00:00:00 --- log: started forth/01.12.07 03:43:32 --- join: aum (~aum@210-86-60-227.jetstart.xtra.co.nz) joined #forth 03:44:05 --- part: aum left #forth 04:30:02 --- join: aum` (~aum@210-86-60-227.jetstart.xtra.co.nz) joined #forth 05:25:35 --- quit: aum` () 07:03:12 --- log: started forth/01.12.07 07:03:12 --- join: clog (~nef@bespin.org) joined #forth 07:03:12 --- mode: sagan.openprojects.net set mode: +n 07:03:12 --- names: list (clog) 07:26:45 --- mode: ChanServ set mode: +nt 10:27:13 --- join: Speuler (akhandel@banane.icafe.spacenet.de) joined #forth 10:27:18 --- part: Speuler left #forth 10:57:33 --- join: Speuler (akhandel@banane.icafe.spacenet.de) joined #forth 11:19:37 --- join: klooie (kloo@213-84-79-23.adsl.xs4all.nl) joined #forth 11:19:39 hello. 11:22:20 --- join: MrReach (~mrreach@209.181.43.190) joined #forth 11:22:30 greetings! 11:23:17 --- mode: ChanServ set mode: +oo klooie Speuler 11:23:17 --- mode: ChanServ set mode: +o MrReach 11:23:23 how goes? 11:24:15 hi MrReach. 11:24:27 i'm fine, still pretending to be a forth guy. :) 11:24:40 don't we all? 11:25:06 every once in a while i remember that i would really like to do something with forth though i don't have the time, so i join this channel. 11:25:21 what would you like to do? 11:25:22 this time reading in phil koopman's book about stack machines did the trick.. 11:25:48 i would like to implement a subset of lisp in forth. 11:26:16 preferably on a real stack machine.. 11:26:41 * MrReach nods. 11:26:54 I was thinking about something similar yesterday 11:27:08 lisp is stack-heavy too, and also has small words leading to lots of jumps/calls that stack machines are good at. 11:27:27 I finally concluded that the resulting language wouldn't look much like LISP at all, although it would be predicated on LISP 11:27:35 well. 11:27:45 i've been reading about both forth and lisp, but have not done significant work in either. 11:27:57 there's this book - the title escapes me - that shows how to implement lisp in pascal. 11:28:13 with that as a guide, i could do it (at least partway) in forth.. 11:28:32 --- join: edrx (edrx@200.240.18.99) joined #forth 11:28:42 just as a learning experience.. the world doesn't really need another toy lisp-alike. :) 11:28:56 I was thinking that computational power of desktops and internet connectivity has increased enough that Douglas Lenat's Eurisko program could be unleashed as a text parser on USENET 11:29:30 * klooie does not know Eurisko. 11:30:00 Eurisko is based on Minky's frames concept of data organization 11:30:35 Eurisko is an interesting gadget ... it thinks about rulesets, those of the outside world, and those in itself 11:30:57 minsky frames are at least vaguely familiar to me. 11:31:15 It is now barred from participating in the Housten Simulated Naval War Games competition 11:32:18 google told me a little about Eurisko, that's very interesting. 11:32:24 * MrReach looks for a reference 11:32:52 hrm, i have to go. 11:33:01 * klooie will return in due course. 11:33:13 ok, be well 11:33:16 you too. 11:34:04 http://www.media.mit.edu/people/minsky/papers/Frames/frames.html 11:34:12 did you get that? 11:35:43 I got it :) hi MrReach 11:35:52 * MrReach grins. 11:35:58 how are you? 11:37:50 fine, and you? 11:38:12 so-so ... seem to be fighting some type of infection 11:40:16 oh-oh... is it of the type that you can forget about when you are doing something interesting, at least? 11:40:30 yes, thank god 11:41:14 do you have a link about Eurisko at hand? 11:41:55 no, I have hardcopy of Lenat's original work 11:42:02 I'm sure I could find something, though 11:42:04 brb 11:43:11 do you read Spanish? 11:44:28 yes 11:44:45 http://w3.mor.itesm.mx/~emorales/Cursos/KDD/node17.html 11:45:13 I stumbled on that link a minute ago on a google search, I think... 11:45:42 aha 11:45:56 dict eurisko gives a link if you have dict-foldoc installed 11:46:01 ok, I used "eurisko lenat" 11:46:23 broken 11:46:26 me too :) 11:47:18 what's AM? 11:48:14 http://www.urticator.net/essay/3/323.html is very nice 11:48:35 AM (Automatic Mathamatician) was the precurser of Eurisko 11:49:06 was rather specialised, but led Lenat to realize that the concept worked, but it needed generalization 11:50:17 THAT's interesting 11:50:42 Lenat went to work for a company, Cyc ... heavily funded by the DoD 11:50:59 I see that they have made an open-source version of his progs 11:51:11 (sometime in the last two years) 11:52:17 I was about to say that it is a pity that we only see descriptions of those programs, not runnable/traceable/debuggable versions... :) 11:52:34 http://www.opencyc.org/ 11:54:34 Doug Lenat, Founder and President of Cycorp, will discuss OpenCyc and 11:54:34 artificial intelligence this Friday, June 29, on "Talk of the Nation: 11:54:34 Science Friday," 11:55:00 that might be too old to play an archive 11:55:31 ? 11:57:19 http://search.npr.org/cf/cmn/cmnpd01fm.cfm?PrgDate=06/29/2001&PrgID=5 11:57:37 show was prompted by the movie "A.I." opening 11:58:13 btw, do you program AI things using Forth? 11:58:28 not much 11:58:32 would like to 11:58:40 I never did anything with AI 11:58:49 it was A.I. that attracted me to computer programing as a kid 11:59:09 for me it was video games 12:00:35 and as the newer games were coming out and they were much more complex and much less fun, I became each time more fascinated by simple designs 12:01:04 g'day 12:01:14 g' 12:01:20 greetings, Speuler 12:01:31 g'reetings 12:01:39 has the forth been with you recently ? 12:01:45 Speuler: I've been doing a bit of research regarding shared libs in Linux 12:02:20 yes, and I even have some to share 12:02:24 you'll be very happy to know that "dlopen" library is only 5k or 10k in size (depending on which version) 12:02:32 that's nice. but not usable for your jornada ... 12:03:10 but for linking to xlibs that would be of great help 12:03:28 unforunately, it uses the glibc mapfile function, rather than the kernel syscall, which rather defeats the whole purpose in the first place 12:03:36 that the mem footprint is very small, is good for handhelds too 12:03:51 making support for x quite feasiblle 12:04:07 the dlopen lib has a lib dependancy on glibc 12:04:16 hmm 12:04:17 (idiots!) 12:04:33 hopefully, for handhelds there exists a rewrite if that fn 12:04:56 Speuler: did you ever do any benchmarking to see how much compute time is needed for glibc initialization? 12:05:06 nope 12:05:08 it's easy in windows 12:05:21 never bothered, because machine was quick enough 12:05:32 not so easy in linux 12:05:38 ok 12:05:45 the init, or the benchmark ? 12:06:06 the init 12:06:31 i move to a table, preparing for food takeup 12:06:36 l8er ... 12:06:53 I realize that it's a zero-time operation to map the glibc lib into the processes' memory, but I have no idea how much initialization it does 12:07:08 ok, Speuler, ejoy $f00d 12:07:15 --- quit: Speuler ("ircII/tkirc") 12:07:31 I'm sleepy 12:07:38 are you? why? 12:08:00 and later there'll be a show of a local glam rock band that I want to see 12:08:16 ok, nap time? 12:08:52 I have been programming too much and sleeping too little 12:08:57 yes, maybe 12:10:09 I've spent the last two days completely immersed in turning my minimal Forth-on-Lua language into something that can compile bytecodes and that has an inner interpreter in C 12:11:21 some postings on the MachineForth/ColorForth mailing lists made me get very anxious to have that working 12:11:51 why the sudden rush? 12:13:36 several people there were discussing ways to have MachineForth and ColorForth environments running on top of Linux, and I've been dreaming with that for at least one year 12:15:02 are they open source? 12:15:38 and I just had to port some old programs to my new Forth-on-Lua thing to start trying to glue their inner interpreters into my stuff 12:15:54 yes, even public domain in some cases (Chuck's stuff) 12:16:13 the problem is that most of the time they don't have proper source code 12:16:28 what language is the source in? metacompiled forth? 12:17:18 I think that most of the source for Chuck's Forths is huffman-coded strings buried into .com files that people use to write to the first sectors of floppies 12:17:39 but I have some links to more readable stuff, one moment 12:18:00 heh 12:18:05 leave it to Chuck 12:18:49 here's one that I've been able to run in a DOS emulator: http://angg.twu.net/e/fortho.e.html#c4th 12:19:33 I would like to be able to run Chuck's systems, but they don't work on my machine 12:20:09 and ideally I would like to run them without leaving linux and to convert the source to text to browse it with emacs 12:20:54 no doubt 12:21:55 there's some mystic around Chuck's code 12:22:19 oh? 12:22:40 oops, weird english 12:22:49 oh! yes ... "it's Chuck's code, there must be something to learn here" 12:23:24 even if it's just Jeff Fox that puts Chuck in a pedestal... I'm curious anyway 12:23:50 Mike Perry relates a story about Chuck's CAD system which ran 12:23:50 on the board whose design it had been written for. Chuck 12:23:50 showed Mike that he had written the core of the CAD application 12:23:50 in about five lines of code. When Mike asked him how long it 12:23:50 had taken, he replied, "Oh, about two years." 12:24:40 heh 12:28:53 and my 40-line Forth-on-Lua thing took me about 5 years :) 12:29:52 a few years more and it will become something that other people would like to use 12:30:24 heh 12:30:36 a few hours more and it will become something that people will just have to download a tgz an run two commands to run a demo 12:30:51 I'm glad to see you byte-code your forth 12:31:13 it should make a substantial diff in both speed and size 12:32:10 I'm doing it the ugly way for now - I produce a nasm file, that is then linked with engine in C... I haven't got yet to the point where the engine becomes a .so that can be dynamically loaded into Lua 12:33:08 but in the predecessor of the current verison it was possible to call lua from bytecoded programs 12:34:13 I haven't ported the demo that did that yet 12:34:37 * MrReach nods. 12:34:53 still storing the bytecodes into a c-string? 12:35:31 and the code for the predecessor is about 20K, including comments, and I every time I see it I find it ugly and don't have any motivation to understand it again 12:35:42 no, not a C string 12:35:51 do you understand nasm code? 12:35:51 * MrReach laughs. 12:36:05 yes, depending on how complex the macros are 12:38:36 look at the .asm and .asm.lst files in the crim/ dir 12:40:18 the bytecodes are just sequences of bytes in the .data section, with _f0 pointing to where they start 12:43:13 * MrReach nods. 12:44:12 * klooie is back. 12:44:23 hi klooie 12:44:27 hey edrx. 12:44:57 http://www.media.mit.edu/people/minsky/papers/Frames/frames.html 12:45:36 thank you, MrReach. 12:45:44 np 12:46:19 i think that for every language, there is at least one web page advocating its use for web programming. 12:46:31 of course 12:46:43 everyone wants to see their pet thrive 12:46:45 in the last two days i have seen mod_lisp for apache, a html spewing library for prolog, and now an article on forth CGI. 12:46:48 ok, nap+shower+dinner+show, with bits of programming in the breaks 12:46:55 * klooie nods. 12:47:11 * MrReach laughs out loud. 12:47:24 --- mode: MrReach set mode: +o edrx 12:47:48 klooie: where did you find the article on forth cgis? I remember seeing some cgis written in gforth and pforth, but they didn't convince me... 12:48:10 on comp.lang.scheme it was recently discussed how a scheme plug-in for browsers could really win the language some territory. 12:48:28 yes, I was not really impressed with Paysen's web server 12:48:47 http://www.geocities.com/oneelkruns/cgishell_paper.html 12:49:10 that article is about specific implementations. 12:49:18 .. it just triggered my observation. 12:50:56 klooie: thanks 12:50:59 time to go, bye 12:51:02 later 12:51:03 --- part: edrx left #forth 12:51:06 be well 12:51:29 somebody wrote a webserver in forth? 12:51:45 yes, lemme get the URL 12:52:01 http://www.jwdt.com/~paysan/httpd-en.html ? 12:52:13 that looks like it 12:52:28 yup, that's it 12:52:43 google, have to love it. 12:53:22 I prefer altavista 12:53:59 i used altavista before, it used to be fine but it's crumbling now. 12:54:12 oh? how so? 12:54:51 they have just merged all the national or language-specific databases they had. 12:56:00 hmmm 12:56:25 also i hear from semi-reliable sources that their database is aging. 12:56:31 time for me to do some physical work for a while 12:56:43 yes, bandwidth is a serious bottleneck 12:57:08 --- nick: MrReach -> MrGone 12:57:11 sounds healthy. :) 12:57:11 see you. 13:36:10 --- join: Speuler (akhandel@apfel.icafe.spacenet.de) joined #forth 14:36:38 --- part: Speuler left #forth 15:38:06 --- join: qless (qless@clgr000977.hs.telusplanet.net) joined #forth 15:38:41 --- quit: qless (Client Quit) 16:02:01 --- nick: MrGone -> MrReach 16:12:54 --- quit: MrReach (Ping timeout: 181 seconds) 16:34:45 --- join: MrReach (~mrreach@209.181.43.190) joined #forth 16:35:20 --- mode: ChanServ set mode: +o MrReach 18:11:22 --- join: qless (~qless@clgr000977.hs.telusplanet.net) joined #forth 18:12:21 --- quit: qless (Client Quit) 19:35:12 --- join: TheBlueWizard (TheBlueWiz@ip-216-25-202-189.vienna.va.fcc.net) joined #forth 19:35:18 hiya all 21:21:35 bye 21:21:38 --- part: TheBlueWizard left #forth 23:59:59 --- log: ended forth/01.12.07