IRC log started Mon May 22 00:00:02 2000 [msg(TUNES)] permlog 2000.0522 * water/#tunes contemplates what "/ /" could possibly mean in slate anyone have any insight? :) if "/ /" turns out to have no intelligible meaning, then that would be a definite point in my design debate 12:10am -:- SignOff hcf: #TUNES (Leaving) * water/#tunes trolls RI for some papers that might help well, no one's learned more than eihrul and i have, apparently, about the subject although some ppl got close in ETYMA 12:50am -:- SignOff smoke: #TUNES (Ping timeout for smoke[16dyn179.delft.casema.net]) abi: etyma is a system for using recursive namespaces for inheritance and composition at http://citeseer.nj.nec.com/banavar96application.html 01:00am -:- SignOff water: #TUNES (The Tao went that-a-way!) -:- Kyle_L [kyle@cr168790-a.nmkt1.on.wave.home.com] has joined #Tunes -:- SignOff Kyle_L: #TUNES (Leaving) -:- eihrul [lee@usr5-ppp198.lvdi.net] has joined #tunes -:- SignOff eihrul: #TUNES ([x]chat) -:- Ghyll [nobody@firewall.prosalg.no] has joined #tunes -:- water [water@tnt-9-251.tscnet.net] has joined #tunes * water/#tunes rejoices that his javascript now works congrats. javascript is the devil. * FufieAFK/#tunes makes the sign of the cross well, certain sites, even opensource ones, demand js 07:30am which reminds me, there's that bootloader that i wanted to download what bootloader? i forgot the name it's been so long did you think about / /? yep ? no good ideas just yet, not ones that anyone would understand, at least hmph ;) 07:40am * coreyr/#tunes stops waiting... waiting for what? for you to go into it -:- SignOff Ghyll: #TUNES (Read error to Ghyll[firewall.prosalg.no]: EOF from client) 07:50am http://www.xosl.org/ -:- NetSplit: irc.linux.com split from lewis.openprojects.net [07:56am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [irc.linux.com] -:- Netjoined: irc.linux.com lewis.openprojects.net -:- FufieAFK [stig@tunnel-44-54.vpn.uib.no] has joined #tunes -:- coreyr [coreyr@net255ip95.parklink.com] has joined #tunes -:- abi [nef@bespin.dhs.org] has joined #tunes -:- I440r [mark4@purplecoder.com] has joined #tunes -:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes -:- thomas [thomas@193.217.63.152] has joined #tunes -:- water [water@tnt-9-251.tscnet.net] has joined #tunes 08:00am hmm hm FFI support in squeak for X11 08:10am * water/#tunes goes to read -:- hcf [nef@207-172-225-114.s114.tnt1.pld.me.dialup.rcn.com] has joined #tunes -:- Kyle_L [kyle@cr168790-a.nmkt1.on.wave.home.com] has joined #Tunes hey kyle h i 09:20am -:- SignOff FufieAFK: #TUNES (Leaving) -:- Fare [rideaufr@lantier.enst.fr] has joined #Tunes ` hi ' lo any update in the Slate docs? i fixed what we were discussing earlier, as well as typographical stuff right now, i'm still making sure i've thought out all the language issues before i explain something that i'd have to revise later and elric has been too damned busy to discuss web design and architecture work 09:30am hum. Can you help me design some low-level stuff? (or a bit higher-level, if you prefer) for what? the web site or the language/compiler issues? language/compiler sure Fare: whatcha calling ur beast? well. basile and I tried to agree on a bootstrap architecture hcf: calling ur beast? Fare: whats the name of ur compiler/lang/thing I have laid down a plan, but a crucial high-level component is lacking specification, and a low-level component isn't so well-understood either the low-level one is easier to fix hcf: let the lang to be named "kluge" wow. and i thought *i* was secretive the low-level lang is i386 assembly 09:50am I'm considering the design of an assembler and am wondering what the "right thing" is to properly factor things, etc i'm dealing with that issue as well (especially with regard to variable-length instruction resolution, assembly segment-merging, symbol handling, etc) (of course Fare's not listening) kluge was intended to be some kind of lisp/ml dialect, but I'm lost in the precise design water: which issue? refactoring and the "right thing" to do hm lisp/ml mixture well apparently we have two separate designs now, though i doubt Fare's would exist if mine didn't I've still got no fully coherent design yet; the difficult thing is still meaningfully integrating linearity in it all i'm dealing with similar issues I'm having a hierarchy of linear allocation zones/resource pools the way i'm currently handling linearity is the way i handle all reference-handling... with namespaces and meta-objects (actually, a tree that need not be statically rooted) that's similar to my idea yes, each of these zone is pretty much one of your namespaces, except that there isn't a worldwide "pointer/reference/name/symbol/foo" type there isn't one in mine either, Fare good then there comes the problem of locking and concurrency, etc oh thank you soooo much for approving linear resource holding points are a generalization for locks -- valued locks oh really? :) they do the interface between linear and non-linear computations maybe you can tell me how to write a parser, too! :) 10:00am maybe i should be more explicit here.... looks like the right thing to do is for them to hold a reference to either the current value (if available) or the holding thread (if not available yet) -- which also kind of implements futures and (delay ...) -:- smoke [smoke@16dyn179.delft.casema.net] has joined #tunes In TOOL, linearity, is the specializing of a context. If there is an 'object' context, then giving it a new name or saying that it is of a specific type, is considered a specialization of the context relative to what is was initially. thomas! heh forget it... you're lost in your own little world water! water: sure hey tom fare: i find it insulting that you insist on a separate design effort I don't insist on it I'm just unsatisfied with the current slate docs well then talk to me I'm talking to you i'm the one with the info and knowledge, not what i put online to help explain things to newbies no you're not of course I am you don't talk, you lecture you're the phd baby, and i'm the amateur * coreyr/#tunes hmms. you're just listening with bad a prioris well, you've set a really bad precedent in the past for not listening to anyone you're the one with an inferiority complex heh * water/#tunes considers inverting that statement does "your lost in your little world" classify as 'listening', nister water? :| of course not, i got pissed off you get pissed off easily. the fact that Fare just sat on his ass for 5 years and only when i'm writing up something as an HLL proposal that he doesn't like because it's not the way he writes, he goes out and starts his own little campaign water: uh? water: I think you're drunk or so heh no, i'm angered that you're not being up-front I'm still doing the same thing (or non-thing) as usual oh ok so this is just another smokescreen of ours? s/ours/yours/ on irc it is really easy to misunderstand others, just talk tunes. I'm not any more jealous of slate than of retro, sphere, or any other thing that popped up in the tunes list. fare doesn't talk tunes, he talks politics I'm looking at it with interest (more so than other things, actually) (I like retro, too) (I'd like clementine, if freely available) so why are you discussing things with basile separately and not being open about slate and those ideas? I too, do not agree with everything said here, and don't allways regard everyone here as as intelligent as they see them self. And sometimes people talk about things that I find booring or irrelevant - things I don't care to read up on. But I don't walk around bragging my ignorance! (actually, the same kind of discussion pops up often enough with water) 10:10am fare, I know! I was interested in TML/TOOL, too, except for the string-based aspects. I know I'm just trying to not embrace passionately anything, least I'm sure it is or can become what I desire water: because basile invited me at his house water: you're invited too, if you come to France. well i'm not overhauling my web docs until i have a better framework for laying out the concepts than what tunes has right now (cvs is the poor man's web management tool) water: or maybe I should forbid you to ever contact Tril privately, although you're both in .wa.us ? whatever yeah! lets get paranoid! that's not what i meant this is silly indeed tunes has been silly for the last 5 years, though the documentation is a joke water: improve it you have cvs access i've been reading fare's stuff every god damned month for the last 5 years overwrite it screw that leave, split, take off i have enough problems dealing with slate docs no, i'm saying i refuse to directly handle the web site then don't you grumble this way there's absolutely no task management in tunes at all stop demanding just do it stop denying I don't deny Fare: why dont u just fucking do it hcf: I don't, that's all you've had 5 years to improve the tunes explanations hcf: you have no right to demand anything from me; I have no right to demand anything from you water: so have you but i never cared about tunes ok, so why complain? until a few months ago when i decided to try to satisfy the spec to get the tunes project to shut up great I applause and await with interest but i have zero experience with command-line utilities and web tools and no time to deal with them... i work 11 hour days and commute 1.5 hours a trip 10:20am water: i would suggest you'd look into command line utilities; i've heard you say some dramatic things about them there has to be separation of tasks for reasons like this and other factors like the need to research heavily to find the right ideas for tunes *** DCC CHAT connection to Kyle_L lost: Connection refused bah water: sure. so what? btw, if given USD 3M, what would you do about it all? fsck you to then given what? 3 million dollars irrelevant... it won't happen raise more money with it Fare: for how long has there been a tunes website? if it happens via VC (most likely scenario), then they've bought you... it has strings attached, and they'll make you pay 5.5 years afaik i mean when it become a group thing s/come/came/ it arose from Moose Fare split from their project under the leadership of people who are now long gone in 5.5yrs, no full time webguy was ever gotten yep i'm starting to think about paying someone myself to do the work water: how long did u have to look for elric? i just ran into him i already knew web coders, but most of them didn't grok tunes principles is there any end in sight to his issues which prevent his use? well, he claims his stress load will be reduced starting today since they're hiring someone else to do the work alongside him 10:30am heh now Fare is doing his classic "pretend i'm not afk and maybe people won't notice that i just read scroll-buffers" routine * Fare/#Tunes is back from a phone call go ahead, read the scroll-buffer! :) people I know propose that I try raise VC fund, and propose me at least USD 3M, which is why I ask VC is dirty money wrt tunes why dirty? As long as the business plan specifies that it's free software, that the CVS anonymous read access is granted from day one with GPL or such license, what's wrong? if a university won't support the process because it's too abstract, then VC will just pretend to support the idea on the basis that they don't understand it fare: what about patents? it's not about GPL, it's about tunes ok, so what to do before to try attract VC? formalize, perhaps? smoke: good remark; should be included in business plan, too water: sure which hasn't happened in 5 years first thing to do with 3M, plan a meeting to work out personality kinks. :) :) it could have happened 2.5 years ago completely personally, I'm still trying to PhD before to try anything i could hire people part-time even on my salary and the best businesses can start in garages water: I couldn't; unless they are thrid-country hackers; but I can't reliably communicate with one i don't mean hackers coders (the ordinary kind) would just get in the way unless they only did web-dev anyway, I wanted to talk about assembler design or more precisely, assembler back-end design any taker? well i've successfully killed off hll discussion (as if what Fare was doing was discussion at all) heh water: want to discuss hll? ok; I can, too my problem was with wanting a symmetrical syntax for constructing and deconstructing statements ejemplo? 10:40am in other words, you want CONS-cells? ... ouch, bad copy/paste I would like pattern transformers be such that you can invert them just by reversing the order of the deconstructing and constructing patterns. is a pattern just another word for a list, though? what's the formal difference? for instance, in the assembler, I have pattern converters like (ENTER (IMM16) (IMM8)) (#xC8 (imm16) (imm8)) (ouch, copy/pasting tab into my ircii setup breaks havoc loose) (yes i know what a pattern is, i just want to make sure we mean exactly the same thing) bleh the problem with this approach is that it requires separate declaration of bindable variables well i know there are great benefits of pattern matching, but what's the particular relevance? * water/#tunes nods declarations of variables are terrible I was just trying to build a coherent pattern language out of such a consideration what for? water: declarations could be helped by the interactive programming system. maybe i'd rather the declarations not be needed in the first place water: or there could be syntactical sorting or defaults as in prolog/commonlisp/Ocaml/fortran (ie don't solve the problem, eliminate it from arising) too complex... is this supposed to be built in to the language? the system i was considering involved using explicit casting of expressions into different typespaces and having functions operate on those mirrors of the original objects (i don't treat expressions as special to the language) er... expression types not sure what you mean, now hm 10:50am the expressions that are fed into the evaluator are just ordinary objects... there's no inherent basis on text or literals i hate this... i had an explanation i was working on earlier that escapes me now take for instance the system i'm using for the natural numbers ok, just like lisp sexp instead of text; just that you have more structured stuff than sexp they don't have representations built in to the system.... those representation primitives plug in separately sure in my case, quoting is just a function that casts between object types and is not a unique or static notion how does that affect the semantics of the base language (ie outside of macro/extension processing) ? well, it results in the odd properties of slate, some of which are described in the tutorial i haven't fully debugged the idea, though.... namespace access hinges on it hinges on it? right now, i'm looking at generalizing namespace access so that it's co-recursive (to support graph structures e.g. ".." access) the tutorial appears not on the -home page it's accessed from the news page for now it's not something i'm advertising yet are your namespaces linear? hm (I suspect they are somehow, since stateful) yes the way you duplicate and share references (retrieved by names) is through cloning linear namespace? and cloning often produces objects whose effects don't register in the original register? kyle: where there's only one way to access any object s/register/mute/ thanks 11:00am I guess my question about linearity was that direct linear pointing yields non-circular datastructures; hence, directly-embedded sub-namespaces yield a hierarchical structure no one can help me with assembly backend? in fact, this cloning happens on method invocation in slate, since binding arguments mutes the method's state yup, that's a critical point in any language fare: most of the assembler programmers don't hang around after i've been discussing slate the fact that function invocation generally implies a non-linear effect. well in slate's case it produces a pseudo-anonymous slot I'm addressing assembler metaprogrammers, anyway oh I'm wondering if we can factor function invocation into explicitly separated linear and non-linear effects although i'm considering having the slot be accessed by the expression text, which is of course not the best idea, but i'm working on it maybe i.e. an assembly-translated function call usually involves allocation (non-linear) and having the PC point to a permanent (non-linear) object. well, from outside the namespace, an object can be referred to from anywhere by sending the selector message to the namespace hm my question is non-gratuitous: it's about allowing users to define new invocation semantics. in slate, i'd simply describe the memory architecture as a mutable object with immutable structure and also about being able to express arbitrary invocation semantics for the sake of e.g. interfacing with foreign subsystems, or doing real-time stuff, or taking advantage of non-generic hardware of some embedded processors, etc you could overlay other abstractions as long as you had ways to programmatically embed them with morphisms ok, but you're talking in too general a scope of application now (for the purpose of discussion) I'm trying to draw a direct link between high-level and low-level i know but you're talking about complicated stuff... like things that require months of work 11:10am even withotu these months of work, I'd like a framework that allows for these later months of work grrrr no kidding i know exactly what you want, i've read the damned tunes docs my question about linear namespaces as also that .. was only well-defined in such a tree then why grrrr? because you missed my point you can define new invocation semantics by, for example, wrapping object types with other object types dynamically uh? there's also dynamic switching of meta-behavior haven't you heard of object wrappers? you mean meta-object wrappers, don't you? heh ok there's a confusion of terms in clos, everything related to the original object that is not the original object is called a meta-object there's another research initiative which has a particular designation for the term meta-object 61453 time actually there are several 61453=? look on the ./slate-references page please to find out what i mean by meta-objects (format t "~x" 61453) ok wtf does that mean? * Fare/#Tunes is away (format t "~x" 61453) for a guy who supposedly has great ideas about user-interfaces, he's a complete hypocrit 11:20am * water/#tunes looks for a compiler book to complement acd&i water: he's off for diner oh how does 61453==dinner? yes what's his code system? and why the hell does he use codes at all? and whats the point if the ppl listening to it dont know it (format t "~x" number) is the common lisp equivalent of C's printf("%x\n",number); so he can criticize us for not knowing it ;) perhaps its a "hehe, i'm frustrating water w/ my codes" thing oic so it's numbers as ascii actually it would be sprintf() but this is already way offtopic wow that's so hypocritical that it deserves no comment :-) sprintf or it's CL equivalent has to be the most un-tunes-like programming feature that i've ever seen s/it's/its/ format is actually quite practical practical? yes it's only portable because everyone and their half-brother's ox uses the exact same format hm well, i meant to say that it's an ugly function (it can do very non-trivial things like print a number in the Roman number system), but if you happen to need such a particularly bloated function you only need to 'find' it instead of 'code' it uh roman numeral printing is pretty trivial... at least it is in smalltalk well, i guess i'm just used to smalltalk now, so n/m 11:40am aha but otoh smalltalk *is* more like tunes than c (or CL in that respect) perhaps it would've been better for common lisp if there would be a clear seperation between library functionality and the actual language perhaps 11:50am hohum tell me about it 12:50pm -:- hcf_ [nef@207-172-225-87.s87.tnt1.pld.me.dialup.rcn.com] has joined #tunes -:- SignOff hcf: #TUNES (Ping timeout for hcf[207-172-225-114.s114.tnt1.pld.me.dialup.rcn.com]) -:- hcf_ is now known as hcf -:- ult [ult@user-38lc646.dialup.mindspring.com] has joined #Tunes >>> hcf [nef@207-172-225-87.s87.tnt1.pld.me.dialup.rcn.com] requested PING 959028214 from TUNES -:- SignOff abi: #TUNES (Ping timeout for abi[bespin.dhs.org]) -:- SignOff smkl: #TUNES (Ping timeout for smkl[glubimox.yok.utu.fi]) -:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes -:- SignOff hcf: #TUNES (Ping timeout for hcf[207-172-225-87.s87.tnt1.pld.me.dialup.rcn.com]) -:- hcf [nef@207-172-225-87.s87.tnt1.pld.me.dialup.rcn.com] has joined #tunes -:- SignOff Kyle_L: #TUNES (Ping timeout for Kyle_L[cr168790-a.nmkt1.on.wave.home.com]) -:- ``water [water@tnt-9-251.tscnet.net] has joined #tunes -:- SignOff Fare: #TUNES (Ping timeout for Fare[lantier.enst.fr]) -:- SignOff water: #TUNES (Read error to water[tnt-9-251.tscnet.net]: Connection reset by peer) how about doing tunesdot? <``water> hm -:- ``water is now known as water theres sort of quasi-netsplit occuring right now s/sort/some sort/ -:- SignOff smkl: #TUNES (Ping timeout for smkl[glubimox.yok.utu.fi]) yeah that's like the idea i was discussing the other day about xml and unifying chat, the mlist, and web-dev no kidding :) hmm although xml isn't the greatest -:- Fare [rideaufr@lantier.enst.fr] has joined #Tunes re? 01:50pm anyway, it'd be an interesting framework to code up would smth w/ existing /.-like sw be worth it? no not the body of slash code "as is" er not specifically slashcode /.-like hm Would you then require usage of a browser to read? theres a few copycats well without the whizbang stuff... moderation could be karma+structured scripting where you can commit statements to the dev-tree from within the discussion -:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes so basically, u'd be too unhappy w/ any existing sw heh maybe sounds somewhat like fare/tril's wait for tunes cop out the bottom line is that i don't like dealing with the web coding directly iow, they didnt do the db poop earlier in tunes.org's life if that means a person as intermediary, then that's good enough for me yet for some reason, excepted cvs yep... nice and arbitrary odd * water/#tunes looks at the tunes site btw tril hasn't been with tunes all that long only 2-3 years i think -:- SignOff hcf: #TUNES (Ping timeout for hcf[207-172-225-87.s87.tnt1.pld.me.dialup.rcn.com]) lovely day, isn't it? 02:00pm -:- abi [nef@206.63.100.13] has joined #tunes -:- hcf [nef@207-172-225-87.s87.tnt1.pld.me.dialup.rcn.com] has joined #tunes hmm, ridot eh? so...? tunesdot would be like a ridot i'm meeting elric tonight, btw, whether he likes it or not ;) heh oic fare-moderation and fare-karma? er whatever * water/#tunes does a little housekeeping near the computer i dont do slashdot, dont know about moderation/karma oh 02:10pm -:- SignOff ult: #TUNES (Ping timeout for ult[user-38lc646.dialup.mindspring.com]) well what ideas do you think i should work out with elric? -:- SignOff abi: #TUNES (Ping timeout for abi[206.63.100.13]) -:- ult [ult@user-38lc646.dialup.mindspring.com] has joined #Tunes damn it not a great time to talk, i suppose i was checking the log was about to respond will the slate web docs be merged w/ tunes.org? -:- abi [nef@bespin.dhs.org] has joined #tunes using the same system, or totally separate? -:- NetSplit: forward.openprojects.net split from irc.linux.com [02:13pm] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [forward.openprojects.net] they should be merged in theory but that's a decision to be made either way, both should have the same format -:- Netjoined: forward.openprojects.net irc.linux.com -:- I440r [mark4@purplecoder.com] has joined #tunes -:- lar1 [larman@adsl-63-204-133-5.dsl.snfc21.pacbell.net] has joined #tunes hey l1440r hey lar hcf: so yes, the same system for both re water brb (vacuuming) How many words does it take to bootstrap a language? 02:20pm