IRC log started Sat Mar 13 00:00:00 1999
[msg(TUNES)] permlog 1999.0313
<Fare> are you there?
03:30am
ωνω smkl [sami@ppp1.dial-in.verkkotieto.com] has joined #tunes
<smkl> "Anyone can come up with a catchy acronym and put together a web site full of buzzwords."
07:40am
<smkl> do you know if Mozart needs emacs?
10:00am
<Fare> Mozart?
<abi> Mozart is a concurrent constraint programming language at http://mozart.ps.uni-sb.de/
10:40am
<Fare> ain't it a compiler for Oz?
<smkl> anyway, i checked it and it still needs it
<smkl> yes
<smkl> (oz3)
<Fare> needs to be added to the review project at the Oz entry
<Fare> it looks like to me Mercury is pushing things in a more interesting direction
<Fare> although maybe I'm mistaken
<Fare> I need to have a closer look at Oz
<smkl> what do you mean?
10:50am
<Fare> Hum, after all, Oz seems more directed towards distributed programming, and that's what I need.
<Fare> Mercury is too much (Haskell/LISP)*Prolog 
11:10am
ωνω tcn [tcn@cci-209150250128.clarityconnect.net] has joined #tunes
<smkl> hello tcn
<tcn> hey how ya doing?
<tcn> who's here?
<smkl> Fare is
<tcn> abi, kut?
<abi> kut is a Kernel Usable for TUNES, by Basile Starynkevitch <mailto:basile.starynkevitch@cea.fr> or ftp://ftp2.tunes.org/pub/tunes/obsolete/work/low-level/kut_0_4.tgz
<tcn> thank you
<abi> pas de quoi tcn
<tcn> abi, clementine?
<abi> i heard clementine was the os being designed with the no kernel model.
<tcn> abi, prototypes?
<abi> tcn: wish i knew
11:50am
<smkl> tcn: what os are you running now?
<tcn> Linux
<tcn> bloated piece of shit
12:30pm
<tcn> I feel betrayed.. Linux is now more bloated than Windoze.
<tcn> What are you running, smkl?
12:40pm
<smkl> do you have ocaml?
<tcn> yeah, I got it just a few weeks ago
<tcn> I like it
12:50pm
ωνω tcn has changed the topic on channel #Tunes to: New Tunes FAQ: http://www.tunes.org/~bineng/faq.html
<smkl> tcn: do you run X?
ωνω pipo [ykieffer@droopy.imag.fr] has joined #tunes
<pipo> Hi there!
<pipo> Anybody awake?
<smkl> i am
<pipo> :)
<pipo> I know you!
<pipo> I've read your name in those IRC tunes logs on the tunes web, just a minute ago
<pipo> :)
<pipo> I imagine you're into tunes. What part of the project is your interest exactly?
<smkl> do you have ocaml and run X?
<pipo> ocaml is not too far away... X is under my eyes. Why?
<smkl> reflection and safety, programming in high level language
<smkl> i need to beta test my game
<pipo> ok
<pipo> is it a multi-player game?
<pipo> or do you lack caml or X right now? :-)
<smkl> you can find it from http://www.vtoy.fi/~smkl
<smkl> multiplayer
01:30pm
<tcn> I have ocaml & X
<smkl> you can test it too
<tcn> which ocaml version?
<pipo> I never used ocaml.
<pipo> (version 1.06)
<smkl> 2.0x
<pipo> So I don't know what I have to do with it (perhaps there's a README in the archive I can just read)
<smkl> the game requires 16 or 24 bpp 
<pipo> ok, forget it :-)
<pipo> I don't think we can get those NCD terinals out of 8bit mode...
<smkl> ok
<pipo> is your game related to tunes in any way?
<smkl> nope
<pipo> ok :)
<pipo> what language are you comfortable with (whe thinking about reflection)?
<tcn> NCD?
<smkl> none
<pipo> Network Computer Device(s) I think
<pipo> are you trying to design the famous HLL?
<tcn> I'm planning to write an ML compiler for Retro/Tunes soon
<tcn> I think ML is a pretty good language, so far anyway. (only used it a little bit)
<pipo> tcn: what is retro?
<tcn> pipo: my OS
<pipo> tcn: don't you like caml light and ocaml?
<pipo> tcn: I c :)
<tcn> abi, retro?
<abi> tcn: wish i knew
<pipo> tcn: what characterizes it best, besides being your creation?
<smkl> tcn: do you want to test my game?
<tcn> abi: Retro is tcn's prototype OS. It's written in assembly.
<pipo> smkl: are you working on the hll?
<tcn> smkl: I just downloaded it
<smkl> pipo: no. i currently have no time to do anything
<tcn> pipo: I'm working on HLL's
<pipo> ok
<tcn> except write games :)
<pipo> tcn: reflective ones?
01:40pm
<tcn> smkl, have you read my latest post to the tunes mailing list? About MOOOS?
<smkl> tcn: partly
<tcn> pipo: yeah, reflective. Reflection is the reason I got into language design
<pipo> tcn: do you know any reflective language?
<smkl> and i don't have time to write games, it was finished four months ago, but i haven't had time to release it
<tcn> FORTH is pretty reflective.. 
<smkl> how are you going to add reflection to ML?
<pipo> tcn: I don't know Forth, but I know HP28s RPL. Is it reflective too?
<tcn> pipo: HP28? I dunno.. I've seen an HP48, and it looked alot like Forth.
<pipo> tcn: same question for HP48 then... :)
<pipo> tcn: let's say its language is a superset of the HP28's one
<tcn> pipo: I guess so. Is it pretty easy to change a program from within itself?
<pipo> tcn: I don't think so, except if you know about binary organisation of objects, which is not described in the manual
<pipo> tcn: how do you change a program with the help of a program with a HP48?
<tcn> Any language is reflective, for example, you could write a C program that generates another C program, compiles and links it.. but that's a pain in the ass.
<tcn> pipo: Never used one :)
<pipo> tcn: yes...
<pipo> tcn: When thinking about recursive program, I first imagine interpretive systems...
<pipo> tcn: well, it doesn't necessarily make a difference, now that you can dynamically link code 
<pipo> tcn: I believe forth is more low-level than RPL, and that's probably the reason why it's supposed to be reflective
01:50pm
<tcn> My idea of "good" reflection is: A program can change itself while it's running, through a formal mechanism (not by directly modifying its machine code)
<pipo> ok
<pipo> I've looked at a language that tried to be reflective up to some point...
<pipo> it's Self... (only works on sparc)
<pipo> It's a prototype-based language (kind of more general than object languages)
<tcn> Some FORTH's (like F-PC) are pretty reflective, at the low level. ANSI FORTH isn't really.. it's just extendable.
<pipo> you can add or remove dynamically any data or method from an object...
<tcn> Sounds like what I have in mind
<pipo> (but then, funny things can happen if you do that)
<tcn> smkl: Will your game run under FreeBSD? I'm having problems with Debian linux :)
<pipo> tcn: what kind of pbs?
<smkl> tcn: perhaps
<tcn> oh, I needed the X include files, and the package management system is fucking up
<tcn> royally
<pipo> oh...
<tcn> Why do you think I installed FreeBSD last month?
<tcn> :)
<pipo> I never compiled anything at home...
<pipo> but I guess some debian people do :-)
<tcn> The binaries for FreeBSD are half the size of Linux binaries, so they load twice as fast, take up half the disk space..
<pipo> I never tried all those BSD thingies. I spend enough time on computers not trying every OS - but perhaps I should
02:00pm
<pipo> is it the same object format?
<tcn> No.. Linux uses ELF, and BSD uses a dynamically-linked a.out format.
<pipo> ok
<pipo> any limitation with that linking scheme?
<tcn> not that I know of
<pipo> perhaps they remove debugging symbols and such...?
<tcn> It has shared memory and everything. 
<pipo> even a different object format shouldn't explain such a difference
<tcn> This is with both OS's using stripped object files, for the same program.
<pipo> ok
<tcn> (Netscape.. a real hog!)
<pipo> so there's probably another reason...
<pipo> you should investigate that and propose Debian people a fix - they are nice enough as to accept it :-)
<tcn> Maybe FreeBSD provides some libraries that Linux doesn't, so they have to include these libraries in netscape
<tcn> It's not Debian. It's Linux.
<pipo> I know :-)
<pipo> did you only compare netscape's binary?
<smkl> it's netscape
<tcn> Now that Linux is popular, it's not good anymore.
<pipo> I don't agree
<pipo> although some things are not very pleasant
<pipo> (the way everybody puts junk in /etc... did that stop in 2.1?)
<pipo> you should have a look at hurd. It's definitely not too popular!
<tcn> I installed glibc2 by hand before I got Debian, and I noticed all my binaries were twice as big.. 
<tcn> ok, so it's GNU :)
<tcn> I'm not investigating this shit, because I'm writing a new OS from scratch.. I'm just putting up with Windows and *nix until my OS does what I need.
<pipo> you sound like you know what you want :-)
<tcn> My kernel is 7k, compared to Linux's 700k :)
<pipo> ?
<pipo> what does it do... just read disks? :-)
<tcn> almost :)
<pipo> or did you put that in the boot-loader? :)
<pipo> Hurd is micro-kernel, you should definitely like it
<tcn> Nahh, that sounds too slow
<pipo> it's got a linux-sized kernel doing half the work, or even less :-)
<tcn> My OS is modular *and* efficient
<tcn> You can have your cake & eat it too
<pipo> you sound like you finished it some weeks ago
02:10pm
<tcn> why yes, I finished it for the time being, and I'm working on college. Life's a bitch.
<pipo> did you release it?
<tcn> My plan was to use assembly language, write a tiny modular kernel with a few drivers and a simple compiler, and build the OS from within itself.
<tcn> I'm working on the drivers still.
<pipo> ok. you're going to compile ML into assembly?
<tcn> All I really need is: screen (done), keyboard (crude, but works), and floppy
<tcn> No, ML (or LISP or FORTH) into machine code
<pipo> can FORTH be compiled?
<tcn> I have a page about it, with a kernel image.. http://www.tunes.org/~tcn/retro.html
<tcn> Yup, FORTH can be compiled. It's a compiler-interpreter.
<pipo> so you cannot produce pure machine code (without any runtime), can you?
<tcn> runtime?
<tcn> library?
<pipo> well, a dinamyc library is the modern replacement for a runtime routine you copy into every executable, right? :)
<pipo> (by runtime I mean routines that are added to compiled program to help do their job, like interrupt-handling routines for turbo-pascal under ms-dog)
<tcn> My OS never loads the same routine into RAM twice.
<pipo> that's a nice feature
<pipo> how do you achieve that?
<tcn> Programs have access to all the routines available on the machine, in RAM or on disk.. unless of course, the routines are protected for security reasons.
<pipo> (or more precisely, when do you say two routines are the same?)
02:20pm
<tcn> Instead of big libraries, I only modules.. each routine stands alone, so to speak.
<pipo> but the same thing can be loaded twice, right?
<tcn> Really tiny modules will be combined into bigger modules, for efficiency's sake.
<pipo> (this reminds me the time when I was programming my HP28)
<tcn> Say a program needs the PRINT routine. If PRINT isn't in RAM, the OS loads it. If another program needs PRINT, it's already loaded.
<tcn> I'm doing persistent storage
<pipo> ok
<pipo> but this doesn't prevent people from writing routines twice
<pipo> (one C-addict could argue that glibc2 is also loaded just once :) )
<tcn> haha
<tcn> but it's.. 650k
<pipo> :)
<tcn> I like the fine-grained approach
<pipo> I understand, and I like it too
<pipo> I understand it got you started with FORTH
<pipo> I don't know how things will evolve with ML and LISP though
<tcn> I hope to dissolve the language distinctions.. FORTH, LISP, and ML represent three main syntaxes
<tcn> ML: 2 + 3  LISP: (+ 3 2)  FORTH: 3 2 +
<tcn> I'd like to use all three syntaxes interchangably
<pipo> what about handling parameters to functions?
<pipo> (I mean, from the point of view of the interpreter/compiler/calculus model)
<tcn> They're all equivalent.. syntax shouldn't matter
<tcn> The compiler builds a syntax tree, and it's the same from there on, no matter what syntax you wrote the program in.
02:30pm
<pipo> in the hp28, you should have different arities for the same word, depending on (part of) the data fed into it
<pipo> oops, I meant you could have...
<tcn> arities?
<tcn> is that a French word?
<pipo> perhaps :)
<pipo> arity is the number of arguments of, say a function
<tcn> oh ok, it's a math word :)
<tcn> I knew I saw it somewhere
<pipo> I've only heard it in the mouth of CS teachers in France...
<pipo> but it could well be different in your country
<tcn> I think I saw it while reading about lambda-calculus
<pipo> (sorry, but I'm going to leave soon...)
<pipo> lambda-calculus ain't maths :-)
<tcn> allright, yeah, I gotta get some work done too
<tcn> good chatting with ya
<pipo> I gotta sleep :-)
<tcn> yeah, must be after midnight there
<smkl> hey, you need to test xskijump!
<pipo> good chatting with ya too... see you some time!
<tcn> smkl: in freebsd
ωνω SignOff pipo: #TUNES (using sirc version 2.21+ssfe)
<smkl> tcn: did you try it?
<tcn> No I'm in Linux right now.. my Linux installation is fucked up
<tcn> No X header files, and no easy way to get them
<tcn> But in BSD, I already have them..
<tcn> ahh, but I haven't installed ocaml-2.0 in BSD yet, dammit :)
<tcn> Another time
<tcn> My *nix OS's are in limbo at the moment :)
<tcn> hey did you look at that new Tunes FAQ?
02:40pm
<smkl> it's ok
<tcn> need a little work
<tcn> I wonder how it looks in netscape
<tcn> the same :)
<tcn> I'm gonna take off now
<tcn> see ya
ωνω SignOff tcn: #TUNES (cramming)
02:50pm
ωνω _QZ [brand@p0wer.qzx.com] has joined #tunes
<smkl> hello _QZ
<_QZ> hi
03:00pm
<smkl> Fare: do you know sequenceL?
<smkl> the need for recursive calls in functions
<smkl> is abstracted out of SequenceL
03:50pm
<Fare> I dunno S*L
<Fare> what is it?
<Fare> do you have to explicitly use a fix-point combinator?
<smkl> not sure yet
<smkl> http://paces.geo.utep.edu/csr/documents/sequencel.html
<Fare> explicit Y is very awkward to use, to say the least.
<Fare> Ever programmed in SKI?
<Fare> (or another combinator family)
04:00pm
<smkl> nope
04:10pm
<Fare> well, you're lucky.
04:30pm
ωνω SignOff smkl: #TUNES (sleeping ...)
ωνω tcn [tcn@cci-209150250118.clarityconnect.net] has joined #tunes
<tcn> So, who's here?
>>> tcn [tcn@cci-209150250118.clarityconnect.net] requested PING 921379120 56289 from #tunes
06:40pm
ωνω lar1 [Avzman@208.255.162.70] has joined #tunes
ωνω SignOff lar1: #TUNES (Leaving)
ωνω SignOff tcn: #TUNES (tcn has no reason)
ωνω NetSplit: koontz.openprojects.net split from forward.openprojects.net [07:41pm]
ωνω BitchX: Press Ctrl-F to see who left Ctrl-E to change to [koontz.openprojects.net]
ωνω Netjoined: koontz.openprojects.net forward.openprojects.net
ωνω abi [nef@bespin.cx] has joined #Tunes
ωνω Fare [rideaufr@esmeralda.enst.fr] has joined #Tunes
ωνω NetSplit: varley.openprojects.net split from sterling.openprojects.net [09:03pm]
ωνω BitchX: Press Ctrl-F to see who left Ctrl-E to change to [varley.openprojects.net]
ωνω Netjoined: varley.openprojects.net sterling.openprojects.net
ωνω Fare [rideaufr@esmeralda.enst.fr] has joined #Tunes
!sterling.openprojects.net!! Remote CONNECT norton.openprojects.net 8005 from lilo
!sterling.openprojects.net!! SERVER Numeric Collision: services1.openprojects.net != services10.openprojects.net
!sterling.openprojects.net!! Remote CONNECT vinge.openprojects.net 8005 from lilo
!sterling.openprojects.net!! Remote CONNECT norton.openprojects.net 8005 from lilo
!carter.openprojects.net!! Remote CONNECT koontz.openprojects.net 8005 from lilo
!vinge.openprojects.net!! Remote CONNECT koontz.openprojects.net 8005 from lilo
!doogie:*! The Great Roaming CONNECT of '99!
ωνω NetSplit: devlin.openprojects.net split from asimov.openprojects.net [11:00pm]
ωνω BitchX: Press Ctrl-F to see who left Ctrl-E to change to [devlin.openprojects.net]
ωνω Netjoined: devlin.openprojects.net asimov.openprojects.net
ωνω Fare [rideaufr@esmeralda.enst.fr] has joined #Tunes
ωνω _QZ [brand@p0wer.qzx.com] has joined #Tunes
ωνω Tril [dem@sloth.wcug.wwu.edu] has joined #Tunes
[msg(TUNES)] newlog 1999.0314
IRC log ended Sun Mar 14 00:00:00 1999