IRC log started Sat Feb 26 00:00:02 2000 [msg(TUNES)] permlog 2000.0226 -:- SignOff air: #TUNES (http://www.qzx.com/ :: sleep) -:- water [water@tnt-9-203.tscnet.net] has joined #tunes -:- hcf [nef@me-portland-us222.javanet.com] has joined #tunes -:- SignOff water: #TUNES (Ping timeout for water[tnt-9-203.tscnet.net]) -:- water_ [water@tnt-9-221.tscnet.net] has joined #tunes grr -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us222.javanet.com]) -:- water_ is now known as water hehe 12:50am -:- _ruiner_ [DIY@ppp008.wi.centurytel.net] has joined #tunes -:- SignOff _ruiner_: #TUNES (destroy what destroys you) -:- smoke [smoke@16dyn229.delft.casema.net] has joined #tunes -:- SignOff smoke: #TUNES (Ping timeout for smoke[16dyn229.delft.casema.net]) -:- water [water@tnt-9-221.tscnet.net] has left #tunes [] -:- smoke [smoke@16dyn229.delft.casema.net] has joined #tunes -:- Fare [rideaufr@esmeralda.enst.fr] has joined #Tunes -:- SignOff thomas: #TUNES (Ping timeout for thomas[193.217.63.152]) -:- rares [rares@wtrb-sh3-port31.snet.net] has joined #tunes -:- hcf [nef@me-portland-us715.javanet.com] has joined #tunes -:- SignOff hcf: #TUNES (Leaving) -:- hcf [nef@me-portland-us715.javanet.com] has joined #tunes -:- eihrul [lee@216.24.141.154] has joined #tunes hoy eihrul 07:10am lo get much reading done? unfortunately, no... fell asleep at 7 PM, very lame 07:20am -:- Fufie [stig@tunnel-44-33.vpn.uib.no] has joined #tunes lo-p 07:40am hi 7? ewww 08:10am yes 08:20am y 08:30am -:- thomas [thomas@193.217.63.152] has joined #tunes ? 08:40am well I gotta go anyhow -:- SignOff rares: #TUNES (KVirc 1.0.0 Millennium BETA 3) 08:50am -:- water [water@tnt-10-135.tscnet.net] has joined #tunes good morning mornin water 09:00am * water/#tunes sees that usenet programming language discussions look for answers to the things he addresses in Slate so do the squeak fellows hm: it would be interesting to use the language tutorial itself to encourage people to refactor by showing how easy it is 09:10am it would be really nice if more people would try smalltalk when they think they don't like oop but smallralk (& self) are just silly languages :) 09:20am -:- SignOff smoke: #TUNES (Ping timeout for smoke[16dyn229.delft.casema.net]) hey eih what were you reading last night that bored you so? -:- smoke [smoke@15dyn69.delft.casema.net] has joined #tunes actually, i fell asleep soon after i left before i ever got to the book heh been very tired the last few days no idea why (i.e. i fall asleep like 6 hours before usual) dehydration? hmmm, never been a problem before well anyway, what issues can we work on with slate? concurrency ok i assume you have thoughts on it. go ahead 10:10am or not? :) not any right now but what have you thought up so far? unfortunately, i'm no expert on using concurrency iow i'm not sure about costs/benefits of various ideas it'd be nice if there were some sort of survey of the subject but the basic stuff seems pretty simple: i'm sure there are on cora yeah but i'm obviously not creative enough with the search keys anyway, we need a forking primitive of course, and we can use oop to generalize its facilities i'l see what i can do however, our evaluation strategy is lazy, so it's a bit strange what an object does when forked hcf: thx man water: were the papers on concurrency i gave u links to not enuf? water: though, how are threads reified? :) hcf: not for a general survey -:- rares [rares@wtrb-sh3-port49.snet.net] has joined #tunes hcf: you know, compare/contrast of methods or frameworks hcf: like a high-level tutorial X suX eih: as streams of computations perhaps? would be interested in looking at one myself possibly water: have u used concurrency? hcf: not extensively... just in the one game engine i was playing with a few years ago 10:20am well anyway, it seems natural to have an object representing the thread as a computation's namespace which would allow us to pass values in a distributed setting using the same namespace idea however, this also seems like an overloading of features into the namespace system, to an extent what would some methods/frameworks be named that one should expect to in a concurrency survey? water: can't be any worse than p-objects s/in/be in/ hcf: hm passive/active data models :) expect to in a concurrency hcf: maybe there aren't that many rares: expect what? hcf corrected you rares: s/you/himself/ oh i'm braindead today I've been thinking hm semaphores, exceptions, mutexes, constraint systems, pre-emption, scheduling, ... how about expected proglangs? hm cool, distributed foo, concurrent foo, ...? synchronization yes, and propagation 10:30am -:- rares is now known as Blitzenstrudel yuck, now to my least favorite topic, active vs passive objects eih: do you see this as unresolved in slate? synchronization is a performance drain blitz: it can be, but i don't think it has to be perhaps the role threads play in slate is very vague :) well i see the active/passive argument as obfuscating eihrul and I had a headache dealing with synch awhile ago while working on emk imo, the argument should instead be about how to handle the causal creation of objects in a distributed system, there will of course be multiple sources of objects gah if we only had a collapsible tunnelable self-optimizing object system um sure nah, synchronization was the easy part reality has that virtual reality doesn't ahem * Blitzenstrudel/#tunes stops the train before it derails i can imagine treating i/o streams as evaluator-driven objects -:- Blitzenstrudel is now known as rares -:- SignOff hcf: #TUNES (Leaving) even though they're "active", their activity protocol is determined by their programming -:- hcf [nef@me-portland-us1036.javanet.com] has joined #tunes (or even non-determined :) but i suppose that just pushes the problem back a level because you still have some data source that your programmed i/o object handles (yes/no?) sure 10:40am although you could look at those objects as being forwarded by another evaluator, i suppose hm no, that's too brittle a model -:- SignOff smoke: #TUNES (Ping timeout for smoke[15dyn69.delft.casema.net]) i doubt mouse events would ffit into that idea hcf: those papers you handed me raised issues, but certainly not enough or certainly not well enough to help formulate an answer off-target (still searching...) but perhaps useful (iow, ICUC): http://www-i2.informatik.rwth-aachen.de/Research/MCS/ http://www-formal.stanford.edu/clt/ http://web.yl.is.s.u-tokyo.ac.jp/pl/meta.html (worth a revisit) hm i suppose graphs are just fine for modelling distributed / concurrent computations but i'm not sure how they'll handle i/o or incomplete / lagged information-sharing -:- smoke [smoke@15dyn69.delft.casema.net] has joined #tunes hm i/o and lagged info sharing sound like they could be unified conceptually 10:50am i/o (etc) related papers, ftp://ftp.cl.cam.ac.uk/papers/adg/index.html mmm good stuff i thot so ;) now to find out just how good * eihrul/#tunes goes off to read. -:- SignOff eihrul: #TUNES ([x]chat) does common lisp have some sort of seek() option? i want to read a random word from /usr/dict/words, but use (read-line) takes about 1 minute to traverse the file :( 11:00am sounds weird what I do when reading large files is to read in larger chunks first getting file size, allocating a buffer and reading it all in a swoop and then reading lines from memory * rares/#tunes wonders about traversing databases randomly is it even possible? line-based reading of several megabytes then take very few milliseconds FILE-POSITION ? thx hcf fufie: ah! great idea, thanks smoke: if you want example code I can give you a pointer he said pointer fufie: i'm having a lot of trouble running hangman.lsp on gcl btw, is that presumably my fault or gcl's? fufie: please http://sds.yi.org/cgi-bin/cvsweb.cgi/sds/src/clisp/base/Attic/xml-base.lisp?rev=1.1.2.5&content-type=text/x-cvsweb-markup&only_with_tag=UpHillDev see EAT-FILE gcl can be odd.. have you tried with clisp or cmucl or acl? clisp and cmucl both work fine. i've tried acl, but the binary didn't work, and i didn't feel much for building from source ok.. use clisp and cmucl then.. grin. 11:10am hm finally a good description of how haskell monads are actually used (and why they picked those weird syntactic symbols for monad operators) 11:20am weird, just got a rash of thesis proposal hits on concurrency? w/ such as a search param, which mean anything s/which/which could/ i tossed in mut* and immut* hm haskell monad syntax winds up being very procedural-looking 11:30am what ever happened to Vadim Antonov's Grail? tunes-rev's link: http://www.alink.net/~avg/ is dead * water/#tunes shrugs ask Fare 11:40am http://www.cora.justresearch.com/cgi-bin/details.cgi?id=57256&from=hier_html.cgi k er sorry, mna, i already have this one figures heh i'm still not done grokking it, though 11:50am http://www.cora.justresearch.com/cgi-bin/details.cgi?id=6080&from=hier_html.cgi heh got that one too yeah i did a lot of trolling through research recently did u go thru the relevent parts of cora? i think i covered it it's not nearly as comprehensive as i've found researchindex to be, or arxiv for that matter hm no it seems concurrency brings up some interesting results yeah, i tend not to bother w/ cora whats arxiv? it's a math/science archive that has a mailing list 12:00pm www.arxiv.org occasionally some really cool stuff gets posted, but mostly just TR's http://x34.deja.com/=dnc/[ST_rn=ps]/getdoc.xp?AN=431234632&CONTEXT=951594951.997457967&hitnum=13 http://www.cse.ogi.edu/~jl/ 12:10pm -:- rares [rares@wtrb-sh3-port49.snet.net] has left #tunes [Segmentation] "blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah blah" :P i just love it when the best research in a n area is nothing but a lot of shots in the dark 12:40pm damn i finally re-located that paper on rewrite for memory management strategies oh? url? abi: rewriting for gc is at http://www.cs.rice.edu/CS/PLT/Publications/fpca95-mfh.ps.gz cora refers to it abi: rewriting for gc is also see cora okay, water. 12:50pm -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us1036.javanet.com]) -:- hcf [nef@me-portland-us1036.javanet.com] has joined #tunes -:- SignOff Fufie: #TUNES (Ping timeout for Fufie[tunnel-44-33.vpn.uib.no]) http://www.cora.justresearch.com/cgi-bin/details.cgi?id=361809&session=847302&from=cora_query.cgi&sara= 01:10pm cora is pretty damn slow in loading my entire net experience is slow right now cora just makes it seem worse hm hm aczel is also a hyperset (co-induction / bisimulation) guy the name seemed familiar i'll bet the radius of the cs research community as at most 3 s/as/graph is/ hmm, RI turns 'of' into w/2 http://citeseer.nj.nec.com/details/chen97rolling.html k 01:20pm i don't suppose you have a wishlist or idea about concurrency for slate? no no?!? i'm leaving it to u gee thx so you've never thought about the idea at all ever used a threads package or something? 01:30pm thot about it, havnt used a threads pkg concurrency mailing list @ http://www.cwi.nl/~bertl/concurrency/concurrency.html "hey, why don't you look that up in these academic journal volumes" ;) hm i suppose i could subscribe and ask semi-newbie questions 01:40pm yikes 750 members nah 01:50pm hm. as always the best papers are the newest of course, you can't grok 'em unless you read a pile of predecessors -:- eihrul [lee@usr5-ppp186.lvdi.net] has joined #tunes or at least the "blood-line" of papers wb 02:00pm water: so rmm, it really could be dehydration, eh? yes which could be due to many things diet, caffeine, smoking, ... i have to drink a lot at work on the ship just to avoid problems just about everything dehydrates you except for fairly pure water -:- Kaufmann [newbie@200.255.108.118] has joined #tunes -:- Kaufmann is now known as Kaufmann_theMP3Addict heh hullo kauf: have you ever considered how tunes would handle concurrency? water, yes hm it seems i've found work on a lazy functional deterministic concurrency core kauf: care to share your thoughts on it? * eihrul/#tunes reads the dehydration faq. back eih: there *is* one? eih: where? well, I assume Faré (was it Faré?) is still backing his idea of on-the-fly compilation and execution as the exclusive way of getting stuff done on Tunes, right? sure i suppose water: hmmm, it only had one question, but it was titled a faq anyway i like the idea as well eih: heh. probably not the real thing, then ah there's the rest 02:10pm *ahem* url, please kauf: continue, sorry no prob water: www.rehydrate.org well, then I guess the natural thing to do would be to have one processor run as manager, compiling portions of modules and having the other processes execute them (or different parts of the same module that are declared to be concurrent); sync points and the like could simply mean that the manager would compile only a part of the module and then have the compiled code generate an appropriate interrupt that doesn't really make much sense hm yeah that needs some work, i think then again, I don't really like the idea of centralised management it's the anarchist in me talking :) I don't see how much this manager scheme would differ from a traditional run-time kernel in a single processor machine hm hm? yes, ala self i think if you change the term "manager" to "service", you might get a better description which has a considerably large kernel of C++ code needed to manage everything it doesn't really scale to arbitrary distributive schemes, though water, but I figured the idea was to eliminate the need for a run-time kernel well do you mean "kernel" as in hardware-protected kernel space or something more abstract? water, I figure the solution is do add synchronisation features in the manager itself yeah well that's the idea of a kernel :) 02:20pm there's the apertos approach which considers it not necessarily a kernel but a particular meta-space "hm that feature's tough to handle co-operatively, so let's put it in the manager(kernel)" no, I mean sync features as in distribution... the coordination is done amongst managers, which then each divide their part of the code to be executed (or the problem to be computed, or whatever) between the processors they manage that word "manager" really bugs me what kind of object would it be? water, by kernel I mean a portion of code which has authority above all others, and which can do things other pieces of code can't do, and which is called at run-time by "normal" pieces of code to do those things they can't do by themselves kauf: shorter blurbs please uh? 7 lines of text at a time makes the conversation rough seven lines of text? grrr maybe if you scaled down, perhaps to 18-point text... :) -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us1036.javanet.com]) -:- water [water@tnt-10-135.tscnet.net] has left #tunes [] wow, I really pissed him off, eh? anyway, I'm off. -:- SignOff Kaufmann_theMP3Addict: #TUNES (A stranger is just an asshole you haven't met yet.) -:- water [water@tnt-10-135.tscnet.net] has joined #tunes ? -:- hcf [nef@me-portland-us1036.javanet.com] has joined #tunes would you care to precede that punctuation mark with a predicate? 02:30pm "BriX - A Deterministic Concurrent Functional X Windows System" :) '95, even trademarked? just a research program where at? dunno, there's a paper on it in a bibliography of this paper -:- SignOff hcf: #TUNES (Leaving) btw, slate's call-by-need, isn't it? -:- hcf [nef@me-portland-us124.javanet.com] has joined #tunes (being lazy and all) www.cs.bris.ac.uk/~ian/Functional/brix.html yep. that's it 02:40pm -:- Fufie [stig@tunnel-44-38.vpn.uib.no] has joined #tunes so you're going to ignore yet another of my questions, eh? thinking k (call-by-need) 02:50pm . -:- mibin [mibin@an1-839.tiscalinet.it] has joined #tunes -:- SignOff hcf: #TUNES (Leaving) hey you should probably read this paper oh i almost forgot abi: rewrite for gc doh abi: rewriting for gc rumour has it rewriting for gc is "Abstract Models of Memory Management" at http://www.cs.rice.edu/CS/PLT/Publications/fpca95-mfh.ps.gz or see cora 03:00pm i still haven't really found a good survey of concurrency, but then i suppose it's a wide-open subject damn it, i really need to get this issue resolved so we can move on with the language -:- rares [rares@wtrb-sh3-port20.snet.net] has joined #tunes 03:10pm perhaps the namespace system should model memory spaces literally similar to /proc with the ideas of that gc paper on rewrite, we could also include a virtual namespace for the low-level model of the memory spaces bbiaf 03:20pm http://slashdot.org/article.pl?sid=00/02/26/1426228&mode=thread Hardware based garbage collection yes it's been done b4 of course, it was in a propretary lab again bbiaf 03:30pm why do they feel thy have to patent the littlest screw in a machine -:- SignOff smoke: #TUNES (z) -:- SignOff rares: #TUNES (KVirc 1.0.0 Millennium BETA 3) 03:50pm eihrul: are you here? 04:00pm oh great someone is *seriously* reading my paper on arrow 04:10pm someone who works for a company that i'd wager 100 to 1 would botch the idea up completely -:- SignOff eihrul: #TUNES (Excess Flood) -:- eihrul [lee@usr5-ppp186.lvdi.net] has joined #tunes dude eihrul: are you there? yeah however transiently did i tell you yesterday about the email i got? some guy read the arrow philosophy paper and really got fired up about it and wrote me nope well, i checked him out, and here's where he works: http://www.intelligenesis.net/ i'll bet you 50-to-1 that guy is telling his boss about the ideas in my paper is that bad? there's no way they'll beat me to making arrow is that not why you published the paper? yes, but still it gives some sort of concreteness to what i'm doing i have to admit looking at their whitepaper makes me feel a little nervous the concepts of their product are similar to arrow on a certain level 04:20pm not to mention their page is really hard to reed in 1600x1200 resolution heh yeah that sort of reason keeps me at 1280x1024 -:- Kaufmann [newbie@dial850.infolink.com.br] has joined #tunes -:- water has changed the topic on channel #tunes to: TUNES, Free Reflective Computing System @ http://www.tunes.org || slate @ http://www.tunes.org/~water/slate-home.html || WebMind @ http://www.intelligenesis.net/ I saw something quite interesting *you* did? yes believe it or not :) sorry, i'm completely blown away by my own little realization continue Oh no. You know I can't. Now you've got me all curious about your own little realisation. You go first. look at that WebMind thing yesterday some guy wrote me praising my paper on arrow today i do a little digging on him and found out where he works i seem to have a competitor or something or maybe a new employer :) -:- SignOff Kaufmann: #TUNES (Ping timeout for Kaufmann[dial850.infolink.com.br]) * water/#tunes checks out their CTO's research in cognitive science -:- hcf [nef@me-portland-us1040.javanet.com] has joined #tunes hcf!!!!! look at the 3rd url in the topic 04:30pm this guy who wrote me yesterday praising my paper works at that company but did he understand ur paper? i'm not sure water: maybe you can get a real job now :) but if he's working on that web-mind product, and he likes my idea over their internal theory, then that must mean something the military is not for you Indeed, foof :) * water/#tunes can barely contain himself how long has the military right to keep you behind bars.. can you leave whenever you want to? http://goertzel.org/ben/ he's their CTO if someone posts enough money (roughly half a million), then i'm free actually i've heard it's a lot less depending on circumstances otherwise i must wait 2 years more sheesh.. that sounds draconic http://goertzel.org/ben/ben_research.html that's the critical page he doesn't get half of what i've been working on, i'm guessing so far yes it seems so, although he wouldn't mind my work's philosophy, it seems wow this guy publishes no more coherent stuff than i do :) 04:40pm hmm the whitepaper is at least a year old... i need to contact these people somehow -:- Kaufmann [newbie@dial457.infolink.com.br] has joined #tunes Hey again can anyone suggest a good strategy for approaching these guys? So water, what was your own little realisation again? kauf: check logs okie then i'm thinking that i could write to either the fellow who wrote me or the CTO (original researcher and author) or perhaps both Someone oughta add a feature to abi that searches in the logs for words or people -:- Kaufmann has changed the topic on channel #tunes to: TUNES, Free Reflective Computing System @ http://www.tunes.org || slate @ http://www.tunes.org/~water/slate-home.html || WebMind @ http://www.intelligenesis.net/ 04:50pm -:- water_ [water@tnt-9-220.tscnet.net] has joined #tunes So I saw on Freshmeat a program called xcruise, which is a file manager for X11 with a twist. lol they're using kqml for agent protocols!! does anyone know of any x11 file managers which use ssh for remote sites instead of ftp as most do? uh? kaufie: what's the twist? Kaufmann: theres 17meg of logs -:- SignOff water: #TUNES (Ping timeout for water[tnt-10-135.tscnet.net]) -:- water_ is now known as water hcf, yeah, I know... but it'd be cool nonetheless :) ok i'm way ahead of them on a lot of technical points -:- air [brand@p0wer.qzx.com] has joined #tunes Fufie, instead of the usual GUI, it presents the filesystem as an universe, where directories are galaxies, files are planets and symlinks are wormholes whatever to say nothing of the fact that it's coded in java Kaufmann: what does? that guy who wrote me had to have figured out how my idea was better than theirs air, xcruise hcf: i'm sending you the email what do you think? Kaufmann: where is that at? i know it doesn't say much i'd bet this guy is some lowly coder at the company air, now I'll have to find the original URL... wait a sec Hmmm, I can see it already... Intelligenesis Arrow (c) LOL For Windows 2000, even man, i don't even give a s*** any more about #tunes noise no, multiple-os release based on a portable slate implementation #tunes noise? air, http://apps.freshmeat.net/homepage/947746907/ water: u blew his mind hcf: it seems pretty obvious that this guy will tell his manager or even the CTO, no? 05:00pm and you know for a fact that he's read the CTO's research papers not totally obvious hm perhaps webmind was easy money for 'em and they wanna keep it easy not necessarily isn't Billye Boy CTO in this company in Redmond? I am sure he writes cool research papers :) i assume selfishness