IRC log started Wed Jan 19 00:00:02 2000 has there been any code yet? ;-) there has been some many ppl have some, tunes has none water: have u heard of bank of the west? the people haven't done the research, for the most part [msg(TUNES)] permlog 2000.0119 air: yeah air: noot that i remember s/noot/not/ they any good? i don't know hell, i'm too steeped in code lately to know what day it is sometime s how does the semantics page look so far, specificity-wise? s/ hte / the / on the semantics page :) oh -:- SignOff washort: #TUNES (brain over - insert coin) fixed "Object-centered programming" == prototype-based? mostly yes i'll write about prototyping in there 12:10am prototype-based == TRULY object-oriented ok, i added a little blurb uh oh, got to mention that functional-style message-passing isn't the default any more k well, it does seem that making the 'call-object-as-function' implicit message can be done without difficulty in implementation can an object receive multiple arguments? sure the current idea is just to use the slots as arguments actually doh what? no, that's not how it's supposed to work why not? i confused it with another idea i had recently but it doesn't seem like a bad idea.. i'll tell you this: i'm not sure how i want to do it yet okay i had been struggling with the same issue myself wel, beta specifically denotes input variables separately from slots (inner variables) which is very arguable an intuitive way to make a function yeah.. 12:20am one could designate certain slots as inputtable yes -:- SignOff air: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep) in fact, that's how I do it. but i hadn't considered doing it for objects AND functions i think i need to unify them some more hm now that i consider it, although references can be mutable or immutable, objects always would have to be immutable to support that hm, true with variability of their own references within yes, and this ultimately reduces to constants vs variables you've got to determine what exactly is *meant* by an immutable reference object vs reference semantics ya well, the immutable reference doesn't respond to outside updates to the stream of object-states, imo stream of object-states? i think i haven't read the same papers you have :) heh well, immutable objects in lisp are usually streams of objects okay.. er i suppose a reference could be considered an object with one pointer variable why? hm i'm trying to make sense of this issue my "references" are just slots my slots happen to be independent objects referred to by other objects hm okay 12:30am i'm just playing with the idea for now what do the slots contain other than a reference to a "real" object? heh. nothing :) see, that way, references to numbers are treated the same as references to ordinary objects at least, it seems to be that way why can't you have a reference to a primitive number without slots as their own objects? -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us713.javanet.com]) -:- hcf [nef@me-portland-us713.javanet.com] has joined #tunes i suppose because i'd like to blur the line between primitive and constructed witten: got a webpage? water: i would oto hcf: http://torsion.org/witten/ too, rather lol nice torsion blurb thanks :) is that an old picture? i figure i should get some working code before i put up a web page for it ;) hey! water: somewhat, why? i've got some lisp code at least hehe i've got a little c++ code yuck don't worry. i'll rewrite it in torsion as soon as possible :P hm 12:40am well, i can include an implicit method in every object by default that performs the beta-style pattern execution thread not familiar with that abi: beta? beta is an advanced object-oriented language that integrates objects and functions into patterns, or at http://www.daimi.au.dk/~beta/ oh, i know what beta is.. i just haven't read much about it because the syntax really grated on my nerves :) my idea about single-object i/o was that the object as function could manipulate that single object, adding or removing slots as necessary to provide the meaningful output (where necessary) yeah, the syntax is pretty weird meaningful output, as in return values? that seems clean enough and close enough to the beta idea without reducing the self-style simplicity factor yeah but as in beta, you don't always execute an object for return values hm it sometimes acts as a thread for i/o or synchronization control seems i should have another look at beta or used for co-routines oh, some mutant of eiffel does that it's even possible to pass continuations with executing objects ok, as much fun as this is, i must sleep hm... so if i "yanked" the output from an object that didn't get an input, i'd intuitively get an expression equivalent to the intialization code with the input lambda oh ok gnite then nite -:- witten [witten@adsl-gte-la-216-86-196-179.mminternet.com] has left #tunes [] hm i never would have expected a person like that heh 12:50am this "jon katz" /. headline generator is a pretty cool little toy hm... got any suggestions for slate work tonight? title slate-semantics should be Slate Semantics i think i'll stick with "Default Slate Semantics" * hcf/#tunes clarifies the title tag hm? oh! there slate-home's should be (Slate Home | Slate Project Home | The Slate Project) <water> ok, the forseeable next steps are defining the default object slots <water> ah <hcf> for (collab manifesto members)'s <title>, prefix 'Slate ' <water> maybe you should take over the web management :) <hcf> i prefer being the suggestor <water> yeah, with no responsibility :) <hcf> bingo <water> geez 01:00am <hcf> what? <water> k, changed <water> hmm better include the obvious object-representation blurb <hcf> r u planning to have a footer and/or header for each page? perhaps things get more automated <water> probably belongs in semantics for now <water> yes, but i don't have anything to handle that yet <water> tril suggested using bineng's python scripts <hcf> oh, the edit page thing? <water> yeah <hcf> have u seen the use of it? the tpp site <water> looking at it now <water> can't say i like the style <water> but the function of the script seems to be good <hcf> i'm the one who suggested that the editpage page should be editable ;) <water> heh <water> maybe i should try zope <water> although it seems a little over-powered <hcf> yeah <hcf> how big do u actually thing the slate pages will get? <water> hm <hcf> oop <hcf> s/thing/think/ 01:10am <water> have you looked at the squeak pages, including the main swiki? <hcf> i tend not to like wiki stuff <hcf> but if u like it, so be it <water> roughly that large, but more organized and coherent <water> well, i like small interconnecting pages <water> especially orthogonal ones <hcf> what is needed right now? <water> mostly, the web stuff should just be as multi-purpose as possible <water> and not confusing to search <water> hm <hcf> have u looked at laml yet? <water> not closely <water> what's good about it? <hcf> i havnt looked at it yet <hcf> r the bisom papers from researchindex enuf or do u want something additional? <water> i haven't looked 01:20am <water> man, my day never stops <water> i mean, work now is annoying because they make me wait around for no good reason when i could be working <water> how do you get the papers that ri references? <hcf> i thot they were linked to, i'l check <water> k, figured it out <water> i had to find a self-citation and get the details on it... pretty obtuse <hcf> query->details->link below title+by lines <water> wtf is query? <hcf> in this case, the results page <water> see, this is why i have you :) <water> i hate ridiculous interfaces like this <water> if i had to wade through that crap to find papers, i'd never have time to read <water> as it is, you're distracting me from adding to slate specs 01:30am <water> :P <water> :) <water> it's fun to be a jerk sometimes <hcf> this 'face iz lame <hcf> they give an url <hcf> last line of each result item paragraph <hcf> but its plain text -:- SignOff hcf: #TUNES (Leaving) <water> hm... i figure 4 implicit slots: meta, parent, input, output -:- hcf [nef@me-portland-us713.javanet.com] has joined #tunes <water> to repeat: i figure 4 implicit slots: meta, parent, input, output <water> that seems to cover it within the basic object <water> am i forgetting something? <hcf> not that i can see <water> hm <water> hmm... meta-objects seem a bit thorny <water> i mean, why would someone use a meta-object as a function? <hcf> shouldnt need to <water> pretty funny, actually... i bet i won't figure out any good use for it, but i just *know* some hacker'll figure it out real quick <water> maybe fare has thought of something similar <water> whoa... a meta-object function would have access to all sorts of objects without having to reflect explicitly at all 01:40am <water> k, uploaded the basic formats for object/metaobject slots <water> *really* basic for now <hcf> ur gonna do a mlist intro right? <water> in a day or two yes <water> should i mention the mlist to the tunes mlist? <water> i did to bill tanksley <hcf> do it for lurkers sakes <hcf> s/s/s'/ <water> k <hcf> meta objects have meta objects? <water> yeah <water> like i said, there's a "top" 01:50am <water> a meta-object that's self-describing <hcf> i recall now <water> the more important relationship is still prototyping <water> k, sent another mail -:- MP8192 [mp@a90-p59-zg.tel.hr] has joined #tunes <water> hm <water> hello -:- SignOff MP8192: #TUNES (Nema me, nema me, ooooooo) <Fufie> water.. <water> hey f00f <Fufie> hi <water> as usual, not getting my sleep :) <Fufie> aren't you a bit early in announcing slate as the hll? <water> not an announcement, as a question <Fufie> still 02:00am <water> see the little "?" <Fufie> :) <water> well, that was the point of slate <water> 4 years ago <water> i just didn't give a damn about tunes then <water> ok i need a nap before work <Fufie> good nap <water> there's a lot of work tomorrow <water> mostly just a lot of stress <water> :P <water> gnite -:- water [water@tnt-10-171.tscnet.net] has left #tunes [] -:- SignOff hcf: #TUNES (Leaving) 02:10am -:- binEng [Anders@j141.ryd.student.liu.se] has joined #tunes -:- thomas [thomas@193.217.63.152] has joined #tunes -:- AlonzoTG [Alonzo@207-172-49-128.s128.tnt7.lnhva.md.dialup.rcn.com] has joined #tunes -:- hcf [nef@me-portland-us140.javanet.com] has joined #tunes -:- jdl [jiml@ultra1.inconnect.com] has joined #tunes <binEng> hi <jdl> hi, bin <binEng> what is your C code supposed to do? <jdl> Have you seen my "C-Like" code? (Same for you, fufie) <jdl> :) <jdl> bin: It's supposed to be isomorphic to ANSI C -- i.e., to have the exact same functionality. <binEng> A code alone is not much to have unless you can do something with it.. or is it just an experiment to see if the metacode sufficies? 08:20am <jdl> bin: There's lots of advantages -- first, C is familiar to almost everybody, so it's a good demonstration of how Prism codes work. <jdl> Second, I'm going to use it as a proof of concept for the Prism Metaprogram Protocol. <jdl> The C-Like code allows Prism to work with ANSI C models. <binEng> "work" in what way? <jdl> Well, the Prism protocol only works with Prism models, not with ordinary source code. So given the C-Like code and a parser, you can translate ANSI C models into Prism C-Like models, thus allowing the models to be shared between Prism metaprog <jdl> metaprograms via the Prism protocol. -:- SignOff AlonzoTG: #TUNES (Read error to AlonzoTG[207-172-49-128.s128.tnt7.lnhva.md.dialup.rcn.com]: Connection reset by peer) <binEng> ic <binEng> I never got the hang on what the metaprogram protocol was..? <jdl> The advantage of using the Prism protocol is that Prism models are a lot easier to deal with than ordinary source code models (at least, that's the goal). <jdl> The metaprogram protocol is a way for metaprograms to share program models that's easier than using ordinary source code. <jdl> It's also a way to shake out the bugs in Prism so it will be tested for the next stage, which is to make codes themselves interoperable. <binEng> like a pipe? <jdl> bin: Exactly. Or rather, the Prism protocol says what you send over a pipe. With the Prism protocol, your command-line might look like this (literally): <jdl> parse_c | lint_c | optomize_c | c_2_x86 | optimize_x86 | emit_x86 a.out <jdl> s/parse_c/parse_c foo.c/ <binEng> ok <jdl> It allows you to write small, focused metaprograms without having to worry about the hassles of parsing and emitting real source code. <binEng> yes 08:30am <Fufie> jim: I am here <jdl> Anyway, I just dropped by to see if anybody had looked at the C-Like code yet. (I can't stay, I've got a job interview in half an hour.) <jdl> hi, fufie -:- AlonzoTG [Alonzo@216-164-138-151.s151.tnt5.lnhva.md.dialup.rcn.com] has joined #tunes <jdl> fufie: see log <Fufie> just did <binEng> jdl: How about if I filtered out the entries in the glossary you don't "need".. <Fufie> I had a quick look at the C-like thing, but I'll have a more thorough look later <Fufie> a few things were missing though <jdl> bin: sounds good... I'm going to go through it later today and mark it up -- can you send me the raw source? <Fufie> and I'll see if I can't dig out a pdf or a ps of the standard for you <jdl> fufie: what were your first impressions? <binEng> sure. I should probably write up my thoughts in an essey instead <jdl> fufie: that would be very good :) <Fufie> my first impression is that it looked a little like a simplified version of the yacc-grammar <jdl> bin: that's my feeling -- BTW, I do agree with the sentiment of your goals entry, just not the location. <binEng> jdl: great :) <Fufie> I presume you already have K&R which includes the grammar <jdl> fufie: that's correct. "simplified" is the goal - the advantage of the Prism metacode is there's no syntax, so you don't have to do funny things just to get the language to parse right. <jdl> fufie: Actually, no, just a desktop reference. :) * Fufie/#tunes things jdl is poorly equiped for a programmer ;) <Fufie> thinks <jdl> :) <binEng> jdl: there.. sent. <Fufie> again.. it's brave to try to make something isomorphic to ANSI C when you don't have either K&R or the standard ;) <jdl> bin: thanks <Fufie> and I presume it is a one-way only translation.. you lacked too much to translate back to ANSI C and be isomorphic <jdl> fufie: well, I'm pretty familiar with C. Also, like all things I work on, I don't expect to get it exactly right the first time. <Fufie> and please use the words used in C, ie it's called statement, not instruction <binEng> jdl: Lemme sak you one thing, how does the versioning work? Do you do it by hand, or have some software to keep track of documents and versions? <jdl> fufie: not so! At this time, you'd lose symbol names and comments, but nothing else (except for the parts that aren't supported) <jdl> bin: by hand ;) <Fufie> I'll have a closer look as I said <binEng> ah <Fufie> afk.. mash <jdl> mash? <Fufie> the tv series <jdl> oh <Fufie> and I'll eat while watching :) <jdl> Anyway, if you guys could type up your comments and send them to the prism mailing list, I'd appreciate it. (I have to prepare for my interview now.) -:- SignOff jdl: #TUNES (Leaving) 08:40am -:- ilona [user1197@212.25.69.210] has joined #tunes <ilona> i cant tune display settings (Deal) for X-Windows in Phatlinux -:- SignOff binEng: #TUNES ( <k!14>) 08:50am -:- hcf 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> -:- ilona [user1197@212.25.69.210] has left #tunes [] -:- NetSplit: sterling.openprojects.net split from lewis.openprojects.net [09:38am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [sterling.openprojects.net] -:- Netjoined: sterling.openprojects.net lewis.openprojects.net -:- abi [nef@bespin.dhs.org] has joined #tunes -:- hcf [nef@me-portland-us140.javanet.com] has joined #tunes -:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes -:- Fufie [stig@tunnel-44-22.vpn.uib.no] has joined #tunes -:- iepos [root@d16.k1-1.tecinfo.com] has joined #TUNES -:- Fare [fare@quatramaran.ens.fr] has joined #Tunes -:- AlonzoTG [Alonzo@216-164-137-207.s461.tnt4.lnhva.md.dialup.rcn.com] has joined #tunes -:- Vilts [root@et14.estpak.ee] has joined #tunes <Vilts> hell oo <Vilts> anyone knows about ncurses? <Fufie> yes <Fufie> but haven't used it in years <Fufie> if you can, choose some other library <Vilts> great... how can i change background of whole window and specific area? <Vilts> can u suggest something other? <Vilts> i already know how to change color of text... <Fufie> s-lang <Fufie> www.slang.org <AlonzoTG> om <Fufie> or even better, a graphical library <AlonzoTG> spare monitor sucks. <AlonzoTG> =\ <AlonzoTG> my normal one needs to be re-calibrated... <AlonzoTG> =(((( <Vilts> but i need it under text mode... not under X <Fufie> see www.slang.org <Vilts> kay... <Vilts> umm... actually it's www.s-lang.org... <Fufie> :) <Vilts> uhh... it's _programming lang_ or what? <Fufie> it's a curses-like library <Vilts> in main screen it says "S-LANG - Embedded Programming Language"... <Vilts> and "S-lang is a C-like programming language..." <Fufie> yes.. that's good <Fufie> it's meant to be embedded in C <Fufie> but is simpler <Fufie> ie you don't need to fight C when doing curses-stuff <Fufie> big bonus point <Fufie> S-lang is what we call a DSL <Fufie> dsl? <abi> dsl is domain-specific language or digital subscriber line <Vilts> DSL? <Vilts> huh... but C is all i know... <Fufie> give it a shot <Fufie> you might like it 10:10am <Vilts> and u say it can do all these background/text color tricks? <Fufie> yes <Fufie> hands-down <Vilts> and it can make windows too? <Fufie> probably <Vilts> ok... i'll see what it can do... <Vilts> umm... but maybe u know something about teletext too? <Fufie> not a lot <Vilts> how much u know about it? <Fufie> seen it on the telly ;) -:- SignOff hcf: #TUNES (Leaving) <Vilts> hehe... but i must make one proggie for it, maybe u can help me on it? <Fufie> not likely <Fufie> my time is limited <Vilts> some hints... maybe... 10:20am <Vilts> okay.... thanx Fufie...it seems to be quite nice library. Thnx <Vilts> gotta go now -:- SignOff Vilts: #TUNES (Leaving) 10:30am -:- SignOff Fare: #TUNES (Ping timeout for Fare[quatramaran.ens.fr]) -:- binEng [Anders@j141.ryd.student.liu.se] has joined #tunes <iepos> hello... <binEng> hello 10:50am -:- washort [washort@d100.narrowgate.net] has joined #tunes -:- thomas [thomas@193.217.63.152] has joined #tunes -:- SignOff iepos: #TUNES (BitchX: the original point-and-click interface.) -:- jdl [jiml@209.140.64.2] has joined #tunes <jdl> hello again <jdl> okay, n/m ;) <binEng> hullo <jdl> Hi, bin, what's new? <binEng> new? Well.. I've watched some tv if you want to hear about it ;) <jdl> :) <jdl> BTW, what does 'binEng' mean? <binEng> Actually I've been at the university to look at a supercomputer rendering graphics.. -:- binEng is now known as bineng <bineng> (n/m about that) 11:50am * bineng/#tunes messed up heguess * jdl/#tunes is away (msg me to get my attention) -:- dalvarez [ircusr@212.68.72.98] has joined #tunes * Fufie/#tunes stretches and heads back to the telly <bineng> hmm, telly? 12:20pm -:- Kaufmann` [newbie@200.255.150.239] has joined #tunes -:- Kaufmann` is now known as Kaufmann_away <dalvarez> someone familiar with theoretical C, ie expressions, boolean results of them and stuff ? 12:30pm -:- SignOff washort: #TUNES (Ping timeout for washort[d100.narrowgate.net]) <dalvarez> thought about a piece of code that open()s a file, then tries to open(...,O_CREAT | O_EXCL) another, checks the results on zero-value in both cases and asks to proceed in case of confirmation by the user if previously the second open() returned an ... -:- Crypton [Crypton@212.150.33.90] has joined #tunes <dalvarez> error indicating that O_EXCL has been "violated" by the file to open being already existent and then as the conditional block of a superior if-construct checking the result of all this calls an error function to display an appropiate msg and <dalvarez> abort - for this I have to know if boolean-resulting constucts can be included in eachother passing their values to the next upper level <dalvarez> can s.b. give me an advice? 12:40pm <dalvarez> of course I could write it simple one thing after each other but that wouldn't really be a challenge... <dalvarez> wow.... got to notice right now how damn long that sentence has become.... sorry for that :) -:- Crypton [Crypton@212.150.33.90] has left #tunes [OmegaScript] -:- SignOff jdl: #TUNES (Leaving) <dalvarez> could, for example, s.th like 'if ( if (' work ? <bineng> uh <bineng> That's some complicated thing you're trying to explain.. or at least that what you make it sound like :) <Fufie> bin: telly as in television or tv <bineng> ic <bineng> baaad thing, Fufie <Fufie> yes <Fufie> but watching news on the telly isn't as bad as watching soap <bineng> soaps are eeevil <dalvarez> bineng: the description has become a bit wierd, I know (actually right now I've difficulcies to get it myself 'parsed', but the problem I was trying to describe is basically concerning the question if boolean constructs can be (don't know the word), <dalvarez> you know .... put into each other <dalvarez> 'encapsulated' - that's it -:- Fare [fare@quatramaran.ens.fr] has joined #Tunes <Fare> G <bineng> hi Fare <dalvarez> hey Fare <bineng> dalvarez: hardly if's <bineng> dalvarez: instead, 'if(..) { if(..) {..}}' 12:50pm -:- Kaufmann_away is now known as Kaufmann` <dalvarez> yes, of course, but two of the boolean-type expressions have nothing to do with each other except that they evtl. lead to perform the same action, so I can just make the depend on each other by || <dalvarez> in the if <Fare> uh? <bineng> dalvarez: so why not? <dalvarez> and one of the possibillities gets branched then and should be substituted by the result of a second open <bineng> dalvarez: perhaps a common subroutine? <dalvarez> Maybe that's impossible (I'm playing aaround with sample code in the background right now) <dalvarez> bineng: how that <Kaufmann`> Hey, I just came out with a scheme to Make Money Now <bineng> dalvarez: well if the same code occurs two times, a subroutine could perhaps do it instead <Kaufmann`> First, you buy the Microsoft check on ebay. <dalvarez> ah.. understood 'foo { does the branch }' then in the main 'if ( thingie_one || foo())' <bineng> right <Kaufmann`> then, you forge a few thousand copies of it, and put _them_ on sale on ebay. <dalvarez> taking a closer look at my idea makes it look somewhat inefficient in the end (could it be that the processor's testing all the time (more than usual)) <bineng> Kaufmann`: great idea. You do that and send me the money - deal? <dalvarez> Kaufmann: mE t00, please <Kaufmann`> LOL <Kaufmann`> I'll hoard it 01:00pm <dalvarez> bineng: that results in the same computing as it'd do the step-by-step way, so there is no way to get it *better* <dalvarez> thanks anyway... <dalvarez> (sometimes brainstroming-like ideas overcome me and I have to get something more elite) <Fufie> kaufie: how is your distributed object protocol progressing? <Kaufmann`> Fufie, well, it's not :P <Kaufmann`> Fufie, I've other things to do, unfortunately <Fufie> how many projects are you working on now? * Fufie/#tunes has too many projects himself <Kaufmann`> hell -:- washort [washort@d132.narrowgate.net] has joined #tunes <Kaufmann`> 5+ personal projects, plus three others from work <dalvarez> Fufie: you asked me last days if I'd like to contribute to the SDS ... a general question: what the hell has it all to do with XML ? <Fufie> dalvarez: that's the format the programs communicate with.. <Fufie> kaufie: heh.. I have stripped down a few projects and I am currently at three/four :) <Fufie> dalvarez: communicate in <dalvarez> and what language are you using for devel mainly (will surely hear a LISP-like thing that makes me fear as the answer) * bineng/#tunes has a half project or so ;) <Fufie> dal: I write in Common Lisp.. Ghyll has written code in Java and C++ <Fufie> dal: so currently the framework makes it really easy to use those languages.. it might be extended with Scheme soon as well <dalvarez> will start to learn scheme and lisp stuff as asap <dalvarez> as asap <:-P <dalvarez> ASAP! <Fufie> going from c to java or c++ is probably a smaller step though 01:10pm <dalvarez> Fufie : already getting familiar with object oriented concepts <dalvarez> like it more low-level somehow <washort> learning OO concepts from C++ and Java? erk :-/ <Kaufmann`> Not a good idea. <Kaufmann`> Get a book on CLOS <washort> to say the least... <washort> Kaufmann`: whee, CLOS makes my brain explode :) <washort> i'm more of a smalltalker at the moment.... trying to work my way thru _Art of the MOP_ <Kaufmann`> well, then get _Art of the MOP_, which describes a rather more understandable CLOSette <washort> i'm reading it. :) <Kaufmann`> heh <Kaufmann`> dalvarez should get it. <Fufie> CLOS requires more or less that you already understand lisp <Kaufmann`> Self has a good object model. So has Beta, to some extent <dalvarez> tell me more about this CLOS thing <washort> Fufie: i mostly understand Lisp. I think. <Fufie> for a person who only knows C and go all the way to CLOS is _huge_ <washort> dalvarez: it's possibly the most flexible OO system ever. <dalvarez> wasn't self this denton thing? <Kaufmann`> CLOS is the Common Lisp Object System <washort> Fufie: *nod* <Fufie> wash: I was more thinking about dal <washort> Fufie: Smalltalk is a good halfway, i think <Fufie> I think Java is a better mid-step as it forces OO out of C.. <washort> dalvarez: if you haven't already, play with Squeak, it's the most fun programming system I've ever used. <Kaufmann`> I think it's a better idea to first learn Scheme <washort> Fufie: but Java has soem huge brain-damages that really aren't worth dealing with, IMO <washort> Kaufmann`: There is that, too. :) <washort> Scheme is good... <dalvarez> prefer learning few languages in depth than many more or less <washort> dalvarez: Start with Smalltalk and/or Scheme, then, both languages are relatively small <Fare> Where's abi when we need her? <Fare> abi: CLOS? <abi> CLOS is, like, 'low-level' for a reason <washort> heh :) <dalvarez> will take a look at euLISP cause there seem to be many ppl fascinated by it and, as you stated yourself, it'll possibly form the base of the 'tunes-lang' <washort> dalvarez: what language background are you coming from? <dalvarez> wasshort: C <Kaufmann`> Anyway, where is Brian? <Fufie> eulisp has a few problems.. one of them is that you'll need a book and eulisp has no books, and you will require people who know it and that's also tough to find <washort> dalvarez: *nod* i'd stop off at Scheme first, because it's so small and there are many excellent books/tutorials available <washort> and the central ideas are the saem ones ase most other Lisps <dalvarez> already downloaded the euLISP tutotials from bath.ac.uk...you know what <Kaufmann`> washort, only better <dalvarez> abi: euLISP? <abi> i guess euLISP is http://www.bath.ac.uk/~masjap/EuLisp/eulisp.html and http://www.aiai.ed.ac.uk/~jeff/lisp/eulisp.html and ftp://ftp.maths.bath.ac.uk/pub/eulisp/ <dalvarez> that's <washort> dalvarez: do you know any other languages besides C? 01:20pm <Fufie> kauf: define better :) <Kaufmann`> People are going to be pissed off at me for saying this, but Perl might be a good idea. <washort> whee, perl =) * Fufie/#tunes gives Kaufie a fwap.. you heretic you.. <washort> Fufie: hey, compared to a lot of stuff out there, perl is not so bad <dalvarez> learning C++ stuff when I occasionally have time and Prolog, too (which is in the very beginnings); then a bit asm stuff <washort> dalvarez: C++? blick, why? :) <washort> it'll rot your brain <Kaufmann`> Perl is easy to understand. <Fufie> perl is bad <Kaufmann`> ...in a good way. <washort> Kaufmann`: *nod* I still have a soft spot for Perl - it was the first language I learned <dalvarez> wasshort: as thing to add functionality to C? <washort> Kaufmann`: (except for GW-basic, which doesn't count) <Kaufmann`> dalvarez, C++ adds no functionality to C. At best, it adds cruft. <Kaufmann`> and syntactic sugar. <washort> dalvarez: C++ is like an octopus made from nailing extra legs on a dog. <washort> Kaufmann`: syntactic saccharine. <Kaufmann`> LOL washort <Fufie> perl has anal syntax, is completely unreadable and is almost devoid of redeeming features <washort> Fufie: i like perl's syntax <washort> Fufie: you can scare small children with it <dalvarez> C is the only thing really all-purpose so it's cool to know <Fufie> wash: you're fond of line noise as well? <washort> dalvarez: Not true. :) <washort> Fufie: i won't defend perl, except to say it's a kluge and kluges are often useful and necessary. :) <Kaufmann`> Fufie, not true. Perl's syntax makes sense once you "get" it. It's only as unreadable as you want it to be. And it's got a whole shitload of features. <washort> Kaufmann`: well - C++'s syntax makes sense once you 'get' it too. big deal. :) <Fufie> perl doesn't make sense <Kaufmann`> Perl may not be conceptually perfect, but it's very useful. <Kaufmann`> Fufie, yes, Perl doesn't make sense. So what? <washort> Kaufmann`: the cool thing is, Perl doesn't *claim* to be conceptually perfect. It's aiming for second-best at almost anything... and succeeds admirably. <Fufie> perl is a lot of features glued together with the cheap glue you have in kiddie school, some cheap tape and a few small nails <washort> Fufie: yes. <washort> Fufie: that's called "useful" :) <washort> Fufie: not pretty, but useful. :) <dalvarez> I'm sure everyone will start using perl for sys-programming ;) <Fufie> perl's usefulness is so short-lived that it's cruel <washort> i can do random hacks in perl faster than anything else <Fufie> it's a kiddie language for random small hacks under 100 lines <washort> which is why i mainly use perl for random hacks. :) <washort> Fufie: yes. exactly. <dalvarez> mail the proposal to the lin-kernel lists and they'll kick your ass <washort> Fufie: maybe under 1000 lines.... but that's about it <dalvarez> C is standart, that's the main argument <Kaufmann`> Fufie, 1) a lot of stuff falls under that category. 2) as washort said, maybe under 1000 lines. <Fufie> try maintaining a 1000 line perl program made by your average perl programmer <washort> dalvarez: Go ahead and jump... 10,000 lemmings can't be wrong <washort> Fufie: heh. i don't care about average programmers <dalvarez> EVERYTHING uses it, there are implementations avaiable for nearly every arch and it's the lang of UNIX-like OSs <Kaufmann`> Ooooh, I'm so afraid of the lin-kernel lists... :) <washort> Fufie: the average programmer turns out crap in any language. :) <Kaufmann`> dalvarez, UNIX-like OSs are evil. <Fufie> what's the lin-kernel list going to do? <dalvarez> Kaufmann: yet standart <Fufie> wash: yes, but perl is the language for the average programmer.. he is at ease there.. <washort> Fufie: do I care? :) <Kaufmann`> dalvarez, standard doesn't mean good. Not at all. <Kaufmann`> Windows is standard. <Fufie> wash: probably not.. but programmers should be shown the Right Way ;) <dalvarez> Fufie: just a example of the absurdness of making perl appear to be an alternative to C <washort> Fufie: of course they should. <washort> Fufie: but don't take my perl from me. :) 01:30pm <washort> dalvarez: it's not <Fufie> dal: nobody here has suggested perl as an alternative for c for all purposes <dalvarez> I'll learn LISP anyway, so you don't need to convince <washort> dalvarez: not for what you're thinking of anyway <dalvarez> me <Kaufmann`> Well, I'm off. <Fufie> ciao <dalvarez> bye -:- SignOff Kaufmann`: #TUNES (kathyanne gets kick'd / as leaves fall on still waters / she is here no more.) <dalvarez> Fufie: tell me then what language makes sense in what <dalvarez> Fufie: for the specifics <Fufie> can you rephrase your question? <dalvarez> (yes, that was bullshit, here it is again:) where have all this LISP dialects and smalltalk it<<<<#s main application fields <dalvarez> it_'_s <dalvarez> its ( what the hell) <Fufie> smalltalk is used today mostly in the business world <Fufie> and it is also used heavily for prototyping <dalvarez> what are its features compaired to C++/java <washort> and is just a really cool toy <washort> dalvarez: way more flexible, at least powerful <washort> dalvarez: easier to write, easier to debug <Fufie> lisp is used for complex problems and most companies which use it, consider the use of the language an advantage and don't tell the world who uses c++ <Fufie> scheme is mostly in use for teaching and in the language research community <Fufie> scheme also seem to be used more and more as a glue in Linux <dalvarez> I know of math students learning algorithms by scheme <washort> thanks to the GUILE ppl. :) 01:40pm <Fufie> the guile people should start writing documentation <Fufie> the main reason guile is not used much is because the people don't document it <washort> this is true <washort> they're too busy integrating generational GC and GOOPS, i think :) <dalvarez> can u give me hints on literature to read to get a fast and detailed insight into the LISP field? <washort> dalvarez: try www.schemers.org <Fufie> "the evolution of lisp" is about eight years old but gives very good background info <Fufie> www.lisp.org should have a link <washort> *sigh* i'm working with the Jabber project, which is building an instant messaging system based on XML <washort> (i'm writing a squeak client) <dalvarez> will take a look at all this sources in addition to our local university's stuff <washort> it drives me NUTS watching the server ppl and the win32 ppl chasing memory bugs :) <dalvarez> then maybe I can go more indepth on a dialect like euLISP <dalvarez> bye for today - I'm tired and tomorrow school will suck again -:- SignOff dalvarez: #TUNES (Leaving) 01:50pm -:- eihrul [lee@usr5-ppp194.lvdi.net] has joined #tunes -:- SignOff washort: #TUNES (Ping timeout for washort[d132.narrowgate.net]) -:- SignOff AlonzoTG: #TUNES (Have Nice Day :)) -:- AlonzoTG [Alonzo@209-122-248-171.s425.tnt8.lnhva.md.dialup.rcn.com] has joined #tunes -:- SignOff bineng: #TUNES (good evening <k!14>) -:- mibin [mibin@62.11.105.175] has joined #tunes -:- NetSplit: forward.openprojects.net split from irc.linux.com [02:45pm] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [forward.openprojects.net] -:- Netjoined: forward.openprojects.net irc.linux.com -:- Fufie [stig@tunnel-44-22.vpn.uib.no] has joined #tunes -:- thomas [thomas@193.217.63.152] has joined #tunes -:- eihrul [lee@usr5-ppp194.lvdi.net] has joined #tunes -:- AlonzoTG [Alonzo@209-122-248-171.s425.tnt8.lnhva.md.dialup.rcn.com] has joined #tunes <Fufie> hmm... crusoe which is 128-bit seems ideal as a cpu for lisp-machines <Fufie> tunes on crusoe anyone? <eihrul> crusoe is released? <Fufie> yes <Fufie> today <eihrul> gah <eihrul> its just another x86 clone-thingy <Fufie> not really <Fufie> (sheesh.. netscape hogs memory) <eihrul> hrmm <eihrul> i'm still reading <Fufie> 4356 stig 0 0 135M 135M 8768 S 0 0.0 54.0 15:12 netscape <eihrul> trying to get through all the marketing hype <eihrul> so what exactly is it? :) <Fufie> it is using 135 M .. it's dead now though <eihrul> i mean crusoe <Fufie> a new processor <Fufie> see /. (level 2) for some decent info <AlonzoTG> parents machine: 16 mb physical; 12.5 mb swap in use... <eihrul> i know its a derned processor <eihrul> Fufie: maybe a programmer's manual would be nice or something :) <AlonzoTG> they want you to apply for it... <AlonzoTG> it's not good enough to publish on their website so they ask that you pay for it... <eihrul> they do this just so i have to download the product overview 03:00pm <Fufie> eih: give them some time.. it's what? five hours since it was released? :) <Fufie> but a 128-bit processor with code-morphing would be ideal as I see it for lisp and for tunes <eihrul> but it appears you can only use it for x86ish stuff <eihrul> but this is more fuel to fare's ubiquitous virtual machine layer argument in review project about x86 :) <AlonzoTG> yes I agree. Surely native mode code could be executed.... <eihrul> Fufie: 128 bit-wide registers? 03:10pm <eihrul> Fufie: ? :) <Fufie> yes <eihrul> wtf do you need registers that obscenely wide for? <eihrul> :) <Fufie> implementing morphing, lisp and foobar? :) 03:20pm <eihrul> geez man... you can fit all x86 general purpose registers in just two of those :) <eihrul> for some reason i doubt they're 128 bits <eihrul> cause there are 64 of them <Fufie> some details here: http://www.zdnet.com/pcweek/stories/news/0,4153,2423983,00.html <eihrul> uhm, i'm reading the manual man <eihrul> i'll trust it over zdnet 03:30pm <eihrul> this is kind of funny though <eihrul> now even the machine language needs to be compiled :) -:- jdl [jiml@ultra1.inconnect.com] has joined #tunes <jdl> fufie: you awake? <Fufie> yes <jdl> fufie: If you don't mind, I'd like to discuss the holes in the C-like code. <Fufie> I haven't checked it out further.. <Fufie> I have been busy with my own code <jdl> Earlier you said there were some things missing -- what were they? <Fufie> I checked your missing list <Fufie> it names most of it <Fufie> in addition you have the combinations of all those <Fufie> ie pointers and arrays <Fufie> and please call it statement, not instruction (I said this earlier) 03:40pm -:- washort [washort@d134.narrowgate.net] has joined #tunes <jdl> fufie: okay, so nothing that isn't already on the list (no problem about "statement") <Fufie> the list is pretty big when you add all combinations as well <jdl> fufie: you said you might be able to get me a copy of the ANSI C spec -- do you have that or know where I could find it? <eihrul> washort: lo <washort> eihrul: hi. <eihrul> jdl: www.c9x.org <eihrul> jdl: it's the yet-to-be C spec <Fufie> and you have the preprocessor <jdl> (I'm not really worried about combinations; everything I implement I'll make sure works properly (to the best of my ability)) <eihrul> freely available <jdl> e: "yet-to-be?" <eihrul> it's still being standardized <eihrul> but at the very least it covers all the already eixsting parts of C <jdl> fufie: yeah, I'm not sure how to handle the preprocessor. :( It's lexically-based rather than semantically-based, which screws things up. <jdl> e: I thought ANSI standardized C a long time ago? <Fufie> eih: on that site I just get the debian standard page <eihrul> jdl: new standard, man <Fufie> jim: C is being revised <eihrul> grrr <jdl> Oh, shows how much I know. :) <eihrul> seems the site got hosed <jdl> fufie: WRT the preprocessor, right now I'm just going to let the parser take care of it. <eihrul> there's the copy i have <jdl> e: thanks, but how do I get it (irc newbie) <Fufie> you also miss elliptic forms <jdl> fufie: the only one I'm aware of is variable number of parameters, and that's on the todo list. <jdl> e: looks like I figured dcc out -- n/m, and thanks for the file <Fufie> trigraphs? <jdl> trigraph? If you mean "x?y:z", that's supported. <Fufie> no.. that's not a trigraph <Fufie> ??= is one <Fufie> ??> is another <jdl> Doh!! I missed all of those. Good call! <Fufie> was enums in there? 03:50pm <Fufie> did C support digraphs or is that just c9x and c++? <jdl> fufie: Actually, I mised +=, *=, etc. (Doh!) I'm not at all familiar with trigraphs... explain please? <jdl> fufie: enums are on the to-do list. <Fufie> enums are not on the todo list I am reading.. <Fufie> maybe I am rreading a cached version <jdl> brb (checking) -- no once a version is published (non-draft) it never changes <Fufie> you seem to miss pointers as well <jdl> fufie: you're right, enums are on my informal list but not in the document. <Fufie> a trigraph is used for keyboards that do not have all 102 keys available.. not used a lot anymore <jdl> Pointers are already supported (see "Type" atom set and "Reference" and "Dereference" statements) <Fufie> I have used trigraphs in a project several years ago <Fufie> needed them back then <Fufie> using the word reference is not good in c <Fufie> the word reference has another meaning in the c/c++ world <Fufie> in compiler-speak it's ok though <jdl> fufie: In C++ anyway... what would recommend? "Address" maybe? <Fufie> pointer <Fufie> noone will misunderstand it <Fufie> making new terms is not good <jdl> "Reference" = "&p" ; "Dereference" = "*p" <Fufie> umm yes.. you're right <jdl> I agree about terms... but I didn't know the actual names for those operators so I just chose some. <Fufie> get c9x and use the names used there <jdl> My point about reference/dereference -- if I said "pointer" instead of "reference" people would think "*p" instead of "&p" <Fufie> yes.. reference is the right word for &p * eihrul/#tunes has always interpreted &p as "address of." <jdl> okay... I've got c9x now, so I'll look at that for names. <eihrul> but that's just me <Fufie> char *p; is a pointer though.. not a reference * Fufie/#tunes ponders on buying a pointer <eihrul> void or laser? :) <jdl> :) <Fufie> a pointer that says woof <eihrul> oh <Fufie> and bites be when I haven't taken it for a walk <eihrul> name him C <eihrul> then you can teach him stupid C <jdl> lol <eihrul> tricks <Fufie> I'll probably name him Schemer <eihrul> and when it runs into walls you can say, "Look how stupid C is!" <Fufie> pointers are pretty smart -:- lar1 [larman@sdn-ar-019casfrMP046.dialsprint.net] has joined #tunes <Fufie> jim: indirection lvel won't do the job <lar1> Hey <jdl> fufie: why not (as part of variable declr)? 04:00pm <Fufie> jim: const char * const char foo; has several layers <Fufie> can't remember if it's correct syntax, but c9x goes c++ <jdl> fufie: doh, you're right. I remember reading about that in a C secrets book. <jdl> fufie: my concern with using c9x is that I want to be compatible with existing source code. Will that be a problem? <Fufie> no.. c9x will work with old non-deprecated ansi c <jdl> okay, good. <Fufie> note the word 'non-deprecated' :) <jdl> I'll look at the spec, I imagine that will give me the details. <jdl> fufie: Any comments on the Prism approach in general? Its goal is to make it easy to create, understand, and programmatically work with codes/languages. What do you think? <jdl> (Given the C-like code as an example.) <Fufie> it's not isomorphic so far <jdl> fufie: no, but I don't think you could argue that isomorphism is impossible, or even that much more effort on top of what I've got already. <Fufie> but if you masnage to rewrite languages to output a simpler frontend, this will make things a lot easier for SDS.. by themselves the formats aren't too useful :( <Fufie> isomorphism is not needed as I see it <jdl> :) no, but still, you have to have the format before you can output to it :) <AlonzoTG> om <jdl> fufie: isomorphism isn't needed for SDS, but it is for Prism's overall goals. <jdl> eihrul: In case you're wondering, we're talking about http://www.teleport.com/~sphere/documents/001a/1/index.html <eihrul> rmm, i've heard of prism before <Fufie> the problem is also that as you introduce the other missing-parts and combinations the complexity will grow <Fufie> lag <jdl> e: not Prism in general, a specific attempt to make C work with Prism (that's what the URL points to)... I don't care if you look or not, but there's the URL if you want to. -:- SignOff washort: #TUNES (Ping timeout for washort[d134.narrowgate.net]) <eihrul> well, i'm not really fond of C :) <AlonzoTG> om <AlonzoTG> anyway... <AlonzoTG> The free hardware I'm assembling will have a FORTH compiler... 04:10pm <AlonzoTG> It could use some LISP too. =) <Fufie> you might end up with a simpler grammar for C, but it will still be complex <jdl> fufie: yes... I suppose time will tell. That aside, I'm still interested in your overall impressions of the Prism approach. <eihrul> forth compiler? don't make me laugh :) <AlonzoTG> It'll have both stacks... <AlonzoTG> It just needs a compiler... <AlonzoTG> =\ <Fufie> the prism approach is nice, but I need to see it handle complexity at a higher level and some better examples <jdl> fufie: okay, fair enough. Will complete C isomorphism be enough? What sorts of examples are you looking for? <Fufie> real code examples <jdl> fufie: As soon as I nail down the C-like code, that's next. BTW, what's the status of SDS's C/C++ front-end effort? <Fufie> however, if the format is disambiguous it will be very useful <Fufie> the c/c++ front-end effort is still at a hold <Fufie> we have a version working for C <jdl> I saw on your web page that cppppppppp was being dropped? <Fufie> and we have three good alternative front-ends we want to re-write <Fufie> them to spit out CSF, and that takes time <Fufie> and currently I have no time this month <Fufie> but if you want to have a go at our alternatives we could take that to private conversation as some of it is "secret" <jdl> What are you looking at? The reason I ask is that I need to do the same thing in order to create a C->C-Like parser, and I'm hoping to reuse whatever I can <Fufie> cppp will probably be helpful for your use <Fufie> we can't use cppp because it lacks full template-support <jdl> well at this time I'm only interested in C anyway. (C++ is just WAY too complex. I hate that language.) <Fufie> then you should download the base sds distro and have a look at the cppp directory and the travcppp directory, as well as the cppp documentaion <Fufie> the travcppp is an example on how we get info from cppp... it is _ugly_ though <jdl> fufie: Actually, the reason I converted to Linux was so I could try out SDS. :) I'll download the distro and take a look. 04:20pm <Fufie> you might want to get ACL to make SDS useful <Fufie> but if you Linux has CLISP, that might do the trick <jdl> ACL - lisp? <jdl> nope, no clisp (that I can tell) but getting ACL shouldn't be a problem. <eihrul> acl - allegro common lisp <jdl> abi: acl? <abi> acl is Allegro Common LISP, free of charge proprietary software at http://www.franz.com <jdl> cool :) <Fufie> there is also a link to a ftp-area on the sds download page with most software you'll need <jdl> Okay, well I'm going to go take a look at the c9x spec, and also cppp -- thanks for your comments fufie. Let me know if you think of anything else. <jdl> Also, thanks for the spec, eihrul. <eihrul> welcome * jdl/#tunes is away (msg me to get my attention) 04:30pm <Fufie> eih: btw, the interesting lesson about transmeta today is that it again proves that gambling on lisp is the Right Thing as better processors will increase the validity of Lisp and make it a viable alternative -:- SignOff AlonzoTG: #TUNES (Have Nice Day :)) <eihrul> Fufie: rmm, i don't think you lisp-on-crusoe idea has any substance to it :) -:- hcf [nef@me-portland-us637.javanet.com] has joined #tunes <eihrul> and what's wrong with your average riscy processor for lisp? <Fufie> they weren't too succesful ;) <Fufie> it's spring in lisp-land.. feel the green grass between your toes <eihrul> who's going to take us into summer? <Fufie> you and me and a few others <Fufie> Lisp TNG <eihrul> maybe f-cpu can have lispy features :) <eihrul> shift and multiply! :) <Fufie> :) <eihrul> that was a joke... note 04:40pm <eihrul> not that such an instruction wouldn't be useful <Fufie> sure it would <Fufie> let's say you have foo*24.. then you shift two and multiply by six :) <eihrul> i'm talking about tagged data representations <eihrul> (you have much the same problem as with fixed point math) <eihrul> which would be another valid use of the instruction 04:50pm <Fufie> hmm.. a local mailwar stopped 05:00pm -:- SignOff jdl: #TUNES (Leaving) -:- Downix [down@d-gnaps-2.ici.net] has joined #tunes -:- air [brand@p0wer.qzx.com] has joined #tunes <lar1> Hey air 05:20pm <air> hey <air> u didnt email the pmode list yet <lar1> No, I am gonna take a look at the alignment first <lar1> And if I still cant get it <lar1> then I will mail <air> the alignment is fine <lar1> Uhh, you said it wasn't yesterday <air> i aligned it like i did in brix and it still wouldnt work <lar1> Oh <lar1> Did you disable NMI too? <air> yup <lar1> Hmmm <air> just email the pmode list and they will find it <lar1> I should clean the code first, no? <air> yes <air> and pack the gdt crap into the file <lar1> Ok -:- _ruiner_ [blah@ppp383.wi.centurytel.net] has joined #tunes <air> before emailing the list setup an idt and load it with null descriptors <lar1> Set up execptions too? <air> that wouldnt be a bad idea either <air> remap the pic and set each exception to a single handler 05:30pm <lar1> remap the pic to do what? <air> nm on the pic remap <lar1> Ok <air> setup a 32 descriptor idt table with each descriptor set to a single handler <lar1> Right -:- SignOff _ruiner_: #TUNES (Leaving) -:- SignOff lar1: #TUNES (Ping timeout for lar1[sdn-ar-019casfrMP046.dialsprint.net]) -:- lar1 [larman@dialup-209.245.130.161.SanJose1.Level3.net] has joined #tunes 05:40pm -:- Kaufmann` [newbie@dial305.infolink.com.br] has joined #tunes <Kaufmann`> Foo! <lar1> Bar! <Downix> hey <Kaufmann`> So what's up? <Kaufmann`> ah yes <Kaufmann`> I've finally gotten around to installing Linux on my li'l brother's computer <Kaufmann`> said computer is a 133MHz Pentium with a 14" monitor and a crappy video board <Kaufmann`> ...anyone? -:- SignOff hcf: #TUNES (Leaving) >>> Kaufmann` [newbie@dial305.infolink.com.br] requested PING 948333928 from #tunes <lar1> lil' bro! <lar1> How lil is yer lil bro? <Kaufmann`> He's a 12 year old who only cares about videogames. <Kaufmann`> *sigh* <lar1> Heh <Kaufmann`> LOL <Kaufmann`> [shiva.va.us.dal.net] -*Fexer02*- DCC Send MyPicture.bmp.vbs (4.54.230.33) <Kaufmann`> [shiva.va.us.dal.net] *** DCC SEND offer from Fexer02 (jimj500@4.54.230.33) (³MyPicture.bmp.vbs², 2560 bytes) <lar1> Uhhhh <Kaufmann`> the sad thing is 06:10pm <Kaufmann`> that there are doubtlessly some people who _will_ accept this "picture", download it and double-click on it. <Downix> oh man * eihrul/#tunes ponders what this strange double-clicking notion is. <eihrul> you mean there's life beyond the keyboard? <Kaufmann`> heh <eihrul> or does something weird happen when keys are clicked twice in rapid succession? * lar1/#tunes smacks eihrul with "Using you mouse" (found on lar's mac) -:- SignOff lar1: #TUNES (Read error to lar1[dialup-209.245.130.161.SanJose1.Level3.net]: Connection reset by peer) <Kaufmann`> ROTFL <Downix> lol -:- lar1 [larman@dialup-209.245.130.161.SanJose1.Level3.net] has joined #tunes <eihrul> karma <eihrul> peer doesn't like mice 06:20pm <Kaufmann`> oh yeah, I forgot to mention it! My karma jumped to 36 yesterday. <eihrul> hmm, off to read... 06:30pm -:- roblimo [roblimo@cc292052-a.hwrd1.md.home.com] has joined #tunes <lar1> roblimo? Here? <lar1> Hola roblimo! <roblimo> Why not? <roblimo> Gotta be somewhere. <lar1> True :) <lar1> So whats up? Any perticular reason for dropping by? <roblimo> I have X-Chat, just like several million others. BFD. <roblimo> I was checking #debian for a poor CNBC reporter, thought I'd drop in. <lar1> Ah, cool <Downix> hmm <roblimo> The latest trend in mainstream Linux and Open Source reporting is to call me and ask for help instead of using a search engine. <Kaufmann`> LOL <lar1> Heh <Fare> hi <roblimo> See, I'm pretty well-known in SPJ (Society of Professional Journalists) circles. <Downix> Hey Roblimo! <Fare> hum <roblimo> Fare - you bastard!!! ;-) 06:40pm * Downix/#tunes hopes you'd read his comments about your "retraction" <Fare> sorry <Fare> I didn't think things would go that way <roblimo> No, no, you did nothing wrong. You were right. <Fare> actually, there were lots of pretty interesting comments on /. that didn't get moderated beyond 2 <Fare> or beyond 1. <roblimo> I got over 200 emails telling me I should have told off Bruce and the rabble even harder than I did. * Downix/#tunes nods, I'm always a 1 <Kaufmann`> hey * Fare/#Tunes doesn't think his mother ever subscribed to the french SPJ or anything like that; she at least once failed to basic journalist double-checking. But she's usually a better editorialist than most. <Kaufmann`> what are you talking about? <Kaufmann`> what did robin do? <roblimo> I'm a +2. * Kaufmann`/#tunes is a 2 * Kaufmann`/#tunes feels like he's missing out on something * Fare/#Tunes peruses dict to understand... <roblimo> I posted a submission from Fare on /. that questioned a minor point in the HOLY GPL and got flamed to death over it. * Fare/#Tunes thanks Rik Faith for dict <Fare> actually, as was pointed out, it's as much a point about corporatism and civil rights as a point about the GPL * Downix/#tunes nods <Fare> (but isn't the GPL all about defending rights where they get stomped off by corporations with the help of state ?) <roblimo> Whatever. Both RMS and I took it as more of an ethical question than a legal one anyway, no? <Downix> roblimo: I agree that it should have been posted. Open discussion on the GPL has to happen. It's not the end-all-be-all licence * Downix/#tunes nods <Downix> Fare: yes <Kaufmann`> roblimo, I know that you posted the submission. But when did you tell off "Bruce and the rabble"? <Fare> the list is named cybernetHics, isn't it * Kaufmann`/#tunes is going to start a band called Bruce and the Rabble. <Fare> (darn, when is my bold when I need it?) <roblimo> Plus, Fare, U.S. law gives corporations more "individual rights" than do most countries. * Downix/#tunes nods <roblimo> Subtle point: Brit press refers to a company as "they" while America press says "it." * Downix/#tunes feels like such a peon here <roblimo> Downix: you're not a peon to me. ;-) <Fare> As was pointed out to me, the difficulties of the "hole" lie in work legislation and the fiscality of the would-be contracts between members and the association <Fare> particularly in over-lawed europe <Downix> Well, I'm only a 1 on /. <Downix> Only a single 2-comment, and that was on the Amiga <air> hmm i thought linus said that his work at transmeta had nothing to do with linux <Fare> anyway, beyond that hole, the point was to defend the rights of employees, which the GPL doesn't do, as it is <roblimo> Well, when I start posting less on Slashdot (very soon) people will think Bruce drove me off. ;-) <lar1> air: So? Microsoft doesn't care about Linux either... ;) <Fare> Linux? What's that? <Fare> Oops, I mean, Microsoft, what's that? :) <Downix> hey roblimo, you could post stories I submit! hehe <lar1> roblimo: Why are you going to post less? <roblimo> I'm now Editor-in-Chief of all Andover sites. <Fare> ouch * Downix/#tunes really should get a press release made for his company <Downix> Speaking of which, my firm needs a new name <roblimo> What company? <abi> company is licencing 2 different chipsets to us <air> Downix: amigo? <Fare> Downix: when will your chipset be ready? Will a further version work with Crusoe? :) <Downix> roblimo: I'm a partner in a start-up computer company <Downix> Fare: a year, and yes <Fare> abi: forget about company <abi> Fare, I didn't have anything matching about company <roblimo> Does it have a name/website? <Downix> air: ha ha, very funny <Fare> abi: forget company <abi> Fare: I forgot company <Downix> roblimo: No and no. Every name we come up with stinks <roblimo> So do most company names. <Fare> abi: company is two <Kaufmann`> Downix, I've already told you, you can never go wrong with "The Cthulhu Corporation: awakening the ancient gods... for your satisfaction!" 06:50pm <Downix> Kaufmann`: hehe <Kaufmann`> okay people, let's help Downix come up with a good name for his company. <roblimo> Just use the principals' initials. <Downix> roblimo: I don't wanna live with an idiotic name. We're not "eBox" or anything like that <Fare> no, the R'lyeh corporation. And you contact address is Chthulhu@Rlyeh.com <Downix> Sure, go as NEKAP * Downix/#tunes laughs <Fare> (hum; whois tells me the name is taken. Darn!) <Downix> Or PEKAN <Kaufmann`> (That's pronounced as "knee cap", guys) <roblimo> Sure. PANKED <roblimo> DEKNAP <Downix> roblimo: Our principles names are Nate, Kelly, Amy, Elyse and Pam -:- AlonzoTG [Alonzo@209-122-248-171.s425.tnt8.lnhva.md.dialup.rcn.com] has joined #tunes <roblimo> DEPANK <lar1> Good day, ATG <Downix> roblimo: Unless you go by last names, then it' Downes, Mabie, VanSingel, Lerner and Hutchins <Fare> rlyeh.org is better. <Kaufmann`> I don't think there are many companies with recursive names. COMPANY Obviously Manufactures Products which Are Not... urhm, Yidiotic? <roblimo> Sounds like a law firm... <AlonzoTG> fucking spare monitor... -:- lar1 is now known as lar-eating <AlonzoTG> =\ <Downix> Kaufmann`: LOL <Fare> (and there _is_ a chthulhu@rlyeh.org, apparently) <Downix> Maybe I should just name it after my companies design <roblimo> The depank group <Kaufmann`> CORP is an Organisation for Ripping off People <Fare> roblimo: while we're here. Ever heard about TUNES? <roblimo> Not until day before yesterday. <Fare> Kaufie: great! <Downix> lol <Downix> Nothing helping, but I'm getting a good laugh <roblimo> Tunes url? tunes.org? <Fare> CORP is an Organization for Racketeering People <Fare> abi: Tunes? <abi> Tunes is a free reflective computing system at http://www.tunes.org or for programming languages what the internet is to networks or like flypaper for hallucinogenic programmers or discombobulated or obnubilated or AOTA <Fare> it's also the greatest vaporware ever. <Downix> Yep * Downix/#tunes sighs <Downix> I'll never find a name I like <Fare> I need someone to kick me so it becomes real <Kaufmann`> T-Corp: We Kick Ass, Fooh! <Fare> Downix: don't. Because you won't be able to sell it afterwards! <Downix> Fare: Hrm? <Downix> Fare: I'm not designing a corp to sell. <Fare> Downix: if it's the name of something you might sell to become rich, don't take a name you like <roblimo> So I look at the tunes.org site and I go "Huh?" <Fare> wouldn't you sell it, EVER? <Fare> roblimo: ok, the webmaster sucks <Fare> someone please shoot the webmaster <Downix> Fare: no, I'm designing a legacy, not a quick buck <Fare> Downix: Downix.com ? <roblimo> I claim more software expertise than Jon Katz, but that's all. ;-) <Fare> roblimo: start at the FAQ * Downix/#tunes laughs <Downix> Sure, name it after my nick <Fare> then, depending on your questions here, we may tell you what to further read (maybe stuff in papers/) <Fare> Downix: Upwin? <Downix> lol <Downix> We've had a lot of names so far, I feel like we'll never find one we like <AlonzoTG> om? <Kaufmann`> Well, just pick a random letter. <Kaufmann`> The S Group. <Kaufmann`> and then draw a really cool-looking "S". <roblimo> This is going to take time to grok. Not something I can do in a few minutes. 07:00pm <Kaufmann`> and trademark all uses of the letter S. <roblimo> DMVLH Inc. is fine. <Fare> roblimo: if you have questions, feel free to come back. <Downix> uh... no <roblimo> Almost certainly copyrightable and the domain is almost certainly available. <Downix> roblimo: Hehe, but how do you promounce it? <Fare> Admittedly, the TUNES project is for people who want a deep grok at computer science. <roblimo> you can't use DSMRM Corp. - That's me & my wife. <Downix> lol <Kaufmann`> even better - Dot-Org Inc. Or Dot-Net, or Dot-Edu. Take your pick. In any case, it'll mean thousands of people you'll be able to rip off. <Downix> Hey, I could do MVHDL Inc <Downix> hehe <AlonzoTG> om <Fare> Dot-VHDL ? <Downix> hmm <Kaufmann`> hey, here's one I bet no one's thought of before <Kaufmann`> Linux Corporation! <Kaufmann`> :) <Downix> lol <Downix> Maybe I should call it Downix. hehe <AlonzoTG> I've been telling you that all along! <Fare> Updos 07:10pm <Fare> Flurivostuginous, Inc <Downix> lol -:- lar-eating is now known as lar1 <lar1> Call it No use for a name, Inc. <roblimo> Have fun, y'all. I need rest. I'll read up on tunes as I get a chance... -:- roblimo [roblimo@cc292052-a.hwrd1.md.home.com] has left #tunes [] <Downix> How about Names Are Messy Explination Systems ? <Downix> lol <lar1> Wasn't that cool? Robin here and all? Interested in the tunes project, no less <Kaufmann`> LOL Downix <lar1> heh <Kaufmann`> oooooh, Robin was here <Kaufmann`> we worship the ground upon which Robin walks :) <lar1> what kind of name are you looking for in all seriousness :) <Downix> Something mythological in nature <lar1> Kaufmann: Nah, I don't worship Robin. Hes a semi-celeb and its cool that he chatted with us. Thats all <Kaufmann`> YHWH Corp. <Downix> no way * Kaufmann`/#tunes chuckles <Kaufmann`> at least you'll get some respect * Fare/#Tunes worships the people Robin talks to * Fare/#Tunes worships himself <Kaufmann`> "YHWH Corporation: Because We Are Who We Are!" -:- SignOff mibin: #TUNES (Ping timeout for mibin[62.11.105.175]) <Fare> YMMV Corp <lar1> Fare: You worship me as well? <Fare> lar1: too late, I can only worship one person. <Kaufmann`> Fare is a monotheist that way :) <Downix> YUMMY corp <Kaufmann`> okay. * Downix/#tunes is a polytheist <lar1> Alyssan, Inc <Fare> Theists think all gods but theirs are false. Atheists simply don't make an exception for the last one. * Kaufmann`/#tunes figures just about every name from either Greek, Roman, Norse or Egyptian mythologies is taken. <Kaufmann`> Fare, yep. <Fare> (fortune cookie, from wherever) <Downix> Eddas is not, that's our chipset's name. <Downix> That's a part of Norse mythology <Kaufmann`> So go with Eddas Inc., dammit <lar1> Am I the only one here that likes the name Alyssa? <Downix> 2 of my partners are for it, 2 against <Downix> lar1: Think so 07:20pm <Downix> Plus there is an Alyssa Inc already <lar1> Well, then, buy them! :) <eihrul> Fare: you already spouted that fortune the other day... your random selection algorithm is pretty sad <Downix> hehe <eihrul> You know nature has not done it's job when it is beaten out by a mere program called 'fortune' <eihrul> s/it's/its <Kaufmann`> Hey, I've got it <lar1> Downix: If you spend all yer time picking over yer name, more improtant stuff won't get done... I think Eddas is fine... <Kaufmann`> Metatron Corp.: For God hath spoken unto the people of Israel: "Thou Shalt Purchase"! <Downix> lar1: Hehe <Fare> eihrul: it's not random <eihrul> nothing is... that's not the point <Downix> LOL, no thanks <eihrul> you must maximize the variety of the fortunes <Fare> Metatrans Corp <lar1> Heheh Kaufmann!!! <eihrul> random in the sense of a close-to-even distribution <eihrul> and right now its highly skewed against the theist quote <eihrul> s/its/it's <Downix> How about "tunes Corp, reflective hardware for a reflective world?" just kidding <Kaufmann`> anyone here didn't watch _Dogma_? <Downix> I watched Dogma <Kaufmann`> So you got the joke <Downix> Hehe, why not "Omni Computer Products" from Robocop? <lar1> I watched _The Wall_ -:- SignOff lar1: #TUNES (Read error to lar1[dialup-209.245.130.161.SanJose1.Level3.net]: Connection reset by peer) -:- lar1 [larman@dialup-209.245.130.161.SanJose1.Level3.net] has joined #tunes <Fare> Kaufie: what joke? <Kaufmann`> Fare, the one about Metatron Corp. <Fare> I didn't. <Kaufmann`> egads <Downix> I did <lar1> Downix: Perhaps a name is not suited to you... a symbol might be better... think, you can be refered to as "The company" (ie The Artist [forerly knowen as Prince]) ;) <Fare> egads? <Downix> lar1: hehehe, not a bad idea <Kaufmann`> ROTFL <lar1> Plus, it saves you this trouble with finding a name! <Downix> How about our name is <tm> ? <Downix> so it'd be <tm>(tm) <lar1> Heh <Downix> I'll think of it <lar1> Perhaps you should persue a latin dictonary and combinde words like "speed" and "think" <Fare> is that some C++ template code? 07:30pm <lar1> Fare: No, its HTML gone horridly wrong <Fare> speedus thinkatus <lar1> Dusatus :) <Kaufmann`> Hey <Kaufmann`> What about just Lambda (the greek letter) <lar1> Already a company <Downix> what would the web page be? <Fare> ok. 3054 <Fare> oops, 3053 <lar1> Fare: 3053 whats? <Downix> lar1: We did that for our first name, Caraid (which is Gaelic for "friendly") <lar1> Downix: Caraid is fine as long as its not speled caraid... <lar1> Carrade has the same sound I think <Downix> Well, my ex-partner got the name when we split up <lar1> Ah <Kaufmann`> well, I'm off. <lar1> In looking at the Burger King nutritinoal values for school, the menu has: Burgers, Snadwitches, Breakfast, Drinks, and condemnts/compnents. Components? What the hell? <lar1> Later Kauff! -:- SignOff Kaufmann`: #TUNES (kathyanne gets kick'd / as leaves fall on still waters / she is here no more.) 07:40pm -:- washort [washort@d146.narrowgate.net] has joined #tunes <lar1> Hey washort -:- NetSplit: tolkien.openprojects.net split from varley.openprojects.net [07:42pm] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [tolkien.openprojects.net] -:- Netjoined: tolkien.openprojects.net varley.openprojects.net -:- Fare [fare@quatramaran.ens.fr] has joined #tunes -:- air [brand@p0wer.qzx.com] has joined #tunes -:- SignOff washort: #TUNES (Ping timeout for washort[d146.narrowgate.net]) -:- Zone [Gra@tc5-1-7.cyberport.net] has joined #tunes 07:50pm -:- water [water@tnt-10-248.tscnet.net] has joined #tunes <lar1> Water! Hola! <water> yo <water> hm no hcf <lar1> He was here, but he left a few hours ago <water> np <Downix> ok, I need to get going <eihrul> water: immutable objects... <water> eih: yes indeed <Downix> see ya'll later -:- SignOff Downix: #TUNES ([BX] Beefcake! BEEFCAKE!) <water> i worked over the problems all day <water> i'm still not sure where to stand on it for the default case <lar1> water: BTW, you missed roblimo... he was here, and is interested in.. tunes! <water> i've met him, i think <lar1> Ah <water> eih: any thoughts or questions? -:- SignOff air: #TUNES (Ping timeout for air[p0wer.qzx.com]) -:- air [brand@p0wer.qzx.com] has joined #tunes <eihrul> nah, have just been reading through books/papers <water> ok, well i'm not sure just yet about the precise semantics <eihrul> you have 5 minutes to figure it out! <water> heh <water> why the deadline? :) <eihrul> because i'm probably going to go read more in 5 minutes <water> hm <water> well, if you're not going to code, i don't care what you think about the persistence spec <water> s/persistence/mutability <eihrul> what would you have me code? 08:00pm <water> just basic mutable objects... see the semantics page at the bottom <water> not much there, but enough to get some idea of what i'd like to have (though it might not be the right thing to do) <water> i'm willing to be flexible <eihrul> this is all just a wish-list <water> no, the bottom <water> forget the top stuff <eihrul> the 2 lines? <water> yep <eihrul> what about them? <water> well, they're necessary object-format descriptors <water> even if drastically incomplete <eihrul> and what am i to do with a format? <water> i don't know... some feedback would be nice, though <water> and it's something easily done in lisp -:- ult [noone@user-38lcn7b.dialup.mindspring.com] has joined #Tunes <ult> Hey! * ult/#Tunes has been considering optimization <water> i'm trying to take the self idea and fit slate into it <eihrul> its just an inert format for data though <water> hey ult <water> i know <ult> Why are decent languages so darn slow? <ult> Why are so many languages, when looked at speedwise, compared to C? <water> eih: i'm looking for ways to complete it <ult> What're we doing wrong/ <water> ult: because c allows you to spend days and days optimizing code manually <ult> water: But even poorly written C code is usually fast. <water> and the other languages don't <ult> Unless it's -really- poorly written. <water> well, learning to write good c code is part of optimization <eihrul> water: you need ways to clone an object, set its state, and evaluate a constructor as well as specify one <water> which is part of meta-object <ult> Why are all the really good Lisp compilers commercial? <water> that's what needs the most extension <ult> And the really good C compilers free? <eihrul> okay, then you need someone to specify those properties on the meta-object <water> *someone* <water> that's me, so far <eihrul> ult: because lisp was ahead of its time <eihrul> compiler technology is just catching up :) <eihrul> C has already worked its insidious grasp in long before 08:10pm <eihrul> water: er, brain fart <eihrul> s/someone/some way <water> oh <eihrul> as i explained the other day... you must wait for corrections before committing to reading any of my statements <water> well, i'm working on it <eihrul> :) <water> doing research on meta-objects <water> ok <water> (geez... beos 5 for free) <eihrul> inconsequential <water> yeah i know <eihrul> yamk :) <water> mk? <abi> mk is microkernel <water> oh sure <water> but i use it <eihrul> water: and don't you need some arbitrary default meta-object to work with to actually specify with? <water> huh? <eihrul> -to work with <water> yes -:- SignOff AlonzoTG: #TUNES (Have Nice Day :)) <water> so perhaps the clos mop would be a good start you think? <water> i was thinking of cleanly defining it over lisp, but maybe that'd work as well <eihrul> i'd prefer the not depending on lisp semantics :) <water> hm <water> ok i confused implementation with spec <water> i'll just specify it in english, but as precisely as i can <eihrul> that's the way it's usually done... 08:20pm <water> cool, my refactoring topic has fared well in the mlist -:- rares [user6417@wtrb-sh4-port187.snet.net] has joined #tunes <rares> it's all in the compiler <water> crusoe! * rares/#tunes steps back <water> transmeta! <water> 8) <water> gimme <rares> yes good things <rares> CHEAP! <water> it could emulate SELF <rares> $65 <water> 8) <rares> do you realize how much that will kill intel <water> yep -:- lar1 is now known as lar-away <eihrul> why would it need to emulate self? <water> gimme gimme... slate processor 8) <eihrul> its just software above hardware <eihrul> and self is already software <water> oh geez <water> self the bytecode spec <water> the language <eihrul> uh... self already does dynamic compilation <water> emulate the language in hardware <eihrul> no different than what the 'code morph' stuff does <rares> here's the thing <rares> now that it works <rares> how do you manage edit mode and run mode on an object <rares> in realtime <eihrul> water: you might as well just target Self as it is to a VLIW processor and reap the same benefit <water> you don't <eihrul> and reading through the manual <eihrul> it seemed it did optimizations conspicuously similar to what self did :) <water> exactly <eihrul> so far as dynamic compilation went <eihrul> except crusoe doesn't do type analysis :) <water> yeah well you can work that out dynamically -:- Zone [Gra@tc5-1-7.cyberport.net] has left #tunes [] <rares> there's a .01% of me that says see it when I believe it wait that didn't come out right <water> heh <water> of course the self mlist immediately starts a discussion about using crusoe <water> why does laurent always seem to miss the point? <rares> not enough neurotransmitters 08:30pm -:- Downix [down@d-gnaps-2.ici.net] has joined #tunes <rares> hey D <rares> show's on <rares> Al has a place to stay <rares> I'll join him a little later <Downix> ok, cool <rares> (cost me $197 for 3 days) <Downix> I just got us all in invite to a club that night, but I don't know if any of you would fit in <Downix> not too bad <rares> which night <rares> i'm coming on Sat <water> hmm... just looked at crusoe details <Downix> Oh? Sat now? <Downix> water: me too, interesting idea <rares> Al \= Fri <Downix> Hrm <Downix> Oh geesh <Downix> THis is going to be hard to work with <rares> what did you say Fri? <Downix> Am I going to have to pick up Alan? <Downix> i got the whole weekend off, till tuesday <rares> Just get him from South Station to the Motel 6 Boston West <Downix> ok <Downix> Well, it'll work <rares> dude it better <rares> $197! great deal but still * Downix/#tunes hopes his car can take it 08:40pm <water> it's a shame apple couldn't market ppc successfully based sheerly on emulation power <ult> Can PPC emulate well? <water> or ibm or motorola, for that matter <Downix> PPC is not a good emulation CPU <Downix> in fact it's horrid <water> why not? * ult/#Tunes nods <Downix> To emulate well you need to make it less complicated underneath, to allow commands to be executed swiftly <Downix> PowerPC's command structure is overly complex. Emulation is slower on it than on even the 68k -:- NetSplit: forward.openprojects.net split from sterling.openprojects.net [08:41pm] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [forward.openprojects.net] -:- Netjoined: forward.openprojects.net sterling.openprojects.net -:- Fufie [stig@tunnel-44-22.vpn.uib.no] has joined #tunes <Downix> Good examples of CPU's that emulate well are Alpha, MIPS and SPARC <Downix> their simplicity makes them much better canidates <ult> In other words, well designed RISC chips? <Downix> ult: Right. VLIW is the king of it however. <Downix> The E9k and Crusoe <eihrul> uh, the e9k is about as solid as the amiga :) <eihrul> (or atleast the 'new' amiga) <Downix> true -:- SignOff water: #TUNES (Ping timeout for water[tnt-10-248.tscnet.net]) <Downix> Ironically, a MIPS CPU can be turned into a VLIW engine w/ little modifications <Downix> Same with SPARC -:- water [water@tnt-10-74.tscnet.net] has joined #tunes <water> bleh <water> sorry, i missed whatever good reasons about ppc you posted, don't worry about repeating them, i'm sure you're right <Downix> Ironically, a MIPS CPU can be turned into a VLIW engine w/ little modifications <Downix> Same with SPARC <Downix> that's all you missed <water> oh <Downix> PowerPC is overly complex, which makes emulation poor <water> hm ok <Downix> notice, through cirtual PC the G4 gets the same equiv PC performance as a 68060 does using PCx on an Amiga <Downix> but both are CPU-centric <water> yeah <Downix> erm, Virtual PC I ment <water> i know <Downix> Anyways, PPC is just not a good emulation machine * eihrul/#tunes wonders how Crusoe handles branches... <water> well, i know for a fact that x86 ain't <Downix> yup <Downix> the best CPU's for emulation I've seen are SPARC, MIPS and Alpha <water> yeah that makes sense <eihrul> so how are branches handled? :) <Downix> I dunno, ask Transmeta -:- hcf [nef@me-portland-us845.javanet.com] has joined #tunes 08:50pm <air> eihrul: crusoe converts the code and places it in a cache. it all becomes native code <water> yeah, but what of branch-predictions? <eihrul> air: but umm... how does it translate branches to specific addresses? * water/#tunes shuts his mouth and focuses on slate/arrow <eihrul> that is what i'm asking <eihrul> air: i read the damned manual... i know about the cache and what-not <eihrul> having to signal an exception at every first jump could stink <air> maybe the jumps goes to the morph code <air> s/goes/go/ <eihrul> that'd slow things down though.... <Downix> Whatever Transmeta made, even released, it's still an enigma <air> uhh and emulation doesnt? <air> and a jump would be faster than an exception <Downix> hmm <rares> until they build interference processors the time/space ti,e/complexity tradeoff will continue to bite <rares> anyway I'm going to sleep -:- rares [user6417@wtrb-sh4-port187.snet.net] has left #tunes [] 09:00pm <water> hcf: today's "meta-object day" -:- SignOff ult: #TUNES (Read error to ult[user-38lcn7b.dialup.mindspring.com]: Connection reset by peer) <eihrul> air: gah, then you have to do really tricky things with memory <eihrul> that would slow everything down as much as exceptions -:- washort [washort@d142.narrowgate.net] has joined #tunes 09:10pm <eihrul> lo <air> eihrul: tricky things? <eihrul> well, the manual says the code morphing software is written for the native isa... <eihrul> and jumping to a dummy address with code stored there would be kind of hackish and expensive <eihrul> and then again, so would exceptions 09:20pm <eihrul> hmmm * eihrul/#tunes is thinking that if you can declare type classes to the compiler, you can just as well declare the result of any expression and not just a predicate to a given compiler. <water> for a static lang, sure <eihrul> so it'd be more useful to allow you to just tell the compiler the result of any computable expression within the languagte <washort> hehe <washort> quote from cbbrowne on slashdot -- <washort> Alternatively, the Lisp Machine people might find it a slick idea if Transmeta provided microcode to provide a Lisp-oriented instruction set that <washort> (notably) provides a really tightly microcoded set of garbage collector instructions. <eihrul> to much more benefit than a type system (totally at one's leisure however) <eihrul> uh... "garbage collector" instructions? <water> really? <eihrul> i can't think of any lisp garbage collection operation that doesn't do a lot of graph traversal <eihrul> really wha? <water> really gc instructions <water> actually that makes sense <water> nm <eihrul> gc instructions don't make sense :) <eihrul> anyone who thinks so is on some seriously strong drugs.... <water> yes if you want to force a gc <eihrul> gc is a massively complex operation... <water> so? <eihrul> and the gc instruction would be in software... <eihrul> just translated to crusoe code <water> you can always do a quick-and-dirty 80% sweep <eihrul> no benefit over regular garbage collection <water> well object memory would have to be emulated <eihrul> water: then it needs to interpret data formats and everything... <water> eih: yes it is when little space is left <eihrul> by the time you're done you have a complete vm layer :) <water> you're just being argumentative <water> no kidding <eihrul> no... how does a gc instruction help when little space is left? <eihrul> especially one emulated in software? <water> if you fill up the space too quickly between gc cycles <eihrul> gc is as expensive as gc either way * eihrul/#tunes sighs. <water> no gc is one single algorithm <eihrul> again, what's the difference of it being an instruction or just part of a vm? <eihrul> s/instruction/*emulated instruction* <water> just shut up and read <eihrul> problem? <abi> the problem is, as usualy, explaining what tunes is or that as you introduce the other missing-parts and combinations the complexity will grow <washort> eihrul: GC instructions worked on the Lisp Machine, didn't they? <water> heh 09:30pm <eihrul> washort: but people are overlooking that crusoe just has a compiler running in software.... <washort> eihrul: eh, sort of <washort> it's more like programmable microcode <eihrul> washort: manual doesn't say anything about microcode <washort> AFAICT, anyway <washort> the impression I got was that it wasn't "software" software :) <washort> an entire x86 emulation is written in it and runs near native speeds apparently - <eihrul> washort: manual just says it loads the 'code morph' software from a rom, then the ,it just translates the code and sticks it in a cache of already translated code <washort> i'd imagine that that sort of level can enhance GC as well <eihrul> the key thing to note is that its just a program written for the native crusoe isa.... <washort> right <washort> just set up a 36-bit architecture, etc :) <washort> apparently at the demo they had a layer for the JVM. i'd imagine a more useful VM/instruction set could be implemented <water> yes <washort> the other question is will we get to play with it or is that layer only available to transmeta.... <Downix> looking over the docs, looks like it's avail only to transmeta <Downix> which is a major pity <washort> that's a bummer <washort> some bright hardware hacker will figure it out within the year, i'll bet :) <washort> the word that the Crusoe chips bring to mind is 'Dynabook' <water> yeah they have to store the software layer somewhere <washort> pity they don't see it :-/ <water> that's what the squeak and self groups are chanting <washort> hee hee <washort> hardware is the main thing holding them back :) 09:40pm <washort> in terms of performance, anyway -:- lar-away is now known as lar1 <water> hm... typeless fp considered as oop with unrestrcted protocols <eihrul> 'unrestricted'? <water> every object answers every message without lookup <eihrul> yah <water> so a bounded-set theory based oo system would be a well-typed (ct-level) fp language <eihrul> or the other way around <water> yeah a=b => b=a 09:50pm <eihrul> <=> <water> i know <eihrul> okay, i'll go shut up and read now <water> but using that symbol gives the symbols too much meaning -:- SignOff air: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep) -:- SignOff eihrul: #TUNES ([x]chat) 10:00pm <water> hcf: you remember the path to tunes.org's wilma? <hcf> http://www.tunes.org/cgi-bin/wilma/ ? <water> thx <water> er... <water> no, the internal directory <hcf> bespin's? <water> yes <water> sorry, i'm in win32 with no good terminal around <hcf> seemingly, /serv/w3/wilma <water> k <hcf> u want bespin's mail archive path? <water> nah <water> got it <hcf> u gonna respond to the joy/lll/hll talk? <water> hm <water> should i? <hcf> yes, since ur putting slate in the same sorta category 10:10pm <water> well, i don't want to havve to interact with laurent <water> and i can't disagree with what anyone else is saying <hcf> at least interact w/ billy * water/#tunes thinks <hcf> hmm, billy isnt on the slate mlist :( <water> oh well <water> no, i don't have naything to say about that discussion <water> s/nay/any <water> i'm definitely considering starting a thread on co-iteration and co-recursion <water> although i'd probably be wasting time compared to actually re-writing arrow docs <hcf> r u gonna do any squeak code for slate or just leave coding up to eihrul? <water> until i have a spec that i'm satisfied with, i won't code <water> but after that, i will code, definitely <water> time is very precious 10:20pm <water> right now, i'm looking really closely at smalltalk's mop and self's mop <hcf> http://citeseer.nj.nec.com/cs?q=mop+and+smalltalk+and+self&cs=1&co=Citations&cm=50&cf=Any&submit=Search+Indexed+Articles&ao=Citations&am=10&af=Any <water> cool * water/#tunes does a few searches of his own <water> :) 10:30pm -:- SignOff washort: #TUNES (b0ink b0ink!) <water> hm smalltalk and self make things more complicated than they have to be, it seems 10:40pm <water> foo 11:10pm -:- SignOff Downix: #TUNES (BitchX: its not for the tabkey impaired!) -:- SignOff lar1: #TUNES (Leaving) <hcf> ho hum <water> welcome to the nitty-gritty of mop-analysis <water> not much of an audience :) 11:20pm <water> the damn self system is way too under-documented <water> i guess it would be simpler if i had a mac or sparcstation lying around 11:40pm <water> the nice thing is that i've got tons of docs on the smalltalk compiler here 11:50pm [msg(TUNES)] newlog 2000.0120 IRC log ended Thu Jan 20 00:00:02 2000