poppavic: i440r puts a NFA pointer at CFA-4 10:09:59 sure. 10:10:10 tcn: that fixes your header to its address 10:10:17 can't move header anymore 10:10:21 or need to patch up body 10:10:27 can be done 10:10:32 Certainly. 10:10:36 i can just move headers around, no patching required 10:11:04 [ [HEADER] SPACE...] [ [ BODY] SPACE.... ] 10:11:13 and I personally don't give a damn about finding the NFA 10:11:16 tcn this looks interesting: -> http://www.cc.gatech.edu/fac/Olin.Shivers/ll.ps 10:11:18 for the penalty that "debug" or "see" takes a bit longer to produce the name of a word 10:11:29 title was A universal scripting framework 10:11:55 tcn: for an interactive app, it matters.. For a "turnkey" or "native-compiled" app, it wouldn't if no interpretation was needed. 10:12:20 there are not many cases where you need header from cfa 10:12:26 I don't use forth for "big programs" 10:12:31 some diagnostics 10:12:33 debugging/patching 10:12:53 where it doesn't matter that it is slightly slower 10:12:56 tcn what do you use for "big programs"? Small-c? 10:12:56 tcn: as it is now, I can understand why 10:12:58 I say, the simpler the better.. don't need debugging aids for little calculations 10:13:23 gil: yeah, some form of plain old C 10:13:28 No.. If I sit down to an interpretive environment, I want to be able to do whatever I need to do. 10:13:40 i basically use "find" for body-to-header lookup 10:13:53 ..otherwise, I might as well use perl or some schreck - which I refuse to do. 10:13:54 , better, find uses "header" 10:14:04 and ">name" uses "headers" too 10:14:20 Speuler: fine - this runs the voclinks/nodes looking for an address match? 10:14:26 "headers" gets an xt, and executes it for each header 10:14:36 PoppaVic: exactly 10:14:45 * PoppaVic HATES all these newish body/field names.. 10:14:55 looks for the header which matches the cfa, passed as parameter 10:15:16 poppavic: you mean 'xt'? 10:15:17 fine.. For debugging/interactive the delay is not significant - can live with such. 10:15:30 that's what i found too 10:15:35 tcn: that and some others, yeah - makes me nuts looking thru my old books & docs 10:15:52 after all, the interpreter uses a very similar lookup all the time 10:15:56 CFA, NFA, that's old right? 10:15:57 OK, so let's back up a sec.. 10:16:02 tcn: right 10:16:19 We sorta' agree a "commonforth" would be nifty.. 10:16:34 i could agree with that 10:16:43 which means a lot :) 10:16:46 How much of such a beast needs to be machine-mentality "defined"? 10:17:05 primitives ? 10:17:09 no... 10:17:23 terminology.. CFa, NFA, return stack, data stack.. 10:17:31 I mean - how much oh the actual data-structures/access need to be codified.. 10:17:35 oh/of 10:17:42 none 10:17:46 tcn: YES, please ;-) 10:18:00 I miss cfa/pfa/nfa, etc 10:18:21 as close the implementation matches the cpu (register use for example), as better its performance 10:18:22 well, ok - can it be written akin to C & limits.h + sizeof(), etc? 10:18:48 speuler: Forth is not a performance language! it's a SIMPLE language! 10:18:50 --- join: kc5tja (~kc5tja@ip68-8-206-137.sd.sd.cox.net) joined #forth 10:18:53 no good to emulate a stack and stack pointer as variable if there's as stack available 10:19:21 OK.. Wait - I am not now, nor next week, going to write in asm - I just hate it.. So, you see this raises the next set of questions - about machine, lang, access, platform, etc. 10:19:28 but you do need to codify a dictionary structure to define CFA, NFA, PFA 10:19:40 Right. 10:20:01 ..and - preferably - in such a way that the CommonForth, asm & C could all deal with the same damn structures 10:20:12 tcn: no, not necessarily 10:20:18 no need to define 10:20:20 Yes. 10:20:30 you can be flexible with that 10:20:42 even to the point of being able to change it while the system is running 10:20:46 The FORMAT needs to be understandable, or you'll never see it used in embedded devices or in a "major app". 10:21:11 i tend to "pack" headers when saving the system 10:21:29 well, the initial code has to know a precise layout, I agree however that the USERS can use words to reach whatever - IF they are proerly written. 10:21:34 just (flags) (len( (header) (flags) (len) (header) ... 10:21:54 so.. LFA, CFA, PFA, and variable length NFA as the last thing in the header.. that's acceptable to C.. 10:22:03 i relocate those down to here 10:22:10 yeah, "byte, byte, byte.." doesn't bother me.. The alignment crap can drive you up a wall, though 10:22:15 --- join: crc (crc@AC8D918B.ipt.aol.com) joined #forth 10:22:21 we're not talking about your forth, speuler..we're talking Common forth :) 10:22:39 hello crc 10:22:41 tcn: if you don't appreciate my input, i can shut up 10:22:45 hello tcn 10:22:46 Personally, I don't EVER want to have to decide IF or WHEN to pad/align - that's not my job as a human. 10:23:37 Speuler: show me an url, I can DL & look at it - the talk can continue, but.. I'm trying to see how this "CommonForth" thing should branch away from the ANS and everyone-for-himself free-for-all. 10:23:38 speuler: i get what your doing.. let's not go off on tangents now.. 10:24:16 I'm pretty sure there are a lot of snippet/ideas out there to hijack - but when, why and how matter, too 10:24:20 crc: we're talking about a "Common Forth" core model/standard to replace ANS :) 10:24:38 PoppaVic: the code containing those vocs, along with binding to libs from fwd refs, is not on the net. i could upload it though 10:25:19 best thing to run it under a "conventional" os would be msdos ... 10:25:32 no linux support in there 10:25:35 A replacement standard for ANS? It's about time. 10:25:38 mainly i/o 10:25:53 and binary image format 10:26:01 linux is easy to deal with, actually 10:26:26 Speuler: well, if the source is uploaded, please let me know - Bog knows, I must have like 15 forths in ./legacy at the moment 10:26:41 PoppaVic: all right 10:26:58 linux is nice & easy.. trying to stay at "posix" is even nicer - then the *bsd folks should also benefit. 10:27:15 how do you mean? 10:27:24 bsd folks benefit from what? 10:27:29 tcn: you done anything much with dlopen/sym/plugins/libs? 10:27:35 not yet 10:27:42 gilbertdeb: the Commonforth thing - and platforms 10:27:52 it's on my todo list :) 10:27:57 I've done a very, very small amount of it. 10:28:07 PoppaVic: I have (in C). It's pretty easy to use, but I dislike their error reporting mechanism. 10:28:18 Yeah, it did look a little whacko 10:28:57 --- quit: mur ("i might be back later") 10:29:09 hmm.. how would Commonforth deal with 16, 32, 64-bit CPUs? a slightly different standard for each, instead of one size fits all? 10:29:15 Hmm... you know.. It also occurs to me that 'types' are possible, too. 10:29:39 * kc5tja used it pretty heavily in writing GCOM, a generic COM implementation for Linux. The project never went anywhere on SourceForge, but the system works great for local (e.g., in-process) objects wonderfully well. 10:29:39 tcn: yeah, I have been thinking about that... GCC seems to deal, though. 10:29:43 for 64 bit and > cpus, one might consider byte tokenized forth 10:30:25 hmm, didn't I mail you a doc awhile ago, Speuler - about that? 10:30:26 regarding types: http://bespin.org/~tom/writings/retro.addendum.txt 10:30:37 it's short 10:30:39 but it may be a bad idea to fix the threading model in the commonforth specs 10:30:49 agreed 10:30:50 hmm.. shit - Xchat wouldn't pick that up... 10:30:58 PoppaVic: i have started a byte token implementation 10:30:59 got it 10:31:10 it is on www.forthfreak.net 10:31:13 yeah, I too agree 10:31:46 Do like FIG, implement working Commonforth models on a bunch of platforms/OS's 10:32:05 however... I believe that we can all agree that our ACCESS to the shit all needs to be codified - but, that the binary-arragement is not the job of the std, right? 10:32:14 Right. 10:32:29 no. do like ans, specify what words are doing, not how they are implemented 10:32:37 and the header format, that's for purposes of explaining CFA, NFA, etc.. not set in stone 10:32:40 ok, note: RTTI is more a C++ thing.. 10:32:57 Perhaps the binary standard could be an adjunct standard? That is, split the access standard from the binary standard, so that implementations MUST implement the access standard, and the binary implementation is optional? 10:33:07 hmm.. 10:33:22