IRC log started Fri May 19 00:00:01 2000 [msg(TUNES)] permlog 2000.0519 -:- water [water@tnt-10-129.tscnet.net] has joined #tunes hey all re corey, mark hm -:- karltk [nobody@firewall.prosalg.no] has joined #tunes hi karl -:- karltk is now known as Ghyll hi water oh hey ghyll I'm suddenly not as interesting anymore, eh ? := ) hehe 12:50am i just figured out some interesting things, but it will take time to process into easily-digestible form for tunesers sounds promising. yeah i've also figured out that if i get Fare to formalize the tunes specs enough, i'll be able to point out what's flawed with the design and distinguish it from arrow btw, i'm not taking shit from Fare anymore and i think i've decided to go to Stanford when i get out of the military, or at the very least apply with fervor he's pulling academic rank on you ? :) he always does implicitly of course isn't attending uni in the us extremely expensive ? well i'm probably looking at $100k over 5 years however i think i'll be able to swing some scholarship money (given that my performance in my junior year at college doesn't cloud their view of me) just have some high official write you a recommendation letter, and they'll overlook that easily (isn't that how it works ?) i believe so, but i want to be sure of it 01:00am i *did* get a full ride promised from MIT when i left high school, so there must be something i'm worth to people it was unfortunate that i could not accept why not ? legal/financial issues i had a trust fund, but it didn't apply to private education so i had the money (technically), but it wasn't applicable in practice you're an orphan ? heh i wish no, my parents are alive and poor, it's my grandparents that i never met who stuck me with that aha. ok. I won't ask more. 01:10am well, i tested out of a little over 2 years worth of classes, so my first year was my junior year and my last this was at which college/uni ? texas a&m 01:20am a&m ? accounting and management ? :) heh agriculture and mechanics (engineering mostly) what did you study/dabble in there, then ? well i was officially declared an aerospace/mech engineering double major however, i spent most of my time doing cs and math research on my own i pretty much planned on dropping out from day one heh. now you're in SF, right ? i ran a computer lab for a while, tutoring engineering students on Maple and math in general and acting for the sysadmin no, seattle ah. next-door neighbour to mr. gates :) :P military service isn't compulsory over there, so why did you bother ? (I haven't asked this before, have I ?) no you haven't asked any regular job would have gotten me spotted as too bright a fellow from day zero which would have caused me some problems i would not like to deal with especially as i like devoting my higher-order thoughts to my more useful ideas I didn't think being bright was a genuine problem over there.. well it's the internet/VC craze how so ? they would categorize me as a programmer, anyway 01:30am isn't programming what you do now as well ? hehe no i do very little actual programming you sit bitching on the actual programmers ? heh no the stuff i work on requires thought i take care of semantics issues and make sure what gets programmed does what Fare or i or whoever says it should do i also write toy code to test ideas quite often and throw all of it out when i'm done with it you work with Fare ? not quite or do you now refer to tunes again ? i work with what he has written ah. ok. working *with* someone implies that they do work as well it's funny... when i was introducing tunes to a programmer irl recently, one of their remarks about the tunes docs was that the writer must have been on drugs when he wrote it... and the fellow actually understood the language ideas well, I personally don't quite connect to the documentation at tunes.org I'm probably too pragmatic these days. 01:40am the documentation is terrible anyone with any sense and some academic knowledge of the sources Fare is using would know how badly he's munged the explanations of the ideas heh i don't even care any more if he hears the fact that he protects his precious documentation with cvs is completely laughable as though muting his words would be some kind of vandalism do I sense a touch of hostility here ? no just frustration because my goals are relatively similar and he has been making a mockery of them for the last 6 years the fact that the site has 4 mirrors doesn't help much I've obviously missed out on something major here.. eh? like what? let's just say that I've failed to notice any of this before. have you been on the mailing list? no. I get enough mail as it is. oh ok heh the tunes mlist is not often busy anyway 01:50am perhaps I should add it, then. it's not likely you would get much out of it not until there's consensus on what the tunes language should be 02:00am * cwr/#tunes is away: watching the shuttle go up in 50mins is there a webcast of that ? 02:30am -:- Kyle_L [kyle@cr168790-a.nmkt1.on.wave.home.com] has joined #Tunes -:- SignOff Ghyll: #TUNES (Read error to Ghyll[firewall.prosalg.no]: Connection reset by peer) water? 03:00am -:- Ghyll [nobody@firewall.prosalg.no] has joined #tunes does anyone here find self, and its lack of classes, messy? 04:20am i wouldn't say prototype based OO is messy prototype? Are you saying SELFs goal is to be a prototyping environment? no i'm saying self is prototype based (classless) OO language SELF says to have no distinction, yet uses Traits and Mirrors to achieve what was lost when classes were removed. I see features, but no benefits. well, it just makes the language simplier simpler to make the language, or simpler to use? to make or learn 04:40am I can see that. I would disagree on simpler to use. The distinction between the definition/model of a thing and the thing itself adds clarity to the design of a program it's not like you can't simulate a class-based language if you really want it.. True. And you see them do that with Mirrors and Traits. Simulating is a hack. :) I can simulat OO in C if I wanted. It is another level of automatic checking. to a certain extent. you wouldn't have encapsulation. although, I'm not really sure I like hard encapsulation.. MMMMmmmmm. Explain, this sounds good. in c++, as you know, you have hard encapsulation, there's no general way to read 'foo' in class Foo { int foo; }; (or rather, there's no general way to reach it in a little more involved example) (in this case, you could just cast to an int) it would be much better if you could get at foo by using a special syntax. however, that would mean that c++ programmers would have to grow up. sounds like reflection, not encapsulation such as Foo.@foo or whatever nastyness Bjarne could concoct. not really. reflection is about a program being able to inspect itself. this is just about a way to reach encapsulated data members without nasty pointer hacks at compile time. 04:50am I thought you meant encapsulation like in encapsulate/inherit if you need to reach them, you should just unencapsulate them It applies there, too. I was just arguing that you can't get real OO in C, but you can think OO. smkl: it's not always what one wants. some times you'd like to inherit access. this is especially painful in C++ with 'friend' Ghyll: I do not think SELF has an advantage over SmallTalk wrt encapsulation. It is only fair to compare reall OO lang. Poor C++ is at an extream disadvantage. Kyle_L: agreed. Kyle_L: but self is more of an experimentation than anything else. and it's cute. Come to think of it I do not know how encapsulation is relevent to a comparision between a prototype and class based OO. 05:00am Kyle_L> I can simulat OO in C if I wanted. <- that's where encapsulation came in. Oh. :) Is there a formal name for C++ like OO and Smalltalk OO wrt encapsulation? probably, but I can't remember any on the top of my head. thanks both. -:- NetSplit: fontana.openprojects.net split from tolkien.openprojects.net [05:07am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [fontana.openprojects.net] 05:10am kyle: you are really fscking amazing i don't think i've ever heard of someone ask about self and think of its features as disadvantages the first time around :) you must have this really twisted perspective of cs Some people have told me that. like seeing everything upside down or something screwed up like that :) where the hell did you learn cs from originally? Myself in high school, then some formal university amazing 05:30am Please don't blam the university for my errors. why not? I think I was like this when I got there. university cs students often come here with screwed up perspectives oh well, this is a pointless pursuit They all don't like SELF? Yes. bbiaf OK, versioning. I looked at a paper I read. It only refers to other papers. LOL There is historic, logical, cooperative and compositive versioning. you actually found a paper on versioning?? * water/#tunes shakes his head Yes!, and it was not hidden behind the proprietary doors! get a life there are so many other more interesting concepts than versioning unless of course you want to continue to mis-use the term Versioning appears to be change, usually associated with some constant. gah! why am i sitting here wasting my thoughts on this? You mentioned that versioning only happens wrt an interface. -:- water [water@tnt-10-129.tscnet.net] has left #tunes [] I guess that was a good mood. 05:40am he's been slighly touchy today :) 05:50am -:- smoke [smoke@15dyn206.delft.casema.net] has joined #tunes -:- SignOff Ghyll: #TUNES (Ping timeout for Ghyll[firewall.prosalg.no]) -:- Ghyll [nobody@firewall.prosalg.no] has joined #tunes -:- Netjoined: fontana.openprojects.net tolkien.openprojects.net -:- smoke [smoke@15dyn206.delft.casema.net] has joined #tunes -:- SignOff Ghyll: #TUNES (bah) * cwr/#tunes is back -:- SignOff Kyle_L: #TUNES (Leaving) -:- Kyle_L [kyle@cr168790-a.nmkt1.on.wave.home.com] has joined #Tunes -:- _out [magsilva@200.201.30.78] has joined #tunes re err 09:10am anyone: I was talking to water, he start to help me understand migration. It is the ability to maintain description of things across platforms. It is the use of views to focus on particular aspects of a thing It is the realization that the views may never be conplete; so migration must support the ability to handle new views. Do I have it right? 09:20am more or less do you say that to be passive? yours are valid point of views on migration what less? Are thier others? other points of view: migration is about being able to reify programs and reabsorb them in different hardware|software contexts water: thanks for your catrewrite links -:- SignOff _out: #TUNES (Read error to _out[200.201.30.78]: Connection reset by peer) -:- _out [magsilva@200.201.30.78] has joined #tunes So reification is an elaboration on the last half of my last point. ? you may see it like that did you read http://tunes.org/Migration/ ? Yes. I know the page is obsolete and all that; if some parts aren't clear, that's a bug that should be fixed 09:30am (and at least reported) yes. A person, me, can read through Migration and only see "hardware independence using meta objects". there's more to it than that. It's about _dynamic_ hardware independence not just static one the difference being? in one case, you move source code, and recompile on each target, but once the program is launched, it cannot move anymore -:- SignOff _out: #TUNES (Leaving) in the other, you move running code, including its dynamic internal state and external connections, while it is running dynamic can still be done without knowledge of my third point, or reification. (correct me if I am wrong) you can't dynamically move running code w/o reification Example? write a small program with one or more running threads; run it on your favorite 21-bit computer; how do you move it to your brand new 64-bit computer? if you want to move w/o reification, all you can do is move whole-programs on same architectures. not small subprograms accross architectures I could have a compiler that takes a hardware description and compiles pieces of code. The system can have a bit-level description of interprocess communication. 09:40am What is being reified? The new computer architecture? yes :) and that's _much_ more expensive than reifying high-level stuff I can see that. plus when you migrate back, you have a hardware emulator of a hardware emulator of yourself, instead of running the program directly constant factors _matter_ yes. (especially very large constant factors) oh yes! -:- _out [magsilva@200.201.30.143] has joined #tunes But, when programs/things are always implemented in source ( not binary ), why not stick to moving source if that is simpler? running state I thought reification was just the explicitness of information, not also the addition of new information. if you move your e-commerce application from your old server to your new one, I bet ou don't want to lose all the existing state of your system then, map state to first order before atempting to move it. same for your ongoing quake24 game, when you upgrade your computer from dual K9 to quad Alpha 21464. ( to migrate a relational database, express it in SQL, and evaluate it when it reaches its destination ) Or when you dock away your VeryStrongARM PDA from your FCPU base station "express it in SQL" is precisely what the system ought to be doing transparently and efficiently it does! if your program is a search tree, or a physical simulation, you sure do not want it to do a SQL query at every computation step ( more or less, anyway :) you want it to save stuff only when necessary, plus periodically, for safety against power failure. No, you would probably use lisp instead of SQL Could be done! anyway. You don't want to do explicit marshalling, and you don't want to do it at every step you want it to be done implicitly, and on demand "first order" is the trick when it comes to moving things in and out of contexts sure, sure. Zeroth order, actually. 09:50am reifying is making code into data -- order zero. Zeroth? Isn't '24' a first order integer? no, it's order 0 fun x -> x + 1 ;; is a first-order function on integers how about in php you can write an array as Array( a, b, c ). I thought that was "first order"? fun f -> fun [] -> [] | h::t -> (f h)::(map f t) ;; is a second order function on integers Array is a first-order function, if you like but the resulting array is order-zero it's a basic data type, if you prefer -:- hcf [nef@207-172-225-102.s102.tnt1.pld.me.dialup.rcn.com] has joined #tunes ok, so if I said '( whatever )' then that particular saying is first order while the effect of the saying is zeroth order? ( bad sentence :) With reflection, then, everything is 0th or 1st order, depending on how you use it? rather, everything can be reified into 0th order data I was thinking that. Ok, then what I mean, is to move things in their form prior to reification, I think what's tricky however is to draw boundaries between object/modules so that you know what is reified, and what isn't is reification slightly related to the word 'absorbation'? (there are also performance issues, but that's a matter of reification-aware compilers and optimizations) absorption is the opposite of reification ok! it takes a 0th order object representation and rebuilds the object from it please explain I would imagine the reification of things can be incomplete. so that "migration = reication;communication;absorption" YES! YES! (too) of course, with proper static knowledge, this equation can be optimized by merging those three conceptual steps I should write this down so that all newbies can see. -:- _out [magsilva@200.201.30.143] has left #tunes [] absorbation/reification are like parse/unparse Kyle_L: I should write this on Migration/index.html 10:00am thomas: yes difficulty: reification must draw the boundaries of objects; absorption must connect the object to an external context that's what references are for not as simple; at least, there must be programming-language support for that Have you seen good solutions? I looked at SELF, they have a primitive (very) migration mechanism based on logging changes I beleive. 10:10am if any one cares: astro linux, http://bima.astro.umd.edu/nemo/linuxastro/talk/ Thanks Fare for another enlightening talk. -:- water [water@tnt-9-88.tscnet.net] has joined #tunes any FORTH-programmed telescope in it? 10:20am catrewrite links? abi: categorical rewriting? categorical rewriting is at http://www.mcs.le.ac.uk/~nghani/catrew.html oh yeah re all, btw err, lo! hey time for a comprehensive clarification on migration docs, wouldn't you say? :) :) btw i have this paper on how the programming language involved affects *what* can be reified water: url? hm i'll look 10:30am btw, you would probably find it helpful to crawl through RI using keyword searches and citation and relevance relations they provide 10:40am hm there's also "reification without evaluation" from bawden, but i can't locate the paper in question atm water: its here: http://achilles.bu.edu/bawden/ well yeah i know where bawden's at... we discussed that yesterday but the paper i originally mentioned isn't there whats the oribinally mentioned paper? Fare: just a suggestion, perhaps you should classify the different uses and levels where partial evaluation is interesting for tunes systems it was a paper discussing how reified concepts are affected by the expressiveness and nature of the programming language used 10:50am i think i found it, one sec "A Reflective System is as Extensible as its Internal Representations: An Iluustration" by John Wiseman Simmons II and Daniel P. Friedman the url eludes me not quite what i remembered it as being, but still related ok i could dcc the paper to you mail it, rather k 11:00am -:- hotboy [asdsddfk@02-moul-04e.dial.optilinkcomm.net] has joined #tunes -:- SignOff hotboy: #TUNES (Leaving) abi: rseir is "a Reflective System is as Extensible as its Internal Representations: an iluustration" at http://citeseer.nj.nec.com/did/59073 and ftp://ftp.cs.indiana.edu/pub/scheme-repository/doc/pubs/iucstr366.ps.gz wow thx hm i did a nearly direct search for it and didn't turn that up oh well * water/#tunes follows RI links, trolling for papers ooh i got it by searching for John Wiseman Simmons well there you go hm lots of p/e/ papers (excellent ones) linked to from rseir er... p/e even 11:10am -:- HenZo [henzo@bespin.dhs.org] has joined #tunes hey henzo sup water how's tril? Just fine so why haven't we seen him lately? :) Hasn't he been around here? hell, no hehe I'll encourage him to pay a visit i've got a web designer who volunteered to work on the site, btw he works at aa.net ... lead sysadmin really... yes i decribed the tunes technical goals and he liked it right off the bat I'll run that by Tril Cool although he thinks that Fare's on drugs :) please do -:- redfive [redfive@bespin.dhs.org] has joined #tunes also, i'd like to know the status of the web db hi, red5 hi Haha 11:40am hm another bespin fellow The web db isn't exactly fully working/done anyway, i've got slate nearly final in specs oh cool and eihrul's working out a compiler Nice (dynamic, optimizing etc) so we just might want to have a logo soon ;) self-synthetized drugs Heh, hi Fare Fare: no doubt gakuk right so anyway, if tril would contact me or at least show up on irc sometime to discuss the status, i'll try to get my friend online at the same time so ppl can hash things out mailto:dem@tunes.org usually when i do that, i get nothing in response ytalk dem when logged in phone him Sure thing, I'll let him know today. it's happened so many times that i just gave up When I want to contact him, I manage thx that's nice Fare Tril is hard to get a response from, but I can do it for you. No worries. k A'ight well I got class now, so I'll talk to you later. I'll get ahold of Tril. Hasta. cya -:- SignOff HenZo: #TUNES (Tunes? What's that?) I'll go with him -:- redfive [redfive@bespin.dhs.org] has left #tunes [] shuttle launch was awesome this morning, nice orange dawn on the smoke trail. i really miss that there was a year when i lived in orlando FL when i went to see the shuttle launches every time i havent gone in a while, growing up here its like an everyday occurance it seems grr all the tourists, "is this a good place to watch?" yeah that's what ppl told me when i was there 11:50am it never got old for me, but then i was into aerospace engineering until i dropped out (i still am, really) every dork in florida wants to work at nasa at one time its just it seems they go up every other day :) 12:00pm cwr: w? 12:10pm Fare: space shuttle launches * water/#tunes starts reading his papers collected on architectural and protocol reflection -:- morton [morton@98AFB88D.ipt.aol.com] has joined #tunes hi morton hi water is there anything you want to talk about? (pretty quiet right now, otherwise) nah ok, i can answer questions for ya anyway what is the meaning of life? heh really? "heh"? some answers require greater access privileges than others ;) 12:20pm wow, "heh" ..... never would've imagined :-) morton: 42 no, that's the answer to life, the universe, and everything i hate that book