00:04:38 [QUIT] nate37 quit: barnes.openprojects.net king.openprojects.net 00:04:38 [QUIT] ult quit: barnes.openprojects.net king.openprojects.net 00:04:43 ult joined #tunes 02:04:41 kev joined #tunes 03:57:16 Fare joined #tunes 07:04:31 Kyle joined #tunes 08:00:58 water joined #tunes 08:02:18 re 08:29:26 hcf joined #tunes 08:33:11 abi: peer to peer is [net] see O'Reilly's P2P directory at http://www.oreillynet.com/pub/q/p2p_category 08:40:34 hm 08:48:13 [QUIT] ult quit: ircII EPIC4pre2 -- Accept no limitations 09:04:45 mibin joined #tunes 09:04:54 hey mibin 09:05:09 hi water 09:09:48 anything on your mind? 09:09:56 do you know if someone is working on the c translator subproject? 09:09:58 water is doing a lot of thinking and reading 09:10:04 hum 09:10:10 no, not directly 09:11:12 but it's an aspect of the hll that's an application of what we're trying to develop in slate libraries (taking slate to be an hll proposal) 09:11:50 im writing a C declaration parser in scheme 09:11:55 e.g. slate objects that directly represent low-level objects and simple slate parsers 09:12:00 ok 09:22:27 [QUIT] mibin quit: Ping timeout for mibin[an1-212.dialup.tiscalinet.it] 09:24:34 bleh, i'm getting ahead of myself... getting ideas about the slate gui 09:24:48 ult joined #tunes 09:25:06 re 10:07:13 [MODE] ChanServ set mode: +o hcf 10:08:48 [MODE] hcf set mode: -o hcf 10:09:01 [TOPIC] hcf: TUNES: Free Reflective Computing System http://www.tunes.org || Slate Programming Language http://slate.tunes.org || Lambda PL weblog http://lambda.weblogs.com || Updated Mailing List Archives http://lists.tunes.org 10:14:41 icac, more p2p, http://www.exocortex.org/p2p/index.html 10:18:46 water: http://citeseer.nj.nec.com/sewell98from.html 10:20:15 k 10:58:36 foo 10:58:40 [QUIT] hcf quit: Ping timeout for hcf[207-172-225-149.s149.tnt1.pld.me.dialup.rcn.com] 11:13:52 [QUIT] ult quit: ircII EPIC4pre2 -- Accept no limitations 11:58:58 kev joined #tunes 12:12:47 hey kev 12:33:28 hcf joined #tunes 12:37:17 in the gee-what-an-improvement department: http://www.spherum.com/news.htm says "The Spherum website has been revised and improved." and has a link to file:///C:/WINDOWS/Desktop/Website/index.htm 12:38:34 lol 12:44:20 does CLOS have a "homepage" url? i dont mean the CLHS 12:44:42 not afaik 12:45:01 it really is most notable as being part of cl 12:46:23 does CL have a "homepage" url? 12:47:30 the official spec is ansi 12:48:02 lisp.org is the next closest 12:48:53 ANSI's site sells the spec right? 12:49:17 yep 12:49:33 so no point in pointing to them 12:49:54 dunno, why don't you check? 12:50:32 i'm trying to be lazy 12:50:56 well, try harder ;) 12:51:17 i could leave the factoids w/o urls 12:51:42 geez 12:51:52 water checks out ansi's site 12:52:11 no, i'l do it 12:52:41 x3j11 iirc 12:54:38 forget it, they only sell it 12:54:59 you may as well point to the CL part of the lisp.org site 12:58:56 ok http://www.lisp.org/table/references.htm points to the ansi x3 spec, x3 site seems down, oh well 13:05:14 [QUIT] hcf quit: Leaving 14:30:51 [QUIT] Freggel quit: 14:40:13 eihrul joined #tunes 14:40:23 hey eih 14:44:42 [QUIT] MysticOne quit: Read error to MysticOne[user-24-214-59-110.knology.net]: Connection reset by peer 14:54:17 [QUIT] diablovision quit: 14:56:31 [QUIT] eihrul quit: Leaving 15:32:41 [QUIT] gREMLiNs quit: Ping timeout for gREMLiNs[195.252.66.40] 15:36:01 eihrul joined #tunes 15:54:23 ult joined #tunes 16:24:00 lar1 joined #tunes 16:40:11 nwsh joined #tunes 16:46:31 anyone alive? 16:48:23 depends on the question of whether anyone can be considered to be truly alive 16:48:42 the short, consistent answer would be 'no,' though 16:50:15 right. 16:50:49 damn, a ufo just landed on my porch, better go investigate. 16:51:00 nwsh left #tunes 17:48:01 [QUIT] water quit: The Tao went that-a-way! 17:50:58 water joined #tunes 18:03:15 [QUIT] water quit: The Tao went that-a-way! 18:06:18 water joined #tunes 18:49:23 [QUIT] eihrul quit: Leaving 18:53:28 hcf joined #tunes 19:52:12 Kyle joined #tunes 19:53:28 eihrul joined #tunes 20:06:26 nwsh joined #tunes 20:08:18 so what exactly would you want to do with a metaprogram. what kinds of other programs would the metaprogram write? 20:13:17 [QUIT] MysticOne quit: Ping timeout for MysticOne[porpoise23.panama.gulf.net] 20:13:25 i mean, how exactly would the meta program know what the user want's to do? 20:18:06 self optimisation and morphing is possible, because the metaprogram woulden't be pulling code out of thin air, it would be talking what is already there and find the best way or another way to acomplish the same task. 20:19:01 eh? 20:19:18 wot? 20:19:40 have you heard of researchindex? 20:20:11 huh? 20:20:15 or run-time feedback through profiling? 20:20:20 abi: RI 20:20:20 RI is [research] ResearchIndex at http://citeseer.nj.nec.com/cs/ 20:20:40 it's a big collection of papers and bibs on cs research 20:21:03 anything practical in there? 20:21:09 like partial evaluation, meta-programming, supercompilation, program specialization 20:21:12 YES! 20:21:43 it indexes almost every CS research paper site 20:21:44 hmmmm 20:22:05 abi: pto 20:22:06 pto is ProgramTransformationOrg at http://www.program-transformation.org 20:22:32 ooo 20:22:40 and have you heard of lisp's macro system? 20:23:15 lisp's macros are meta-programs that allow you to (among other things) define your own new programming language over common lisp 20:23:32 well, not necessarily common lisp 20:23:33 nope, i'm just some kid. 20:23:46 abi: common lisp 20:23:48 common lisp is [pl+lisp] a dialect of the LISP family of languages, see http://www.lisp.org 20:24:09 abi: self 20:24:10 self is a prototype-based object system at http://www.sun.com/research/self/ or old mirror at http://www.cs.ucsb.edu/oocsb/self/ 20:24:57 self's sparc implementation demonstrates orthogonal optimization of high-level language programs using dynamic profiling information from the runtime system 20:25:47 hang on, i gotta learn what that means. 20:26:47 i wouldn't say "orthogonal" perse 20:29:15 ok transparent to the source program 20:29:46 umm, in other words, that self thingie looks at a HLL program while it's running and takes the bits that aren't usefull for the program out and keeps the rest? 20:31:11 nope 20:31:21 read up on it, kid 20:31:42 pffft, reading. 20:32:34 no, self does some gross profiling of the program, then optimizes optimistically the parts the profiling says should most likely be important :) 20:33:45 why not just optimise the whole thing? 20:33:54 because, then that wouldn't make the program faster 20:34:03 you're neglecting something... 20:34:33 lets say we have a reference "running time" for a program 20:34:42 which basically is a totally unoptimized version 20:34:53 k 20:34:59 now... 20:35:21 if you want to introduce optimizations while a program is running 20:35:32 you must not neglect the fact that optimization takes time :) 20:35:40 ah 20:36:07 so, total optimized running time = optimization time + optimized running time 20:36:32 which must be less than the unoptimized running time, for optimization to be of any benefit 20:36:44 now... consider, the "principle" of locality: 20:36:45 fair enough. 20:36:53 90% of the time is spent in 10% of the code 20:37:17 so, if we optimize the other 90%, we're just inflating optimization time, without decreasing optimized running time at all 20:37:27 iow, much work for little gain 20:37:52 now, if you can accept the static compilation model 20:38:02 k 20:38:06 this is all fine, since you just "forget" about optimization time 20:38:20 but static compilation model is very restrictive :) 20:38:42 necessitates either poor optimizations or impoverished languages 20:40:05 so the challenge is to find a better way of optimising code, right? 20:40:29 they've already been found, in some cases :) 20:40:50 ideally, the "most optimal" conceivable adaptive optimizer i can think of 20:41:23 would, via some form of profiling and heuristics, determine which parts of the code are most usefully optimized (somewhat easy) 20:41:59 then, determine what pieces of data can be declared run-time constant to provide the most improvement in the important parts of the code (much more difficult) 20:42:00 ) 20:42:15 and then partially evaluate based on that information 20:42:38 hmmm, k. 20:43:06 after you have the partial evaluated code 20:43:18 it is a matter of elementary static optimizations from there on out :) 20:43:43 (though, you need to be careful about how you do them, because as i stated before, you don't want to increase the total running time of a program) 20:44:08 fudge. 20:44:39 though, that is probably the subject of a PhD thesis right there :P 20:45:34 ahh. so, in other words, anything that i think of has been thought of way before i did, and a better way has already been found and implemented. great. :) 20:45:43 no 20:45:56 the "most optimal" conceivable adaptive optimizer i speak of does not yet exist :P 20:46:15 mainly due to the difficultly of determining what to partially evaluate :) 20:46:28 and also 20:47:04 nary a one tries to do very complicated static optimization in addition to exposing run-time constant information 20:47:37 [QUIT] Kyle quit: Leaving 20:47:38 still a long ways to go :) 20:48:01 hey, i think i get it. :) 20:49:28 so are you all uni professors and stuff? 20:49:33 hah 20:49:42 water is a college drop-out and i'm a college freshman :P 20:49:54 voluntarily 20:51:00 in general, tunes crowd tends to be slightly "academic" in a sense, but not established academic or anything 20:51:09 mostly arm-chair academic :) 20:51:41 ahh, well, i'm one of those too (college freshman), but i'm lazy at learning so's i'm not too knowledgeable about anything. 20:52:10 lazy at learning or lazy at being educated? :) 20:52:28 learning. 20:53:15 ouch, bummer 20:53:29 i guess saying sheet straight up is better that making excuses, so, yeh. 20:54:32 hmmm, how about converting the machine code into maths, simplifying, then converting back? 20:54:46 hmm, scratch that. 20:57:18 eihrul shrugs. 20:57:30 you're always representing stuff in some language 20:58:53 what you'd call "maths" would just be another "machine code" in the end 20:58:53 basically, with any optimization, you do represent the code in some form convenient to optimize, but what is convenient to optimize 20:58:53 unless you're a masochist 20:59:33 not exactly. yeh, i see yer point. 21:01:55 pffft. a serious case of 'best left to people that like doing what they're doing' 21:09:15 will tunes ever be a commercial or rebel OS or will it be mainly for research/the hard stuff? 21:09:43 tunes is more of a spec than an implementation 21:10:06 (imo) 21:10:14 why not go all the way? 21:10:19 heh 21:10:26 we're trying :P 21:10:39 ah 21:10:45 no, i mean that the actual code base we develop need not be the Unique Only Way to do it 21:11:01 in fact, i can think of a couple of ways to do it 21:11:30 that wind up having different flavors, but wind up being the same thing in the end 21:12:07 k. 21:12:37 would it help if i learnt to code in ia64 assembly? 21:13:41 yeah, if it makes you feel better ;) 21:13:42 hangon, would it help if i stopped asking brainnumb quesions and found sheet out for myself... :) (yes, ok...) 21:14:08 you're welcome to ask us for links 21:14:40 k. 21:18:28 but still, the problem is i won't learn nearly as much just reading stuff on the internet versus people that go to university and have access to funky tech and lecturers (and other reasons that i won't get into) 21:18:43 moot 21:19:14 on with the games and tv. 21:19:25 thanks people. 21:19:28 [QUIT] nwsh quit: Read error to nwsh[dialup-41.visp2.unite.com.au]: Connection reset by peer 21:46:04 [QUIT] water quit: Ping timeout for water[c207-202-221-160.sea1.cablespeed.com] 21:47:24 water joined #tunes 22:17:59 [QUIT] _ruiner_ quit: Read error to _ruiner_[24-216-192-59.hsacorp.net]: Connection reset by peer 22:51:06 nwsh joined #tunes 00:02:34 [QUIT] eihrul quit: Leaving