IRC log started Fri Mar 17 00:00:00 2000 [msg(TUNES)] permlog 2000.0317 -:- dalvarez [ircusr@212.68.72.98] has joined #tunes * dalvarez/#tunes notices that somehow #tunes is not particular busy these days... -:- SignOff Plundis: #TUNES (Ping timeout for Plundis[130.238.23.252]) -:- Plundis [plundis@130.238.23.252] has joined #tunes -:- SignOff Plundis: #TUNES (Ping timeout for Plundis[130.238.23.252]) -:- Plundis [plundis@130.238.23.252] has joined #tunes -:- SignOff Plundis: #TUNES (Ping timeout for Plundis[130.238.23.252]) -:- smoke_ [smoke@vengeance.et.tudelft.nl] has joined #tunes hi 04:30am -:- SignOff smoke_: #TUNES (One day sheep will rule the world) -:- thomas [thomas@193.217.63.152] has joined #tunes -:- SignOff smkl: #TUNES (weekend ...) -:- SignOff dalvarez: #TUNES (Leaving) does haskell support some sort of EVAL (as in lisp) ? 07:00am -:- water [water@tnt-10-122.tscnet.net] has joined #tunes hi tom, smoke hi water did you find out that info about haskell? (your join reminds me to check out the slate pages :-) okay :) hm i wonder what you mean by `that', since i asked over 3 different things about haskell today+yesterday :) does haskell support some sort of EVAL (as in lisp) ? i just read today's logs that one being the most recent question i haven't found out yet, no :( i'm looking through the past couple of days to see who/what i've missed i believe it has one, but it's not simple dalvarez had some interesting ideas on dynamic scheduling the other day, but i don't know if he worked those out like what? from the Slate homepage i understand that slate _will_ have an EVAL feature, right? yes although i'm not sure i'll call it eval 07:10am is that the key ingredient to making a pl reflective? the faq & tutorial will definitely address that (but from within the mobius side) hm yes it helps with linguistic reflection but it's not a panacea btw, i'm definitely considering jecel's advice of multiple kinds of meta-objects but it shouldn't affect the amount of reflective power in them, just the architecture do you have a particular reason not to make slate purely functional? yes because people hate typing in "lambda" ;) haha. but seriously the type systems of functional languages are terrible (for tunes) are they? i think so they seemed like great fun to me for instance, they hide complexity of state behind function declaration and don't even get me started on boolean logic in funclangs :/ keep in mind that slate is very mucha functional language it just expresses it differently * Higher-order functional programming supported by the language itself yep 07:20am i'm fully trusting http://www.tunes.org/~water/slate-detail.html it's not pure everywhere, but you can get pureness if you need check the semantics page ah. well hidden! :) i didn't spot that link at fisrt yeah well it's where i put experimental ideas will slate be a language definition or a language implementation btw? oh good squeak is finally getting down and dirty and optimizing morphic and writing a good tutorial both i hope separate, but i'd like both keep in mind though that i'm trying to design it to minimize the size of the "system code" 07:30am does that mean the code interfacing with the OS, or the code to run the language? -:- AlonzoTG [Alonzo@216-164-138-140.s140.tnt5.lnhva.md.dialup.rcn.com] has joined #tunes to run the language hm also by extension to the os, but i haven't addressed that directly yet btw, the arithmetic / logc / text system in slate will be modular modular from the ground up, or using a common base? a lot of this isn't described yet on the site from the ground up common bases suck so even addition is an option? :) (see common lisp for details :) yes but you can reconstruct new versions easily haha. i agree that the common lisp standard defines a LOT, but i can't find good arguments for that being good or bad you simply have to interface it correctly with primitives that the implementation actually supports heh note that in order to make these systems really modular, i have to re-think the use of such systems in the core system code is `the object' the most elementary datatype in slate? (if so, isn't that a bit abstract?) hm 07:40am yes and i like it being that abstract abstract things would serve as a good base for the modularity virtually you'd have no base then well keep in mind the arithmetic / logic/ etc will be built on the logic of objects no base is the best base what logic do the objects adhere to then? basically they have a "result" slot, they are namespaces, they respond to messages, they can be made active, they are sets, ... etc and they can have meta-behavior :) they're sets? yeah they're not lists * water/#tunes cackles ah yes hum ok hehe i got scared you could take intersections and such (done a bit too much topology last week :-) yes you could topology rocks :) what'd be the use? for one, you wouldn't need to have a finite protocol for another, you could model things more directly with objects themselves yup it sure does. but i'm studying it from a mathematical point of view: i haven't the slightest notion of how to see programming languages' semantics in terms of topology oh i definitely do :) haha :) one of the reasons i got started on this language back in college was that i hated symbolic algebra systems but i realized my beef was more about how people think about programming in general you hate symbolic algebra systems? urf? :) like maple and derive and mathematica such? & yeah i used to work on them too 07:50am i loved them dearly until i discovered how they worked brute force? :) i thought at least mathematicians would have a better idea about how to runa computer than programmers would it turns out that they're just as clueless haha most mathematicians i know are complete loonies when it comes to programming seriously i thought that mathematicians would naturally have built tunes when they wanted to make a programming system or slate or arrow or something like that it seems i'll spend my whole life as a big argument with humanity about hierarchies it's a simple thing to derive that "hierarchies suck, cybernetically" i recall the word hierarchy being used on the slate webpages for resolving namespaces in your object system.. :) it's quite another to use such reasoning to prove all such consequences to people in a way from which they will benefit yeah well that's the best bet i have right now i'm also considering a graph architecture oh no please not graphs! which is looking more and more likely actually, it's needed really? normal oo languages just have hacks built in well if my objects are namespaces (just like in self or smalltalk or yada yada)... then ultimately the inner method has to be able to receive as an argument an object from anywhere in the system Gakuk which breaks the hierarchy hi fare since the hierarchy is what provides objects their access to functions, and since i'd like to unify methods with objects, i have no choice but to address this formally instead of with a hack 08:00am btw self does it with a rather explicit hack called the "lobby" what's that? what i'm looking at basically involves giving objects that act as methods (possibly bounded) access to the enclosing module-like namespaces (up to the root namespace) the lobby is the self system's root namespace aha (see the self programming reference or run self on a mac or sparc) no mac nor sparc here :) i might buy a mac just to see self run for once what does "unify methods with objects" mean? hell it's been six years since i first heard of it aren't methods already objects in, say, LISP? no objects as in self what gives? nothing, i'm merely confusing you with terms i really agree with you, but you'll never grok how self runs on Mac? In my youth, it only ran on beafy SparcStations. yes self runs on macs if you'd spoken with jecel in the last year then you'd know 08:10am hm that killed conversation pretty quickly smoke: where were we? water: dunno :) water: ah yes i could ask a stupid question now shoot :) water: will slate be useful for realtime interactive applications? :) don't tell me it wasn't stupid enough actually i recall talking about circularity of namespace structure, but ok yeah eventually yeah, but that got me bored since i know but very little about programming language design :( i'm mostly interested in a Useful language in which coding is fun again.. ok have you used squeak at all? yes, i fiddled around a little with demo code -:- SignOff AlonzoTG: #TUNES (Have Nice Day :)) the current implementation was completely useless for what i have in mind you mean the "play with me" windows? why? it seemed too slow well they're working in a nice jit compiler this year and everybody's working on optimizing things in particular also i didn't really like the graphical interface, and couldn't find out how to omit using it you have to remember that squeak is still in prototype stages right now i'd like to use my own graphic libs; the language shouldn't handle that for me did you really try to use morphic? yes i know.. well squeak's supposed to abstract that for the programmer so you can run the same app on 12 oses ah then i was right in thinking it was useless for what i want :) bah the programs i write don't need to run on 12 os'es. in particular, it's good enough if they only run on one or two machines well slate won't have that aspect 08:20am okay you obviously didn't look at squeak with an eye for using it the things i like to write are either 1) useless, 2) of temporary use or 3) easily portable hm 3 true i was looking agressively for something i could use right away.. in the future i might need to take a look at squeak again (but then for other projects) i can't see why squeak wouldn't help you with 2 or 3 (the programs in category 3 consist of no more than 50 lines usually) well i just wanted to know what you thought of the overall idea they'd probably be 10 lines in squeak hahahaha :) ok this is going nowhere true the things i want are always considered 'stupid' by people who are designing programming languages or operating systems so what the hell do you code? (unless those designers are children themselves) demos/intros/eyecandy, a pixelpainting program and a module tracker i only program for a hobby. for serious work i wouldn't dare to touch a computer. squeak has a pixel painter, btw yeah, but it's probably not what i like to work with bleh forget this hm perhaps you should have a look at what i paint with it www.casema.net/~smoke/bizarre.png do you want to discuss slate, tunes, or arrow? yes i would love to, but because of my background i would probably spoil potentially important discussions :( * water/#tunes waits for this very slow site to show the damned picture aren't you lucky it's only a very small picture :) wtf is that? view it in 320x200x8bpp www.casema.net/~smoke/vomit.png , www.casema.net/~smoke/default.png .. * water/#tunes invokes "zoom in" on his browser oic xcom-style art xcom? yeah the ufo defense game from the early 90's aaaah yes! well the art i like most predates that game :) it's abandonware now xcom sure was fun yep if you HAPPEN to have svgalib+libmikmod+OSS sounddrivers on a linuxbox you may want to look at a fullfledged intro of ours - http://die.die.ms/intro.html no if you want to understand my interest in the Tunes project best, you should have a look at the sourcecode to the `sesamstr' 4k intro there.. after you've browsed through that you will probably never talk to me again i'd best be blunt... if it can't benefit slate/tunes/arrow, i don't give a rat's patootie could you just summarize? 08:30am it's handcoded assembly, which was more fun to code than anything else i've done in any other language. why are you interested in tunes at all then? imho some tunes members could benefit from merely looking at this stupidity. um why? well, i got the impression there are some very intelligent people here no, there are a very few and i'm very scared of the future for computers well you're part of that scariness true, i said `some'. yes if i were to design something like tunes, i would make 'minimal hardware' an important issue. however, most of you here are not. my interest now lies in finding out why. define minimal hw investigating in the programming languages you all like, i have found out a part of it :) minimal hardware = hardware with very low processing power and very little memory say, a c=64 or a mobile phone for example that's circular good god that would only run forth seriously you can't expect more than a good forth on something like that yes, it'd be unpractical to design a decent programming language for such hardware. so your point is? well, you the c64 was just an example to get an idea. a better example might be having an account on a machine somewhere on the internet well i never have doubted tunes (or whatever) would run fine on a 386 it's not very common now, but it might be in the near future yes i know it wouldn't be weird if that account had the equivalent processing power of a c64 but still it could run a full fledged os, since the combined accounts run on a big machine. no, but all the same it's a bad idea to do anything other than forth on it or some similar language well, it might be possible to write a decent language that uses `minimal hardware' wait a sec yes i just said the opposite. why the hell would someone want an account with a lot of lag time and low processing power? but that was for the c64, this isn't. that's total bs that someone could lack the money for a decent account 08:40am as most of the worlds citizens do yeah with terabyte servers the net is not about equal access, btw *ahem* so have you already stated your point? yes, several times. then shush ok 08:50am why are you people obsessed with outdated machines? c64 or 386s ?? because not every can throw down the cash for a computer and people will still need software in a depression depression? where? not *where*, *when* it might happen yeah right computers are cheap and will be even cheaper oh of course there's always been money to go around what if the manufacturing companies go bottom-up? 09:00am they won't lol 'cause fufie said so because there is a _big_ market for computers if one topples, two will arise from the rubble i know there is, but markets don't go bad because people think they shouldn't (aka demand) markets go bad because of people's basic fears and prejudices people depend on computers for _everything_ (even sex in extreme cases) and that won't suddenly disappear of course not but i'm saying it's not impossible imo it's naive to think that progress will stop and we suddenly will start using c64 and 386s and there's always going to be low-poered machines do you know how much a 486 chip costs today? yeah well then you're dense because they were only examples a few dollars.. and i'm an idiot for getting caught up with someone who wants to start an argument for the sake of it less thana big mac.. :) and people in somalia can afford big macs every day of course probably not eating big macs every day isn't very healthy though oh and lets not forget those russians and but one doesn't need to buy a new computer every day.. s/an// typo-day yeah but it also takes much more than a cpu to make a computer or maintain it or give it access to the net 09:10am are outdated machinery cheaper to maintain and connect to the Net? examples, dense man they were examples besides i'd be fairly certain the upgrade cycle won't affect a lot of the people we're discussing I don't know.. maybe you have a very different philosophy from me.. I write sw for the machines we will have next year and not those we threw out last year there you go you write sw for rich people rich??? people in the first-world technology club they may not have money but they're connected to an environment that makes it easy to get into tech yup the first-world but that's not necessarily the rich -:- rares [rares@wtrb-sh2-port191.snet.net] has joined #tunes rich was a metaphor did the second-world disappear when the commie states died? they have resources in their environment that aren't formally quantified it may have so we're just the first wortld and the third world these days? you want an argument in predicate logic or something? eek no buffer to keep the morons out from both sides hi rares :) hey w: no.. just curious about how the world is perceived in different parts of the world fufie wants to prove that what i said is what i meant and that it's wrong if he finds it amusing more power to him yeah well i got sucked in :( * Fufie/#tunes gets blamed for just about everything.. yeah well if you're not talking about Tunes on #tunes, something's wrong in my book i thought i had the record for discussion derailings well it's not as bad as #slashdot (out goes coherenceinteresting tho) -:- dalvarez [ircusr@212.68.72.98] has joined #tunes hey it's dalvarez (aka IRC User :) 09:20am according to the logs you are having the day's first conversation merathon in here thus I though I'd join ah well you'll have to add fuel to the fire to keep in going :) Windows sucks (how's that) hm its a start hey, smoke, continue what we begun yesterday yeah the logger missed it rares: it must be something people disagree about :) smoke: I think you have understood me somehow wrong and taked it for machine internal scheduling only, but the idea I presented you was ment for use with migration scheduling water: did you have a good look at drscheme and the plt docs? yeah a little the best docs i could find were the r5rs which dissappointed me, but then i don't understand how to browse drscheme code I was more thinking of the plt articles.. not the scheme docs the plt stuff requires you to grok r5rs -:- AlonzoTG [Alonzo@216-164-138-133.s133.tnt5.lnhva.md.dialup.rcn.com] has joined #tunes umm.. true.. I thought you knew r5rs or r4rs i do but it's a bad idea to rely on it for language docs and it's definitely not a squeaky thing 09:30am it's a very schemish thing well sure and half of it is made that way to amuse/annoy the more common lispers but how do you browse code that's not in the r5rs? I gotta go earn my self esteem or some shit cyall hm ok I don't know the drscheme graphical environment very well, ask at the mailing list -:- SignOff rares: #TUNES ([x]chat) no thanks i joined and it didn't seem like the kind of community i like but I think drscheme has units as a package concept one can browse um how?? is anyone interested in an idea of distributed scheduling speedup (we could form a query-group then to not to distract) ? in lisp one has packages which are more or less symbol tables one can inspect and do stuff to.. plt scheme has units which I think you can browse i like the demand-driven concurrency idea of using function calls and distribution to drive concurrency that doesn't help much the environment is terrible if there's no immediately obvious way to browse code it seems like it's emacs or tough luck now I think you're unfair well i've been trying to figure it out for a few days now an environment might be excellent to use even if someone totalyy new dosn't get it typo day argh no even squeakers admit that their environment sucks because you can't find the right tool for the job sometimes but they admit it and try to fix it lisp/schemers just point to emacs it seems because emacs is very helpful to lispers and schemers i don't doubt it but it's not obvious to new users but so is morhpic to squeak right that is why one usually gives new lispers and scemers a tutorial in the environment to show how one uses it and (no offense) it's part of their or someone's job to make improve the learning curve who? improve? other lispers? other lispers yes i don't know any good lispers at all requiring word of mouth proves the environment is bad it doesn't require it.. it just shortens the learning curve by an order of magnitude? by a lot the learning curve should be as short as possible if i have to read r5rs every time i want to find how the environment does what i want it to do, i'll go somewhere else r5rs doesn't describe the environment i'd rather ask the environment itself 09:40am it's a document of the scheme standard so the question is how to ask the environment the question? in squeak you use the browser, inspector, explorer, change set, debugger, dictionaries C-h i g r 5 r s RET http://www.cs.rice.edu/CS/PLT/packages/doc/drscheme/index.htm they are all available from the user interface lol great that's soooo obvious all docs: http://www.cs.rice.edu/CS/PLT/packages/doc/index.html