IRC log started Sat May 29 00:00:00 1999
-:- SignOff HickServ: #TUNES (Read error to HickServ[209-68-229-54.dialup.cust.tfb.com]: Connection reset by peer)
-:- HickServ [thrustit@209-68-229-54.dialup.cust.tfb.com] has joined #tunes
<HickServ> blarg
[msg(TUNES)] permlog 1999.0529
<_ruiner_> heh
<_ruiner_> wb
<_ruiner_> did you catch all that?
<HickServ> i don't think so
<_ruiner_> I'm trying to figure out a way to make drivers transparent to each other....so they can share resources without conflicting
<_ruiner_> that way I can go with a compiled kernel like *nix os's.....but use win/dos/linux drivers for stuff I haven't written drivers for
<_ruiner_> kinda layering them
-:- SignOff Tril: #TUNES (Ping timeout for Tril[sloth.wcug.wwu.edu])
<_ruiner_> ever hear of flux?
<HickServ> nope
<_ruiner_> they tried to do something like that
<_ruiner_> using drivers from other os's
-:- Tril [dem@sloth.wcug.wwu.edu] has joined #TUNES
-:- mode/#Tunes [+o Tril] by ChanServ
<_ruiner_> I'm thinking of going with buddy lists for memory management too
<HickServ> i'd have to look deeper at the concept to fully understand it
<HickServ> buddy lists?
<_ruiner_> yeah
<_ruiner_> have billy explain it to you
<_ruiner_> I get most of it, but I"m not sure about all of it
<_ruiner_> basically though, its a couple double linked lists that divide memory into smaller and smaller pieces
<HickServ> brix has cool driver support
<HickServ> oh
<_ruiner_> heh heh heh
<_ruiner_> if you really want to make something just insanely fast....go with an exokernel
<HickServ> why would you divide it up?
<HickServ> yeah i heard
<_ruiner_> to keep track of it and allow programs to use it
<_ruiner_> and it simplifies it a lot
<HickServ> i'll talk to billy about it
<_ruiner_> yeah.....I didn't get some of it....I was pretty tired at the time
<_ruiner_> how old are you by the way?
<HickServ> have you heard of nanOS?
<HickServ> oh
<HickServ> 14
<_ruiner_> cool
<_ruiner_> and no I haven't
<HickServ> oh
<HickServ> it's basically i really small kerenel
<HickServ> things are getting out of hand!!!
<_ruiner_> wait.....maybe I have
<HickServ> but it will have crappy downtime
<_ruiner_> is it on the big list on the tunes site?
<HickServ> uh
<HickServ> no
<HickServ> it's on the os webring
<HickServ> yeah im still debating how i want to organize my drivers
12:10am
<HickServ> of course im still debating on how i will organize the damn kernel
-:- SignOff BlackPhoenix: #TUNES (Read error to BlackPhoenix[adm.univd.kharkov.ua]: EOF from client)
-:- BlackPhoenix [black-phoe@adm.univd.kharkov.ua] has joined #tunes
<_ruiner_> my first os will be how I described.....my second....I think I'll give an exokernel a shot
<HickServ> yeah but they look unstable
-:- SignOff BlackPhoenix: #TUNES (Read error to BlackPhoenix[adm.univd.kharkov.ua]: EOF from client)
<HickServ> but of course people will like it if ou find an efficient way to keep it stable
<HickServ> damn why'd  i use the word look
<_ruiner_> yeah, I think (from how it was explained to me) that it would introduce a lot of bugs itself
<HickServ> i must be getting tired
-:- BlackPhoenix [black-phoe@adm.univd.kharkov.ua] has joined #tunes
<HickServ> but microkernels and mmonlithic are overdone 
<HickServ> yeah i'd go with exokernels, chache kernels, or no kernels (or mabye even your own idea)
-:- SignOff BlackPhoenix: #TUNES (Read error to BlackPhoenix[adm.univd.kharkov.ua]: EOF from client)
<_ruiner_> no kernels?
-:- BlackPhoenix [black-phoe@adm.univd.kharkov.ua] has joined #tunes
<HickServ> ask fare 
<_ruiner_> how would you go about that?
<HickServ> im totally vague on the concept
<_ruiner_> fare?
<abi> it has been said that fare is the one to ask about emacs or french and they like to have dead cricket spitting contests
<_ruiner_> helllllloooooooooooo?
<_ruiner_> lol
<_ruiner_> leonardo dicaprio?
<abi> hmmm... leonardo dicaprio is the fourth Hanson brother
<HickServ> the guy who is sleeing
<_ruiner_> I'm just playing with abi now....
<HickServ> i'll give you a good(hah yeah sure) url though
<_ruiner_> I don't know how in the world you'd go about that
<_ruiner_> especially if you wanted it to be multitasking
<_ruiner_> I could see if he meant more like microkernel.....
<HickServ> http://www.tunes.org/papers/Glossary/index.html#nokernel
<HickServ> well i must go now
<HickServ> goodnight
12:20am
-:- SignOff HickServ: #TUNES (bye)
-:- SignOff BlackPhoenix: #TUNES (Read error to BlackPhoenix[adm.univd.kharkov.ua]: EOF from client)
-:- BlackPhoenix [black-phoe@adm.univd.kharkov.ua] has joined #tunes
-:- SignOff BlackPhoenix: #TUNES (Read error to BlackPhoenix[adm.univd.kharkov.ua]: EOF from client)
-:- BlackPhoenix [black-phoe@adm.univd.kharkov.ua] has joined #tunes
<_ruiner_> wb
<BlackPhoenix> always glad
-:- SignOff BlackPhoenix: #TUNES (Read error to BlackPhoenix[adm.univd.kharkov.ua]: EOF from client)
-:- BlackPhoenix [black-phoe@adm.univd.kharkov.ua] has joined #tunes
-:- SignOff BlackPhoenix: #TUNES (Read error to BlackPhoenix[adm.univd.kharkov.ua]: EOF from client)
-:- BlackPhoenix [black-phoe@adm.univd.kharkov.ua] has joined #tunes
12:30am
-:- SignOff BlackPhoenix: #TUNES (Read error to BlackPhoenix[adm.univd.kharkov.ua]: EOF from client)
-:- BlackPhoenix [black-phoe@adm.univd.kharkov.ua] has joined #tunes
-:- SignOff BlackPhoenix: #TUNES (Read error to BlackPhoenix[adm.univd.kharkov.ua]: EOF from client)
-:- SignOff _ruiner_: #TUNES (Leaving)
-:- _ruiner_ [nate@ppp080.wi.centuryinter.net] has joined #tunes
-:- SignOff _ruiner_: #TUNES (Leaving)
-:- insight [black-phoe@adm.univd.kharkov.ua] has joined #tunes
-:- SignOff insight: #TUNES (Read error to insight[adm.univd.kharkov.ua]: EOF from client)
-:- _ruiner_ [nate@ppp375.wi.centuryinter.net] has joined #tunes
-:- SignOff _ruiner_: #TUNES (Leaving)
-:- smoke [smoke@8dyn3.delft.casema.net] has joined #tunes
<Fare> hum
08:20am
-:- ruiner [blah@ppp296.wi.centuryinter.net] has joined #tunes
-:- AlonzoTG [Alonzo@209-122-220-23.s277.tnt5.lnh.md.dialup.rcn.com] has joined #tunes
-:- SignOff ruiner: #TUNES (Leaving)
-:- NetSplit: koontz.openprojects.net split from sterling.openprojects.net [09:38am]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [koontz.openprojects.net]
-:- Netjoined: koontz.openprojects.net sterling.openprojects.net
-:- Fare [rideau@quatramaran.ens.fr] has joined #Tunes
-:- _1BadDude [Alonzo@216-164-225-25.s279.tnt6.lnh.md.dialup.rcn.com] has joined #tunes
-:- SignOff AlonzoTG: #TUNES (Ping timeout for AlonzoTG[209-122-220-23.s277.tnt5.lnh.md.dialup.rcn.com])
-:- srjones [user7476@pm4-1-469.orf.infi.net] has joined #Tunes
-:- srjones [user7476@pm4-1-469.orf.infi.net] has left #Tunes []
-:- SignOff _1BadDude: #TUNES (Have Nice Day :))
-:- AlonzoTG [Alonzo@209-122-252-32.s286.tnt1.lnh.md.dialup.rcn.com] has joined #tunes
-:- tcn [tcn@cci-209150250096.clarityconnect.net] has joined #tunes
<tcn> help
<tcn> anyone here know anything about IDE controllers?
<tcn> I can read the disk ok.. the "identify" command works too.. but after I do one IDE access, IRQ 14 stops working..
<AlonzoTG> =\
<AlonzoTG> never tried it. 
<tcn> maybe abi knows
<tcn> abi: help!!
<tcn> fare?
<abi> fare is, like, the one to ask about emacs or french and they like to have dead cricket spitting contests
<tcn> abi, seen fare
<tcn> abi: seen fare
<tcn> seen Fare
<abi> Fare was last seen saying something on IRC Sat May 29 08:10:28 1999
<tcn> dohhhh!  :)
<tcn> abi: time?
<abi> time is 12:17:59 1999
<tcn> well, I'll figure it out myself
<tcn> anyway, I almost have it working
12:20pm
-:- SignOff tcn: #TUNES (back to work)
-:- iStormy [stormy@rain.futuresouth.com] has joined #Tunes
<Fare> hum
* Fare/#Tunes is back
-:- hcf [nef@me-portland-us225.javanet.com] has joined #tunes
-:- SignOff smoke: #TUNES (Read error to smoke[8dyn3.delft.casema.net]: EOF from client)
12:40pm
<Fare> hum
-:- HickServ [thrustit@209-68-229-62.dialup.cust.tfb.com] has joined #tunes
-:- SignOff Fare: #TUNES (Ping timeout for Fare[quatramaran.ens.fr])
12:50pm
<HickServ> hello
-:- Fare [rideau@quatramaran.ens.fr] has joined #Tunes
<HickServ> wb fare
<AlonzoTG> om
<HickServ> ommmmmmmmm
<HickServ> how's it going mr grimes?
<HickServ> #os on efnet sux0rs now :(
<AlonzoTG> damn bot channel
<HickServ> yeah
<HickServ> oh well
<HickServ> at least we have #tunes
01:00pm
<HickServ> so how's al's great os coming?
<AlonzoTG> om
<AlonzoTG> suck
<AlonzoTG> unless I want to make a DOS clone =P
<HickServ> no
<AlonzoTG> stuck
<HickServ> where you khaving problems?
<AlonzoTG> linking and loading. =\
<HickServ> could i help?
<AlonzoTG> perhaps
<HickServ> you tried coding any of it yet?
<AlonzoTG> yepers
<AlonzoTG> just havn't figured out how to link it...
<HickServ> no what exactly are you trying to link again?
01:10pm
<HickServ> now = no
<AlonzoTG> objeckt files
<AlonzoTG> to make loadable modules...
<HickServ> oh like linking many of em
<HickServ> i see
<HickServ> i'll go check
<AlonzoTG> then to actually load the modules...
<AlonzoTG> there are some linkerz avail but I need to write/integrate my own...
-:- tcn [tcn@cci-209150250096.clarityconnect.net] has joined #tunes
<tcn> well I got my IDE driver working
<AlonzoTG> k00l
<HickServ> alonzo: I seem to have misplaced my c++ book
<HickServ> what peice of hardware does it dirve?
<AlonzoTG> oh well
<HickServ> damn that arrow system paper is large
<tcn> HickServ: a hard disk
<HickServ> :)
<HickServ> for retro?
<tcn> yup
<HickServ> cooly
<tcn> yeah, sweet
<HickServ> where'd you get the specs?
<tcn> well, the best specs are at http://tunes.org/~qz/borg.qzx.com under the OSDP directory
<HickServ> yeah i've read those
<tcn> the one I started with is there, too (it's called ide.txt)
<HickServ> cooly
<tcn> I looked at some docs from Quantum too
<HickServ> im writing a keyboard driver right now 
<tcn> hehe
<HickServ> just for practice
<tcn> kbd drivers are fun :)
<HickServ> yeah
<tcn> ide is a piece of cake, by comparison
<HickServ> i used t think sriver writing is realy hard and crap
<HickServ> it's not that hard actually
<HickServ> depending on the hardware
<HickServ> someone just needs to write a text to get people started
<tcn> as long as you got decent docs
<tcn> Art of Assembly is pretty good
<HickServ> definately
01:20pm
<HickServ> i've only read half of it though
<HickServ> the url won't  work anymore
<tcn> qz's got it archived on his site ;)
<HickServ> what dir?
<tcn> under blah/
<HickServ> found it
<HickServ> yeah
<tcn> everybody leech bespin :)
<tcn> make sure ya use binary mode :)
-:- SignOff hcf: #TUNES (Leaving)
<HickServ> "Email(notice to spammers: I will get you...don't screw with me)"
<HickServ> hahahaha
<HickServ> I like your attitude
<tcn> yeah
<tcn> speaking of spam, you got any lately? methinks someone already got 'em..
<HickServ> no
<HickServ> mmmmm forth
<HickServ> at first i over looked forth
<HickServ> it's actually quite a nice language
<tcn> nice little language, yeah
<tcn> once you get used to it
<HickServ> even though i don't have a compliler 
<HickServ> im writing my own lll like forth though
<tcn> that's good practice
<tcn> i've done it twice already :)
01:30pm
<HickServ> yeah well i wanna get a little loower than forth
<tcn> just get started, you'll see how it works
<HickServ> yeah
<tcn> you could look at Retro's forth, but that ain't the only way to do it
<HickServ> i want a language that will control all the modules that are in asm
<tcn> in Forth, I have to write a lot of wrapper routines to use non-Forth code
<tcn> some use C calling conventions, some use registers, etc..
<tcn> hehe.. I was looking at the Oberon IDE driver.. you know, Oberon looks a lot like FORTRAN..
<HickServ> heh never leanred fortran(like learning cobol)
01:40pm
* Fare/#Tunes is back
<tcn> hey fare
<HickServ> hello good fare 
<HickServ> :)
<Fare> :)
<Fare> tcn: so may I commit that glossary entry about reification?
* Fare/#Tunes is trying to get vietnamese locale to work under Linux... locales are a BITCH
<tcn> Fare: yeah, that clears it up
<HickServ> heh The International Brotherhood of Assembly Programmers
<HickServ> YES!
<tcn> how many 'letters' are in the vietnamese script? will it fit in an 8-bit character set?
<HickServ> vscii :)
<tcn> i think it has a lot of accent marks
<tcn> kinda like arabic or hindi
01:50pm
<HickServ> you'd have the do 16 bit 
<tcn> Fare: do you need to be in graphics to display vietnamese?
<HickServ> like unicode
<HickServ> i'd think so tcn
<tcn> i've used sanskrit/hindi fonts before
<tcn> metafonts, actually
<tcn> it used escape codes
<Fare> tcn: I have some viscii 1.1 fonts for console
<Fare> problem is with the keyboard.
<tcn> hmm.. you would type the hindi in (transliterated to latin) but it would display in devanagari script
<tcn> hey, mate birkas joined the tunes ml
<Fare> the keyboard doesn't look like it works as it should
<HickServ> did'nt he just like introduce himself?
<tcn> i was talking to him last month about his reflective assembler
<tcn> interesting idea
<tcn> it's the same philosophy as forth.. simple but powerful
02:00pm
<tcn> damn, the news.. "and as usual, kosovo"
<tcn> when are they gonna run out of bombs?
<Fare> whose reflective assembler?
<tcn> mate birkas
<Fare> there's a country there that has ethnically purified its land, and lives in racial segregation, yet the US is its ally.
<Fare> it's Israel.
<tcn> hehe
<tcn> you could say the same for Turkey
<Fare> or for China
<HickServ> i won't say anything bad about idreal
<HickServ> isreal
<Fare> well, Turkey has not exactly racial segregation as Israel has; however it does prevent speech of minority tongues.
<Fare> HS: why not?
<AlonzoTG> om =\
<HickServ> i like isreal  becasue they know how to get things done
<Fare> HS: fear of being shot by the Mossad?
<HickServ> nah
<tcn> he's just a kid :)
<Fare> HS: "I like Stalin, because he knew how to get things done"
<HickServ> hehehe
<tcn> hey, India took a cue from us.. air strikes on pakistan :)
<HickServ> well where else would jews live and not be persicuted
<AlonzoTG> suckey
<tcn> hs: here
* AlonzoTG/#tunes ph33rs da B0MB
<HickServ> those two countrys are gonna get each other in big touble
<tcn> hs: long island :)
<HickServ> trouble
<tcn> nah, they won't use the nukes
<tcn> (hey what the hell was that loud bang i just heard?)
<HickServ> doesn't matter
<HickServ> look at the US's funds
<HickServ> we're about 4 trillion dollars in debt
<HickServ> that's what the cold war did to us
<HickServ> and look at russia
<tcn> heh
<HickServ> even worse
<tcn> here it's just a number
<tcn> well, we won that game
<HickServ> now think about india and pakistan
<HickServ> they're gonna spend all their own cold war 
<HickServ> then they'll want us to bail em out
<tcn> they been at war for.. what, 40 or 50 years?
<HickServ> it's just now  that it's esculating though
<tcn> nah
<tcn> it's been plenty worse before
<HickServ> of course there have been the usual border skirmishes
<HickServ> but now "D4 B0MB" is in the game
02:10pm
<tcn> d4?
<HickServ> da
<tcn> oh
<HickServ> hax0r slang :)
<tcn> i dunno
<tcn> didn't they sign the test ban treaty?
<HickServ> yet they test em
<HickServ> the ussr did the same 
<tcn> so do we.. but we didn't sign it
<tcn> last I remember
<tcn> I thought india & pakistan tested a few bombs each, then signed the treaty
<HickServ> ah hell i don't know
<HickServ> i like talking about tunes beter
<tcn> yeah
<tcn> ok, i'm writing an fdisk program
<Fare> hum
<tcn> Retro only needs 1 partition per disk
<tcn> or it can use a file in another OS's partition
<Fare> where else would the palestinian live and not be persecuted?
<HickServ> dat be one big file
<Fare> why 1 part per disk?????
<tcn> persistent storage.. all storage is treated as one
<Fare> tcn: yeah, but why *need* a partition on each disk
<Fare> sometimes you just plug a friend's/backup/alien/foo disks
<tcn> on each disk you want to use for Retro's storage space
<Fare> ah, that's not the same!
<tcn> you know what I mean :)
02:20pm
<tcn> wow.. partition tables are redundant :)
<tcn> ok, back to work..
-:- SignOff tcn: #TUNES (tcn has no reason)
02:30pm
<HickServ> oh my
<HickServ> what times is it in france
<HickServ> there is a 9 hour difference from you and i i think
<Fare> it's 23:30 here
<HickServ> not too late
<HickServ> hehhehe
<HickServ> only 11:30
<HickServ> yep that's 9 hours
02:50pm
-:- SignOff iStormy: #TUNES (Disconnecting)
-:- core [core@core.suntech.fr] has joined #tunes
* AlonzoTG/#tunes greets "Salutations core!" :-)
<core> people hi
<core> hi alonzo's auto greet
<AlonzoTG> om
<AlonzoTG> =\
<core> :)
<AlonzoTG> hey, I press that button every thyme someone comes in...
<core> oh, thought it was an auto greet scripted thing :)
<core> but hi anyway :)
<AlonzoTG> topick
<AlonzoTG> ?
<core> what with it?
03:20pm
<AlonzoTG> pleeze rite a paper for the OS prototyping symposium. =)
<core> i have written several 'papers' and I have a working OS prototype ;)
<core> (and maneesh if you read logs, yes it's preemptively multithreaded/protected, no you're not stuck with a model of anything, all can be replaced, and no i'm not pushing it as the tunes lowest layers :P)
<AlonzoTG> oh
<AlonzoTG> how does it werk?
<core> it works well ;)
<AlonzoTG> what methods does it use?
<AlonzoTG> how are progies loaded?
<core> to solve what problem?
<core> they are loaded when they resolve the dependencies of another one
<AlonzoTG> how?
<AlonzoTG> how?
<AlonzoTG> how?
<AlonzoTG> how?
<AlonzoTG> how?
<AlonzoTG> how?
<AlonzoTG> how?
<core> what do you mean how
<AlonzoTG> H O W ? ? ? ? ? ? ? 
<core> they're pulled from wherever storage they are at, partially or fully, and their exports bind to the imports of whoever requested them.
<core> how what?
<AlonzoTG> and their exports bind to the imports of whoever requested them.
<AlonzoTG> how?
<core> both describe what they export and what they import using a language that is compiled and linked with the code, data, etc. of the component.
<core> later at runtime this information is extracted from the component, put back into a tree, and used to satisfy requests.
<AlonzoTG>  that is compiled and linked with the code   < HOW HOW HOW HOW HOW ????????
<core> we have written a compiler, that's how.
<core> it parses the interface definition language and outputs it in a binary form
<AlonzoTG> =\
<AlonzoTG> *-sob-*
<core> then this information is linked by ld to the executable
<AlonzoTG> *-cry-*
<core> i made ld recognize the executable format that allows for that, i told you that already
<AlonzoTG> LD?!?!?!??!?!?!?!?!?!
<core> (well, ld and everything using binutils)
<core> yes
<AlonzoTG> =(((((((((((((
<core> it's just part of the XCOM format
<AlonzoTG> oh that....
<AlonzoTG> =\
<core> what's so sad about it?
<AlonzoTG> I have been stuck at the " I need to develop a linker" stage for aout 8 months!!!
<core> well that's indeed a necessary step before writing any runtime.
<core> that's why i wrote XCOM and associated tools/patches
03:30pm
<AlonzoTG>  i wrote XCOM and associated   <<< HOW HOW HOW HOW HOW HOW HOW???????????????????????????????????????????????????????????????????????????????????????????????????????
<core> how what?
<core> more specifically?
<AlonzoTG> =\
<core> what do you want to know?
<AlonzoTG> E V E R Y T H I N G!!!!
<core> okay, that's your problem.
<core> you can't program "everything", you need to break it down until it is small enough pieces to be understandable
<AlonzoTG> BUT WHAT AM I TRYING TO UNDERSTAND??????
<core> the big problem of "how to link code and meta-information such as an interface, into an executable, that I can use at runtime in my operating system". right?
<AlonzoTG> yep
<AlonzoTG> I'm not going to "just call int 21h"... I need this stuff linked!!!
<core> yes
<core> that's the "big problem"
<core> you need to break it up
<AlonzoTG> =>
<AlonzoTG> how
<core> well let's see
<core> if you want two objects to talk together, you need an interface
<core> right?
<AlonzoTG> probably...
<AlonzoTG> not really sure how or wut....
<core> well, an interface if by definition a way of having two different things talk
<core> the two objects (programs, whatever) don't know anything about eachother
<AlonzoTG> how about I steall all your XCOM stuff?
<core> except they can rely on an interface
<core> sure
<AlonzoTG> hehe
<core> so to link two objects at runtime
<core> you need
<core> 1) a way to describe your interface
<AlonzoTG> =\ agreed
<core> 2) a way to transform that human-readable version into a form that the machine can process
<core> 3) a way to store it
<core> 4) a way to retrieve it at runtime
<core> 5) code at runtime that can act as a matchmaker
<core> xcom will give you 3) and 4)
<AlonzoTG> 5 is the tough part...
<core> yes, it is indeed.
<AlonzoTG> I have a book on database theory...
<AlonzoTG> that would probably do it...
<core> yeah, you can picture it as a database
<core> xcom doesn't limit how you design your interface system, it just stores and retrieves it like a good dog.
03:40pm
<core> i don't want to sound like i know all though, my interface system can use a lot of improvement, but at least it works and i don't know of anything else like that, applied at system level.
<AlonzoTG> k
<AlonzoTG> =\
<core> do you have a better idea of how to approach it now?
<AlonzoTG> is there OMF > XCOM linker?
<AlonzoTG> slitely...
<core> my brain isn't all that great, so i have to break things up a lot to understand them, but then it usually works for me.
<core> omf? like the dos object format?
* AlonzoTG/#tunes has nothing but DOS....
-:- SignOff AlonzoTG: #TUNES (REBOOT!!!!)
<core> if binutils can grok OMF, then yes, no problem.
-:- Zenguy [Zenguy@p16t1.std.dialup.ncf.carleton.ca] has joined #Tunes
<core> hi zenguy
<Zenguy> hi, w'sup?
<core> hi, not much, saturday night :)
-:- AlonzoTG [Alonzo@209-122-199-75.s329.tnt4.lnh.md.dialup.rcn.com] has joined #tunes
<core> alonzo: if binutils can grok OMF, then yes, no problem.
<AlonzoTG> k
<AlonzoTG> =]
<AlonzoTG> or I could use GCC...
<Zenguy> you guys trading MP3's or what :)
* AlonzoTG/#tunes greets "Salutations Zenguy!" :-)
<core> zenguy: not right now ;)
<AlonzoTG> we are writing OS
<core> zenguy: i usually compress mine from my own CDs out of laziness to swap cds later :)
<Zenguy> no kidding - OK I'll bite - what's the project?
<core> zenguy: tunes is a loose group of people with projects revolving around languages or computing systems in general
<Zenguy> K, examples?
<core> zenguy: kernelless, component based operating systems; new, reflective languages; language and OS reviews; academic papers..
<core> zenguy: it goes from the fun low-level hack to the high-level theories.. hence the 'loose' group of people :)
03:50pm
<Zenguy> woah man, you guys are definitely over my little hacker head :):):)
<Zenguy> well, looks like yer friendly anyhow :)
<core> zenguy: maybe not.. i, for one, do not grok all the math theory some people use :)
<core> zenguy: yeah, just view it as a bunch of people that talk about computing systems )
<core> :)
<Zenguy> OK, works for me - anybody here use FreeBSD?
<core> zenguy: the hope is that out of those many loosely related projects, we'll be able to combine one, two or more to make a new computing system that improves a lot of areas :)
<Zenguy> OK, what about something like HURD?
<core> zenguy: no, i've been wanting to try it for a long while though
<Zenguy> never tried it myself, but I hear it's not ready for prime-time yet...
<core> zenguy: i won't comment before i give hurd a try, but from reading their [outdated] documents it looks like Yet Another Microkernel, and usually microkernels are more complicated and less stable than monolithic kernels, without any extra configurability
<core> zenguy: (talking about hurd i assume, not freebsd?) there is a debian distribution at least for developers who want to try and improve it
<AlonzoTG> =(((
<core> alonzo: i'm talking about microkernels, not modular kernels. microkernels as in kernels that run most services in independent address spaces and processes.
<Zenguy> I wanna try Debian & FreeBSD (least that's my NextStep :)  )
<core> zenguy: what are you using right now?
<core> i use debian (gnu/linux, not hurd), and am very happy with it
<core> alonzo: to answer your question, you could use gcc for xcom, yes
* Zenguy/#Tunes gives an embarrassed look :)
<core> zenguy: windows? why are you embarrassed
<Zenguy> RH6
<core> oh ;)
<Zenguy> hah!
<core> well.. that's the training pants for linux :)
<Zenguy> exactly
<core> well it's still linux
<Zenguy> which is why Debian is the NextStep :)
<core> it gets in the way when you want to do anything useful with your system besides upgrading packages by hand, but otherwise it's linux ;)
<core> yeah, i ran redhat till version 4.2 and then upgraded to debian whenever i was setting up new boxes
<Zenguy> sure, not a problem -- it's just that philosophically I'm headed for Debian, y'know?
<core> yeah, i understand. i'd rather use a cooperatively developed system too.
<Zenguy> politics is important to me, tho not to everyone
<Zenguy> brb
<core> and it irks me to support a company that makes so much money off people's backs.
<core> well, i'm french, we are about politics all the time here, it's a national occupation.
<Zenguy> back
<core> :)
<Zenguy> yeah, I hear you.  I'm into Free Software, and always have been.  So, RH serves my purpose as a newbie, but I'll be switching to Debian soon enuf :)
<core> debian is harder to setup but so much easier to maintain
<core> i know you can always play with your RH install to get the same level of maintainability, but i've got other things to do :)
04:00pm
<Zenguy> this is exactly what I keep hearing over and over again -- now my question is -- why?  Is there anything that DEB can do that RPM can't?
* Zenguy/#Tunes throws down the gauntlet :)
<core> well, technically it's two package managers
<core> can redhat scan for newer packages on the net and auto upgrade?
<core> (asking sincerely, i haven't used it in so long)
<Zenguy> hmm, I'm still exploring this, but there's something called AutoRPM which may have this kind of functionality; also GNOME which I'm using presently has a utility called GNOrpm that I still have to try...
<core> yeah.. well debian integrates that as part of the core system, not as an add-on
<Zenguy> excellent :)
<core> packages are also built with a lot of care about what they depend on
<core> and they don't install all over the place
<Zenguy> How about FreeBSD?  Anybody here tried that?
<core> they are sorted sensibly in categories, not just thrown into one big directory
<Zenguy> right I  hear ya
<core> well, it's a completely different beast and kernel
<core> it's a server operating system
<core> basically it means that applications for the desktop are even more scarce than on linux
<Zenguy> the advocates of FreeBSD really swear by it it seems
<core> but i heard that it is an awesome server platform
<core> outperforming linux by a factor of 3 or 4 on some operations
<core> it also has a very nice package system, i heard.
<Zenguy> yeah that's my impression too - FreeBSD = server side; Linux = desktop (?)
<core> also all sources of the system are part of a cvs repository where you can see its entire history of development. not just the kernel, but also libc and all system utilities
<core> yeah, that's what people seem to say
<core> i am very happy with linux on the server side though
<Zenguy> i still think that you could run a Debian server OK
<Zenguy> right :)
<core> i serve about 900 megabytes daily with linux boxes
<core> yeah, debian on a server is pretty good.. usually don't spend more than 15 minutes daily to administer 5 debian-based servers :)
<Zenguy> thanks for the chat - it's supper time for moi :)
<core> ahh.. where are you?
<core> and no problem, you're welcome here anytime
<Zenguy> Canada, eh?
<core> ahh.. where in canada? (go to supper if you need to :)
<Zenguy> (my daughter just lost her first tooth)
<core> ahhh.. cute. i don't have kids yet, but i'm planning on ~ 2 years :)
<core> alonzo, still around?
<Zenguy> somewhere in the Great White North :)
<Zenguy> See ya folks :)
04:10pm
-:- Zenguy [Zenguy@p16t1.std.dialup.ncf.carleton.ca] has left #Tunes []
<AlonzoTG> em
<core> yeah?
* Fare/#Tunes is back
<Fare> core!!!!
<core> salut fare!
<core> :)
<core> not in bed yet?
<Fare> alors, manu, kwa29?
<core> toujours la moitie de 18 :)
<Fare> sur arte y a le concours Reine Elizabeth
<core> c'est lequel deja? je confonds toujours avec le concours de violon de la reine de belgique
<Fare> les deux: violon *et* piano
<Fare> c'est celui la-meme
<core> ahhhhh
<core> oui, d'habitude les finalistes sont carrement doues :)
<core> tu joues un des deux?
<Fare> malheureusement non
<Fare> mais j'ecoute les deux :)
<core> hehe.. c'est deja pas mal
<core> je me debrouille au piano, mais c'est tout
-:- FareWell [rideau@quatramaran.ens.fr] has joined #Tunes
<core> pas du niveau des finalistes en tout cas
<core> mais j'ai des doigts a la E.T. donc je triche
<Fare> si si, j'ecoute actuellement du niveau des finalistes
<Fare> s/du/quelqu'un du/
<core> heheh
<core> t'en es ou de ton compilo? :)
04:20pm
<Fare> 0
<Fare> je suis en train de faire du vn
<core> vn?
<core> virtually nothing?
<core> je viens de comprendre pourquoi nickserv aime pas mon nick
<Fare> vietnamien
<Fare> pourquoi?
<Fare> mon frere utilisait tamere
<HickServ> back
<core> ca a pas l'air distribue, si je vais sur le serveur ou je l'ai enregistre, il accepte mon passe, pas sur celui la. space!
<HickServ> deseo hablar frances
<HickServ> hehehe
<HickServ> yeah right
<core> hickserv: too bad :)
<core> hickserv: not discussing anything important... just what was on french tv
<Fare> deseo?
<HickServ> i see
<HickServ> it's spanish
<HickServ> deseo means "i wish"
<Fare> si, pero no conozco este palabra
<HickServ> :*
<core> donde esta la plaza de catalogna?
<Fare> core: que era la expresion que me aprendi? "chupar la ..." ?
<HickServ> suck the
<core> fare: chupa me la poya (well, not you)
<core> :)
<HickServ> heh
<HickServ> ah
<core> mi espanol esta oxidado
<HickServ> uh you only use mi after a prepostion ;\
04:30pm
<core> see, my point exactly :)
<HickServ> para mi para ti
<core> true
<core> well, i only know some spanish from school :)
<HickServ> i still need to learn a lot more about conjugation of verbs
<HickServ> how many years?
<core> my german is better and i want to hope my english is much better ;)
<core> only 2 years of it, and it's been 8 years :)
<HickServ> im learning german
<core> all the street spanish slang i know was learnt from a spanish girl tho ;)
<core> (that, and ralf koenig)
<HickServ> yeah i've only taken one smester
<HickServ> im taking like 2 years when i get to high school
<HickServ> mabye more
<HickServ> jajaja
<core> you're not in high school yet..?
<HickServ> no
<core> oh.. okay
<HickServ> yes
<HickServ> yes indeed
<core> just curious :)
<HickServ> you are i sly one
<HickServ> i'll keep my eye on you
<HickServ> ;D
<HickServ> grrrr ms word is so damn annoying
<core> what did I do now? :)
<HickServ> nothing
<Fare> HS: what? You use that evil proprietary software?
<core> yeah, well, that's m$ for you
<HickServ> yeah i do
<HickServ> i have no choice
<HickServ> this isn't my computer
<core> hickserv: sure you do. you can reverse engineer their stupid format and write word documents with 'echo' then ;)
<HickServ> =\
<HickServ> mabye i'll just finish whatever im doing and not bother cracking an m$ program
<core> that too :)
<core> (i was just kidding, btw)
<HickServ> np
<HickServ> so core what is your involvement in tunes?
<core> low-level i guess
<HickServ> ah
<core> i have a functional kernel-less, component based OS
<core> and i designed the XCOM format on which it is based, on which retro also will be
<HickServ> you are just the person i would like to talk to
<core> and that can hopefully be tailored to meet tunes' needs
<Fare> too many things to write for monday!!!
<HickServ> i am really vague on the kernel-less idea
<core> (whenever Fare submits this promised email about it to tunes@tunes :)
04:40pm
<Fare> Rachmaninoff's third!
<HickServ> could you like um
<core> well, it's just about not having a single block of code with like a few bottleneckish entry points anymore
<HickServ> explain it to me
<core> you still have code running in a privilegied processor mode
* Fare/#Tunes wonders how finalists are discriminated from each other...
* Fare/#Tunes is sure not able to be member of the jury
<core> fare: female, and they know how to play wind instruments
<HickServ> so are messages passed?
<Fare> yuck
<Fare> why would "messages" need to be passed????
<Fare> (unless you're talking about RPC, of course)
<HickServ> so what would you do 
<core> hickserv: only when calling code in other processes or network nodes
<core> hickserv: let the OS decide the best method
<HickServ> does your threads decide whether to preempt or be cooperative?
<core> hickserv: clementine directly calls the function you want if it's in-process and at the same processor privilege level; else, it inserts code to call through a gate if calling authorized privilegied code; else, it inserts code to wrap your parameters and the result in messages in the situation above
<core> hickserv: mine are preempted
<core> hickserv: again, making a no-kernel system doesn't mean you have no privilegied code
<HickServ> fare likes the idea but i don't see any safe way to make taskswitching cooperative
<core> hickserv: make the compiler take care of it
<core> hickserv: indeed, it's not possible to safely do that in C for instance.
<core> hickserv: most of Fare's ideas are to be implemented in conjunction with a system that uses a safe language and safe code.
<HickServ> oh
<core> hickserv: i think i did a decent tradeoff. clementine is written in C and still has no kernel
<core> hickserv: when fare comes up with his compiler for his dream language, then probably more of his ideas can be implemented in it
<Fare> :(
<core> fare?
<abi> fare is sometimes looking at another screen and interruptible with a beep (/msg Fare ^G) or at http://www.tunes.org/~fare/
<Fare> monday, I have a meeting with that guy from Marseille who implements another reflective OS for his thesis...
<core> fare: oh, yeah, him.. what's sad about it?
<Fare> and unlike me, he's rather a coder than a theorist...
<Fare> sad? why sad?
<core> you were frowning above
<Fare> Henri.Lesourd@lip6.fr
<core> ah yeah, lip6 are usually clued people
<core> well, unlike you, he is alone on the task? :)
<Fare> frowning? Not against him, if I were.
<core> oh, what about then?
<Fare> I've contacted him only recently, and we look like we are tuned
04:50pm
<core> that's cool :) maybe we'll see a tunesish implementation from which we can derive good ideas then?
<core> did guillaume email you back?
<Fare> guillaume? not yet.
<core> oh ok.. he's away for the weekend, but i was curious :)
<core> he will.. he's interested in your ideas :)
-:- FareWell [rideau@quatramaran.ens.fr] has left #Tunes []
* Fare/#Tunes is still here, tho
<core> he also has an academic background like you
<HickServ> that leaves FareBadly
<core> hehe
<core> i'll build a statue to kevin p. lawton someday.
<Fare> well, in the past he implemented various systems to do AI
<HickServ> :)))
<Fare> including a LISP system, and an "operating system" sitting on top of DOS
* HickServ/#tunes loves AI
<Fare> he also hacked the linux kernel to intercept syscalls and do fun stuff
<core> cool! do you know _how_ he implemented his reflective OS?
<core> well, intercepting syscalls is easy :)
<Fare> he sent me source for his DOS thing, and I read a paper about his linux hack.
<core> my mind isn't nearly as good as yours so the math stuff escapes me quite.. but the low level bits are easy :)
<Fare> core: sure it is.
<core> well, it's not trained as much then ;)
<Fare> uh, I mean, sure intercepting syscalls is easy
<core> oh.. hehe
<core> yeah, it is
<HickServ> i love hacking the linux kernel
<HickServ> but i only do little stuff
<Fare> I hate it -- no metaprogramming at all
<core> i should probably be hacking the linux kernel some more, but i have my name in it so it's good enough and i moved to something else ;)
<HickServ> really core?
<Fare> the linux kernel *begs* for metaprogramming, but none of its developer realizes it
<HickServ> which file?
<core> i couldn't figure out how to really improve something as interdependent and complex to maintain 
<core> hickserv: /usr/src/linux/drivers/video/fbcon.c
<core> hickserv: i did more work than just that, but that's where i'm credited and it's good enough.
* Fare/#Tunes has his name, too, but isn't too proud of it :(
<core> fare: why, the asm-howto? it's useful.. you just miss a proper explanation of the machine model and register aliases (which is in the gcc info page anyway)
05:00pm
<core> hmm hmm hmm hmm
<AlonzoTG> fare: what do u want done to liNuKs?
<Fare> ATG: transform it into Tunes :)
<Fare> machine model?
<core> fare: yes, the format the target model in gcc uses for describing parameters i mean.. like m, r, g.. for memory/register/generic.. and then A, C, etc. for registers
<Fare> oh. I think the HOWTO lacks a precise pointer to these ones...
<core> fare: it's well explained in the gcc info pages
<Fare> the Asm-HOWTO is really mismaintained
<core> fare: it kind of lacks in the HOWTO though.. maybe you could give a short rundown of them :)
<Fare> as are all my HOWTOs
<core> hehe.. well at least it exists and it is useful
<core> it was useful to me when i coded the assembly parts of ggi at least
<core> your HOWTO and the gcc info pages were enough.
<core> hey Al, figured out your interfacing problem yet? :)
<core> hey hick, still around?
<Fare> xfsft not in debian :(
<Fare> thanks god, there's alien
<core> sure it is
<core> xfstt
<core> Package: xfstt
<core> Status: install ok installed
<core> Priority: extra
<core> Section: x11
<core> Description: TrueType Font Server for X11
<core> that one? :)
<core> alien is evil
<core> converting from RH packages is dangerous, the dependencies will be only mostly right. i ran into lots of problems using alienified packages.
<Fare> no, xfsft is better
05:10pm
<Fare> are is it not?
<core> what's xfsft?
<abi> xfsft is better
<core> thanks abi
<Fare> abi: no, you're very stupid
<Fare> abi: what are you?
<abi> i am the little girl borg with a big brain or #tunes' personal little whore or an infoslut or gaklosmontic or flurivostuginuous or xyvarestoplik or an Artificial Idiot or braindead
<core> the only thing that could be better than xfstt would be implementing multicolor pixmaps for fonts in X so they can be antialiased.
<AlonzoTG> om
<AlonzoTG> =\
<Fare> mu
<core> peta
<core> fare: what's xfsft, and how is it better? or was it just a typo on your part :)
<Fare> well, a friend of mine works on it, so it must be good :)
<core> what is it?? ??? :)
<Fare> a rival to xfstt
<core> oh
<Fare> that uses the freetype library
<core> so does xfstt
<Fare> and is integrated to the standard xfs
<core> well, xfstt is a seperate server, but what's the benefit of integrating it into xfs?
<core> i like to be able to load truetype support only when needed
<Fare> it can serve any type of fonts: truetype, bitmap, type1, etc
<core> and i like that i only need 1 truetype server for all my X11 boxes
<AlonzoTG> om
<AlonzoTG> =\
<core> al?
<abi> al is silly
<AlonzoTG> no, abi, I'm lame
<AlonzoTG> too lame to link
<AlonzoTG> too lame to load
<core> i'm?
<AlonzoTG> =\
<core> i?
<abi> somebody said core was developing a kernel-less, component-based OS, Clementine, and the XCOM file format (http://xcom.tunes.org)
<core> heh.
<AlonzoTG> i?
<abi> you are lame
<core> heh
<AlonzoTG> =P
<Fare> I?
<abi> you are sometimes looking at another screen and interruptible with a beep (/msg Fare ^G) or at http://www.tunes.org/~fare/
<core> i didn't know that feature of abi
<core> scary to think it's a b33y0tchX script
05:20pm
<AlonzoTG> I wish a hydraulic konversion kit wasn't soo ekspensive...
<AlonzoTG> =\
<core> what do you need that for?
<Fare> finale of Rach*'s third!
<Fare> the 12 finalists have played!
<core> yeah, i have that as background noise :)
<Fare> oh, so I needn't comment, then
<Fare> my these VietLUG people are SOOOOO lame!
<core> well, you can always comment :)
<core> i don't have your classical music knowledge
<core> i am a junkie of caffeine and electronic music, i need them both to code.
<Fare> the current musical instrument is older than classical music: it's hand clapping
<core> heheheh
<Fare> prix mille neuf-cent nonante neuf
<core> the queen elizabeth contest was a very big thing when i lived in belgium
<core> heheh.. yeah. reminds me of good times to hear numbers spelled this way
<Fare> it still is
<core> yeah i know but i mean, it was a complete media black-out, they only talked about that :)
<Fare> the hardest and most prestigious classical music competition in the whole world
<Fare> so you lived in .be? that explains a lot...
<core> yeah, very respected.. and indeed people who get it, deserve it
<core> hehe.. like what?
<Fare> rien, une fois.
<core> my lack of south-french accent, my beer belly, or my stupidity? :)
<core> i am french and born in montpellier though. just travelled a lot on my own
<Fare> hehe... I have family in .be...
05:30pm
<core> I evolved from a jerk to a socialising jerk by travelling ;)
<core> Do you? cool! where? wallonie, flanders?
<core> people in belgium are 350 billion years ahead of us. well, in flanders anyway.
<core> too bad mafia has such a grip on the entire country
<Fare> some in wallonie, some in flanders
<core> i guess it's a french tradition to despise our own country in favor of others though. i know it amazes all foreigners.
<Fare> the french wouldn't accept vietnamese students in the end...
<core> there aren't that many non-caucasian people in belgium tho.. didn't see many of them at least, that shocked me :)
<core> they got accepted there?
<Fare> yeah. and the latter as refugees
<Fare> France, terre d'accueil? Mon oeil!
<core> that's good.. i guess people who grow up with two languages are more open
<core> the secretary at suntech is from flanders, she speaks 7 languages including french with no noticeable accent.
<core> and she is 25
<core> and i guess since you can master a language as hard as vietnamese, combinatory logic and things like that must be a piece of cake to you :)
<Fare> y en a un qui boude ... il s'attendait a mieux que 6eme
<core> ils ont du bosser comme des furieux, ca peut se comprendre
<core> moi je les mettrais tous 1er :)
<Fare> flemish people tend to learn lots of languages -- at least 3.5 (flemish/dutch, french, english)
05:40pm
<Fare> core: je croyais que tu mettais les filles devant?
* Fare/#Tunes *cannot* master vietnamese :( :( :( :(
<core> what do you know, an asiatic wins :)
<Fare> they *all* win!
<core> yeah, they do. they impress me so much with that
-:- _ruiner_ [nate@ppp112.wi.centuryinter.net] has joined #tunes
<core> my boss in belgium spoke 5 fluently and understood 3 more
<core> he was flemish too
<_ruiner_> he was also a big dork
<core> the marketing director was from wallonie and barely spoke french :)
<Fare> yeah. Wallons don't learn much
<core> hehe.. my point exactly
<core> ruiner: lol
<core> ruiner: no, he's the best boss i ever had.
<Fare> why didn't you stay? climate?
<_ruiner_> dorks tend to be better bosses than assholes
<core> that's one yeah.. +8 C in may is harsh
<Fare> or because "best" isn't "good"?
<core> two my entire family and all friends are here in montpellier.. after 18 months it starts to make one sad
<AlonzoTG> om
<core> and three i wanted some change.. doing the same thing for a year and half is annoying :)
<_ruiner_> no kidding
<_ruiner_> I spent six months just taking off cases, putting in an ethernet card, and putting the case back on
<_ruiner_> sucked
<core> ruiner: that would suck bad
<core> ruiner: i was running a network, so it's better, but i was missing doing programming
<_ruiner_> well....I also got to install NT and some apps on every single one of them too
<Fare> je declare le concours mille neuf-cent nonante neuf cloture.
<core> hehe, yeah
<core> i love that accent and spelling
<core> and there he goes speaking dutch fluently.. grr! :)
<_ruiner_> masses moi les couilles
<_ruiner_> check out the french.....oh yeah!
* core/#tunes kicks ruiner in the gonads
<_ruiner_> french people suck
<_ruiner_> did I spell it right core?
<core> ruiner: at least the female part of the population does
<core> well, and some of the male population too, but i won't get into that
<_ruiner_> ack!
<core> ruiner: it was perfect :)
<_ruiner_> ok......whew!
<core> heheh
05:50pm
<_ruiner_> why is this a guy only chat? just out of curiosity....
-:- _1BadDude [Alonzo@216-164-230-109.s109.tnt7.lnh.md.dialup.rcn.com] has joined #tunes
<_ruiner_> wb zo
<core> ruiner: because not many girls seem to be interested in writing OS code?
<core> ruiner: i don't know
-:- SignOff AlonzoTG: #TUNES (Read error to AlonzoTG[209-122-199-75.s329.tnt4.lnh.md.dialup.rcn.com]: Connection reset by peer)
<core> ruiner: there are many talented girls out there.
<_ruiner_> but not enough to make this a gender balanced chat room
<_ruiner_> I guess men are just superior
<_ruiner_> yay me!
<core> nah, men suck
<_ruiner_> men rule! especially me
<core> my girlfriend learnt HTML during a lunchbreak
<Fare> men suck? I thought it were women.
<core> and she isn't really programming inclined
<_ruiner_> I learned html in 10 minutes from a book
<_ruiner_> easy stuff
<core> well, can i still brag about her? :)
<Fare> so you needed a book to learn html???
<core> fare: some do, i heard
<Fare> you suck!
<core> she learnt it off me
<core> people had shown her WYSIWIG editors but she didn't like it
<core> she likes to play with stuff ;)
<_ruiner_> give me a break, I learned it 4 years ago
<core> i'm just teasing :) i just think she is immensely more intelligent than i am.
06:00pm
<_ruiner_> could very well be, but you yourself said she wasn't programming inclined
<Fare> don't tell her. She might get ideas...
<_ruiner_> which just seems to add another small bit of credibility to the idea that technology tend to be the domain of men
<core> not really.. she does some bash scripting because she does use unix (hp/ux and linux) and doesn't want to do the same thing 50 times, but that's it (and no, she isn't a fat truck driver either :-)
<core> ruiner: programming as stupid as it is today is, yeah
<_ruiner_> ouch, harsh
<core> ruiner: if we can achieve a system such as the tunes pipe dream, maybe that will change
<_ruiner_> tunes is unachievable
<core> ruiner: i don't know, women don't seem to want to be satisfied with explanations such as "it's for legacy code support".
<_ruiner_> I'd love to see you guys prove me wrong, but I know you never will
<core> ruiner: don't look at me, i'm a low-level person.
<core> ruiner: it is achievable to some extent.
<_ruiner_> prove it
<_ruiner_> do it
<_ruiner_> make it
<core> ruiner: i have a working OS prototype that is quite advanced now, and implements the tunes idea of no-kernel.
<core> it's not perfect, but it works
<_ruiner_> does it live up to the ideas of the tunes project?
<core> i'm not an official member of the tunes project
<core> just part of the loose group of people gravitating around it
<core> well, it implements a fully component-based, kernelless system, that allows for migration and persistance
<_ruiner_> uhhuh
<core> it allows for runtime reconfiguration and specialisation
<core> if fare ever writes his language, it would get even closer.. most of his ideas can only live within his language
<core> for now i think it's a decent tradeoff between current computer reality and his pipe dream
<core> besides some guy in marseille has implemented a reflexive OS for his thesis using apparently similar ideas
<core> and a friend of mine independently came to the same conclusions as fare and brian rice concerning the implementation
<Fare> ruiner: do you have anything to propose? or any conclusive argument for why it cannot be done? or for what should be done instead?
06:10pm
<_ruiner_> nope
<_ruiner_> basically I'm just trying to bait somebody into an argument
<core> ruiner: i'm not an argument type of person :)
<_ruiner_> that or into actually sitting down and really getting to work on it so they can make me eat my words
<core> ruiner: and you can safely entice me to 'make it', i have been writing OS code for ~ 1 year
<core> ruiner: clementine only implements some of fare's ideas, but i can actually see where he's getting at in some ways.
<_ruiner_> I disagree with the whole no kernel idea
<core> how do you understand this idea? what is the no kernel idea for you?
<Fare> 10 GOTO 100
<Fare> ...
<Fare> 100 SLEEP
<core> 110 PRINT "NIGHT FARE!"
* Fare/#Tunes is away (sleep)
<_ruiner_> the same way its defined on the tunes glossary
<core> explain it with your own words, and why you disagree
<core> i'm interested :) maybe you've got a valid point :)
<_ruiner_> I disagree because I think that with all the calls that will have to go back and forth to fully integrate programs with each other will slow down the whole thing
<core> what does it have to do with the idea of no kernel?
<core> the 'no-kernel' idea is simply doing away with a monolithic block of code that has a few bottleneck entry points (syscalls). that's the contrary.
<_ruiner_> it also seems like a pretty archaic way to run things.....a single shoddily written program could wreak havoc
<Fare> just before I goto sleep: just because it's integrated *at the high-level source* doesn't mean the low-level implementation should naively do lots of calls back and forth!
<core> that has nothing to do with no kernel
<core> and there can't be a shoddily written program in fare's idea.
<core> because it uses a safe compiler producing safe, checked code.
<core> not arbitrary c code eh, but a language specifically designed for provable correctness.
<_ruiner_> you're living in a dream world.....
<_ruiner_> you can't check code like that.....
<_ruiner_> well you can....but it won't be 100%
<core> you can't check it with arbitrary C code
<core> that's what i'm saying
<core> you need to have a language specifically designed for that
<Fare> even java (a lame lang) can ensure lack of pointer-related havoc
<core> for instance, java sucks, but it's checkable 100%
<core> exactly fare
<core> why? because there are no inferent pointers
<Fare> LISP did it 40 years ago, btw
<_ruiner_> I bet I could write a program that could get around that
<core> ruiner: if the language enforces typing and semantics, you can't wreck havoc
<Fare> those who do not understand LISP are condemned to reinvent it, poorly.
<core> ruiner: try to write a malicious program that escapes the JVM verifier. i'll give you my respect if you do.
<_ruiner_> it can be done
<core> ruiner: then do it, i'm watching you.
<_ruiner_> especially with the older versions of the vm
06:20pm
<core> ruiner: no, with the latest version of the VM.
<_ruiner_> ack!
<Fare> just because htere's a bug in the vm doesn't mean the design is bad
<_ruiner_> it can be done...
<core> ruiner: maybe there were omissions at some point but that doesn't prove your point, ruiner.
<Fare> in a buggy implementation of Unix (or whatever), there might be problems. So what? The fault is in the bug, not in the design!
<core> ruiner: then do it. if you make a program that violates the JVM security i'll admit i'm wrong.
<core> ruiner: note that even so, clementine is currently written in C and still has no kernel and is still secure.
<_ruiner_> why would you write it in C?
<core> ruiner: because it only implements the part of the idea that makes the system modular, runtime replacable and expandable.
<core> ruiner: as opposed to what?
<_ruiner_> asm
<_ruiner_> I suppose time saving.....?
<core> ruiner: because it is 10 times easier to change things?
<core> ruiner: right
<core> ruiner: asm is fine if you know exactly what you are going to write
<core> ruiner: as i had never seen such a system implemented, that would have been a bad choice
<_ruiner_> hmmmm.....
<core> ruiner: besides, ia-64 is around the corner, and i wanted it to be portable, easy to understand (even more important) so its ideas could be used by others
<_ruiner_> I'm going with assembly language....but I don't have any notions that the first thing I write is going to be the one I go all the way with
<core> ruiner: starting to write an OS in assembly now is silly
<core> ruiner: by the time you're done, the hot platform will be ia64 or something else
<_ruiner_> I'm thinking I'll have 10 different versions of the same thing before I'm happy with one of them
<core> ruiner: if you write an OS by yourself, you'll have something usable with applications around, say, 2001. seriously. i've been working on clementine for a year.
<core> ruiner: and i figure it'll be another year before i have a fully functional system that i can use as my runtime environment
<core> of course it will probably go much faster when i open it to other people.. but still.
<_ruiner_> as soon as I can prove to a few certain people I'm serious and can and will carry through with it, I'll have help
<core> probably
<core> there is no end to the asm vs. C argument
<core> i am a fluent asm programmer as well. some of clementine is assembly for speed and control
<_ruiner_> but yeah, somebody else told me it was silly to program in asm now
<core> and tom is doing fine in assembly with retro
<Fare> read the asm HOWTO about that
<Fare> it makes the point for asm vs C vs other
<_ruiner_> but I want to do it in asm....
<core> since retro is a test project not destined to live on its own as the assembly version, i guess it's ok
<Fare> I want to do it in TTL
<core> ruiner: nobody says you can't. i'm just telling you my opinion
<Fare> (transistor-to-transistor-logic)
<core> ruiner: really i'm an assembly and low level fan. fare can tell you that i was really a lot when we met.
<Fare> I want to do it in subatomic particle structure of the computer hardware!
<core> lol
<core> fare: i'm not throwing all of the code in clementine soon. therefore C is a good choice i think. it allows me to experiment things very quickly.
<Fare> yup
<core> fare: and well written C code compiles to almost identical assembler
<Fare> as long as you're not hit by the limitations of the C calling convention, C is a right choice.
<core> fare: and the more i code, the more i see the usefulness of meta-programming and of a safe, intelligent language.
<core> and the more I wish fare would start writing his compiler :)
<Fare> (and by the limitations of the compiler wrt concurrency invariants)
<core> yeah, there are some tricky bits in C. everyone seems to find them normal.
06:30pm
<core> software can't go on forever with piles of asm hacks
<_ruiner_> why not?
-:- SignOff _1BadDude: #TUNES (Have Nice Day :))
<core> because a computing system is encompassing more and more things
<_ruiner_> I don't like the idea of languages going to higher and higher levels
<core> do you realize how much an OS is expected to do today?
<core> humans aren't scalable
<core> this is only the beginning
<_ruiner_> I don't want it to end up someday where any buttlicker 14 year old kid can write an os in an afternoon
<core> in 50 years, programmers will only build programs from libraries of interoperable components
<core> ruiner: it will be irrelevant by then
<_ruiner_> I mean hell, you can insert a fookin web browser in any vb app....its a friggin control
<core> ruiner: you'd rather write a new buggy browser for every new application?
<core> ruiner: components are the way of the future
<_ruiner_> argh!
<_ruiner_> never!
<core> ruiner: technically in clementine you can make an entire OS with simply selecting components
<core> ruiner: why reinvent the wheel all the time? you can change OSes policies by just replacing components
<_ruiner_> I'm also against the idea of making computers more dependant on networking
<core> ruiner: why not? offline doesn't make sense anymore.
<_ruiner_> I saw something on slashdot a few months ago about dumb terminals for home computing
<_ruiner_> core: have fun with your web tv, I'll stick with my pc
<core> ruiner: and even so, you can't go against economics. and economics say it is more profitable to build an application from components rather than writing everything from scratch again with the thousands of hours of debugging
<core> ruiner: nobody is saying they have to be webtvs
<core> ruiner: i don't think thin clients will work
<core> i don't want one at least.
<core> my dad doesn't want one either, so that spans all population types :)
<core> ruiner: again, you're talking to someone who likes assembler a lot and programs as a job.
<_ruiner_> I'm also big on bbs's......I want to help them make a comeback
<core> bbs'es ruled, so did the amiga, but they are dead now
<core> progress stomps on a few good things
<_ruiner_> there are still some around
<core> ruiner: besides, the IA64 will be very hard to directly code in assembler
<core> ruiner: sure, there are also amigas around
<core> ruiner: but it's on the way out, unfortunately
<_ruiner_> I figure if you can write a hot enough app....give it to the bbs's free of charge....then sell the client software...you could have some awesome quake style gaming
<core> ruiner: you are in the US. in europe, local phone is not free.
<_ruiner_> ack!
<_ruiner_> but.....think about it.....less lag
<core> ruiner: i have 300 mbps, what are you talking about? :)
<_ruiner_> less lag in gaming = good
<core> i don't game much anymore
<_ruiner_> those of us with modems don't have that luxury
<core> i don't like 3D games
<core> they don't make my kind of games anymore
<_ruiner_> games are what makes windows what it is
<core> or i'm not aware that they do at least
<_ruiner_> if it wasn't for games we'd all be using dos still
<_ruiner_> or os/2
<core> no
<core> games worked way better under dos
<_ruiner_> os/2 rocked when it was running a 32 bit app though
<core> programmers just got tired of writing sound and display drivers for 350 boards all the time
<core> and the m$ marketing machine did the rest
<core> you can't blame programmers for not wanting to write code for every stupid pc device on earth
06:40pm
<core> well, winnt and os/2 evolved from the same codebase
<core> knowing that, i'll keep clear from os/2
<_ruiner_> I would try to make that the hardware manufacturer's responsibility...put out the code creative labs!
<core> ain't happening
<_ruiner_> I konw
<_ruiner_> and I know
<core> creative labs hired one of my GGI fellows to make their drivers actually
<core> binary-only releases for linux :/
<_ruiner_> why is creative labs so anal about that?
<core> ruiner: anyway, i love assembler, i wrote a kernel in m68K code a few years ago that got used in lots of amiga demos, and i wrote lots of x86 assembly code for various companies. but you can't help the majority of people evolving into higher level languages/OSes.
<core> ruiner: we are well beyond the point where a single human can program an entire OS in assembler and maintain it for years without hindering growth
<_ruiner_> all I need is 10 years.....
<core> ruiner: dos and windows 3.x are mostly assembler
<core> ruiner: does that tell you anything about locking people out from improvement?
<_ruiner_> I loved dos though
<core> ruiner: besides, the ia64 needs a lot of information an human can hardly provide
<_ruiner_> I'd still be using it if games didn't require win95
<core> ruiner: ie. the ia64 instructions need information like branch probabilities
<core> ruiner: of course, dos is just a shell, it's not an actual OS. is that what you want to design?
<_ruiner_> ok ok....dos was a shell for an 8 bit os
<_ruiner_> blah blah blah
<core> no
<core> i'm not using blind critic here
<core> seriously, it is a filesystem, and a shell
<_ruiner_> no....
<core> all applications implemented everything themselves
<_ruiner_> I liked how it was easy to configure though
<core> of course, there was not much to configure
<_ruiner_> I could change anything by knowing how to edit 2 files.....autoexec.bat and config.sys
<core> only the drivers for the filesystem, a few system flags..
<core> and then all the pile of kludges for memory management
<core> yes, but there wasn't much to change, was it?
<core> i'm not saying it was a bad environment, just not an actual OS.
<_ruiner_> exactly....you could set how many files could be open concurrently
<_ruiner_> true....there wasn't much to configure
<core> yes. it's a filesystem, and a shell.
<_ruiner_> but what there was was really easy
<core> yes, it was
<core> but how many years did it stay without improvement?
<_ruiner_> thats what I liked about it......I didn't have to go buy a new version every 2 years
<_ruiner_> os/2 is only on version 4.0
<_ruiner_> and how many years old is that?
<core> of course, everyone provided kludges for what DOS didn't manage.
<core> os/2 is C
<_ruiner_> os/2 is awesome
<core> if you say so
<core> i need the *nix command line for work, so that limits my choices
<_ruiner_> I like command line interface...but I won't give up a gui
<core> that's not what i said
<core> i have a GUI
<_ruiner_> there's a time I would have
<core> with lots of xterms open
<_ruiner_> I'm not talking about what you use
<_ruiner_> I'm just commenting
<core> okay
<core> well, see, so you wouldn't use dos anymore?
<_ruiner_> I would so long as I could use win 3.x
<core> so dos couldn't evolve into the new computing needs
06:50pm
<core> dos and win* are full of undocumented kluges, illogical api calls, limited values due to register passing (like the infamous BIOS harddrive limitations that led to kludges like LBA)
<_ruiner_> it could be rewritten so that it could though
<core> rewritten is the key word
<core> you will end up rewriting your os many times
<core> i rewrote little of clementine, and it still took me one year to get it from nowhere to a functional system
<core> and i'm working on it about 6 to 10 hours a day.
<_ruiner_> I don't doubt that I'll rewrite it many times
<core> so it'll take you 10 years as you said
<core> in 10 years x86 will have disappared; ia64 is just around the corner
<_ruiner_> I'm not worried
<core> why not?
<_ruiner_> because once its done, it gets easier and easier to do again
<_ruiner_> even if the architecture changes
<_ruiner_> true, code will have to be drastically changed
<_ruiner_> but.....
<_ruiner_> the ideas remain the same
<core> you'll have to rewrite every line of it though
<_ruiner_> not EVERY line
<core> in assembler, sure
<_ruiner_> I bet I could leave at least one line the same
<core> if you move to ia64 i mean
<core> well, besides comments.
<_ruiner_> might not be in the same place.....
<core> all code then.
<core> ia64 instructions are completely different.
<core> it will take years to humans to finetune how to directly program the ia64 in assembler
<_ruiner_> still just ones and zeros when you get down to it
<core> you still have to rewrite the code
<core> already, what you don't seem to realize is that you have to be a pretty good asm programmer to beat a modern C compiler
<core> that's on x86
<core> on merced, the difficulty will increase tenfold
07:00pm
<core> optimising code in assembler has to be done differently for every cpu of the intel family
<core> what will be a big win on 486 will be a big performance penatly on pentium II
<_ruiner_> thats a given
<core> and probably most asm code you can write by yourself for merced will be a big performance penalty as it will take hours to figure out the right branch prediction, predicate etc to give it
<core> well, that means you need to write 4 versions of your operating system if you don't want to be slower than C
<_ruiner_> I was thinking of optimizing for the k7 first actually
<core> well, that would be 5
<core> and 6 if you want to support cyrix cpus
<_ruiner_> I just seem to see how something would be done in asm, that I can't really figure out how to do in c
<core> like what?
<core> there are things that need to be done in assembly for control; but i think the speed issue is irrelevant and wrong
<core> and that most of the OS shouldn't be implemented in it
<core> especially considering that the cpu will spend most of its time in a few routines, that can then be hand tailored, rather than losing time on the entire thing.
<_ruiner_> I'm totally all for C for some parts, but for speed you have to optimize with asm
<core> that's wrong
<core> you will optimize for one cpu
<core> again your k7 optimized code will run dog slow here
<_ruiner_> so you think its a waste of time to optimize a loop for 6 different cpu's?
<core> depends on what loop
<_ruiner_> a chink in the armor
<_ruiner_> heh heh
<core> no
<core> i implemented some bits in assembler
<core> you are talking about writing the whole thing in assembler
<_ruiner_> I was thinking of it
<core> that's what i think is time consuming
<core> you need some assembler in it of course
<core> at least to bootstrap the system
<_ruiner_> or for accessing hardware
<core> and abstract hardware a minimum
<core> if your loop gets free memory pages, i suppose optimizing for every cpu is relevant
<core> if it's an init loop, it makes no sense
<core> and even so, you can probably win bigger by optimising the design
<_ruiner_> good point
<_ruiner_> and thats where I'm at now...design
<_ruiner_> I'm thinking of buddy lists for mem management
<core> yeah, that's what i use for clementine
<core> for physical memory management
<core> just like linux and freebsd, with a few changes to make it go faster
<_ruiner_> I'm thinking of something like that for the file system too
<core> (or maybe freesbee changed their mm recently)
<_ruiner_> it seems like a pretty good idea
<core> hll, that could be interesting
<core> s/hll/hmm/
<core> space wasting though probably
<_ruiner_> probably
<core> depends on your top-level block size
<_ruiner_> actually, I was thinking of something more like a bunch of multidimensional arrays
07:10pm
<core> what do you mean?
<_ruiner_> instead of a couple of linked lists...I'd have an array of partitions, and from there an array based on.....ummmmm.....file types perhaps......and under there based on name.....
<_ruiner_> or size
<_ruiner_> or I dunno
<_ruiner_> still thinking about it
<core> well buddy lists would only be for getting blocks in a filesystem
<core> they are only good at partitioning space
<core> not for sorting files, or did i miss what you meant?
<_ruiner_> seems a lot faster than going through a fat table in a linear fashion
<core> it is
<_ruiner_> I'm talking buddy lists for mem management
<_ruiner_> multidimensional arrays for a filesystem
<core> you can probably not waste too much space actually if your smallest block size is the size of a disk block
<core> well, you could apply buddy lists for block allocation in an fs as well
<core> and yeah buddy lists are fast
<_ruiner_> it would keep track of the location and size at the least significant array level
-:- _QZroot [root@p0wer.qzx.com] has joined #tunes
<core> i can get free pages with a few shifts
<_QZroot> this is KEWL!!!!!!!!!!!!!
<core> _qz: what is? ircing as root so all script kiddies can exploit you?
<_QZroot> redhat is the best
<core> that's debatable
<_ruiner_> I'm in his system right now
<_ruiner_> whats all this gay porn stuff?
<_QZroot> hahah xchat said that ircing as root was stupid :)
<core> _qz: indeed it is
* core/#tunes looks for xchat exploits
<_QZroot> i havent created any users yet and my system is a fresh install
<core> _ruiner_: background pictures for brix's desktop :)
<_ruiner_> lmao
<_QZroot> so fucking me over wont do any good
<core> _qzroot: but you are running as root
<_QZroot> i know im root
<core> _qzroot: so if i can get xchat to do something, it will do so as root :)
<_QZroot> look at my nick
<core> _qzroot: duh
<_ruiner_> planning on launching brix in san fran?
<_QZroot> gnome is wicked
<core> _ruiner_: yeah, they already called henri bellolo so the village people will be there
<_ruiner_> the indian is my favorite
<core> oh yeah, let's not forget the indian
<core> village people ensured enough money for all french people for a few years ;)
<_ruiner_> french people suck!
<core> ruiner: the female part only. well, some male too, but..
<_QZroot> ACK its not using mouse focus
<core> _qz: put on your site that brix got lost in the backup, that'll buy you more time :)
<_QZroot> hahaha
<_QZroot> brix was the one thing that cant get lost
<core> why?
<_QZroot> i put copies all over the place before formatting
<_ruiner_> its like a virus
<_ruiner_> its a pain to get rid of
<core> use the best backup system on earth
<core> release it!!!!! :)
07:20pm
<_QZroot> hah
<_ruiner_> yeah, release it and all kinds of people will have it
<_ruiner_> that'll ensure that even if you do lose it chances are somebody else will have it
<_QZroot> god damn, gnome bitches about everything when root
<core> of course
<_ruiner_> lmao
<core> it's stupid to run X as root :)
<core> or mostly everything for that matter
<_ruiner_> its stupid to be logged in as root unless you're adding users
<_QZroot> ARG
<_QZroot> this uses that passwd crap that bespin has
-:- SignOff HickServ: #TUNES (Ping timeout for HickServ[209-68-229-62.dialup.cust.tfb.com])
<_QZroot> i like 3 char passwds
<_ruiner_> nin
<_ruiner_> oh yeah!
<_ruiner_> I use that a lot
<core> 3 char passwords?
<_QZroot> i use my computer as my alarm and its harder to type really long complicated passwds in my sleep
<_ruiner_> heh heh
<core> so basically i can find your password in hmmmmm...
<_QZroot> a short easy passwd and i can turn it off and be back in bed without losing much sleep
<core> 60^3 attempts
<core> 216000 attempts
<core> not too bad
<_QZroot> BAD PASSWORD: it does not contain enough DIFFERENT characters
<_QZroot> ds9xxx
<core> good dog
<_ruiner_> I used zzzzzz once
<_QZroot> how do i turn this shit off
<_ruiner_> oops
<_ruiner_> one too many z's
<core> you wouldn't think you found my passwords if you find them
<core> _qz: find actual passwords?
<_QZroot> it wants me to type totally random letters/numebrs
<core> good boy.
<_ruiner_> booooo
<_ruiner_> those suck
<_ruiner_> hard to remember
<core> i remember them
<_QZroot> my last passwd that i used for years and years was 'cold'
<_ruiner_> my gramps has his logon id and pw on a piece of paper taped to his desk under his mousepad
<core> if i can remember the contents of 200.000 lines of code, i can remember a few passwords.
<_QZroot> and root was 'yup'
<_ruiner_> I go insane when I program
<_QZroot> core: is clementine 200000 lines?
07:30pm
<core> _qz: let me check..
<_ruiner_> I can think abstractly and remember what some wierd line does.....but when I'm not coding, I can't remember what I had for lunch unless I burp
<core> okay, only 100.000 so far, but i'll get there :)
<_QZroot> my god
<_QZroot> brix is like 8000 lines
<core> brix doesn't do much?
<_QZroot> of asm
<core> so far
<_QZroot> yes it does
<_QZroot> what does clementine do?
<_QZroot> d
<_QZroot> brb, gotta change users
-:- SignOff _QZroot: #TUNES (xchat exiting..)
<core> let's see.. it's fully component based, using interfaces and runtime binding.. has protected virtual memory.. preemptive multithreading and multitasking.. synthetic recursive files and a synthetic filesystem.. ggi drivers.. remote gdb support..
<core> probably QZ didn't write a single line of comment in brix.
<core> mine probably account for 35% of the number of lines.
-:- _QZ [brand@p0wer.qzx.com] has joined #tunes
<_ruiner_> comments are evil
<core> let's see.. it's fully component based, using interfaces and runtime binding.. has protected virtual memory.. preemptive multithreading and multitasking.. synthetic recursive files and a synthetic filesystem.. ggi drivers.. remote gdb support..
<_ruiner_> unless somebody else will read your code
<core> probably QZ you didn't write a single line of comment in brix.
<_ruiner_> then its a must have
<_QZ> i use lots of comments
* core/#tunes waits for _QZ to tell him brix does all of the above plus tcp/ip and he's running it now
<core> so what does brix do?
<_QZ> however, most of them are along the right side
<core> i have blocks of comments and most of them are seperate lines. c is longish
<_QZ> so my line count is mostly code
<_QZ> i dont know what yer synthetic crap is but brix does all of that
<_QZ> cept for the gdb shit
<_ruiner_> brb
<core> you have ggi drivers and remote gdb support (not that this one accounts for much code) ?
<_ruiner_> maybe
-:- SignOff _ruiner_: #TUNES (Leaving)
<_QZ> and i use my own gfx instead of ggi
<core> ggi drivers take some room.
<core> and of course you can demonstrate that on a bootable binary image now?
<_QZ> ya
<core> email it to me.. core@suntech.fr
-:- Wandrr [Wandrr@client-151-200-118-28.bellatlantic.net] has joined #Tunes
<Wandrr> Hello all 
07:40pm
<_QZ> hello
<abi> hey, _QZ
<_QZ> shutup abi
<Wandrr> core - I was taking a look at the logs - you wrote a kernel-less OS? 
<_QZ> ya hes gonna write the kernel later :)
<Wandrr> lol QZ 
<_QZ> s/hes/he is/
<Wandrr> I (tried to) finish reading Brian Rice's article on the Arrows system 
<core> qz: eh?
<core> :)
<core> wanddr: yes
<core> _qz: did you email the image
<Wandrr> I'm wondering if anyone has made a nano-kernel of 'intentions' like in the micos0ft IP system 
<core> wanddr: there is room from improvement of course, i'm sure
<Wandrr> hey core! 
<core> wanddr: sorry, i was away
<core> wanddr: what's the m$ intentions systems? requests for functionality?
<_QZ> gradient desktop, nice :)
<core> _qz: did you email your image?
<Wandrr> It's looks a lot like an abstract syntax tree manipulation system 
-:- Nyet [IDT@ppp03-225.algx.iadfw.net] has joined #Tunes
<_QZ> core: yes
<core> _qz: okay, checking mail now
<core> wanddr: sounds good so far
-:- HickServ [thrustit@209-68-229-88.dialup.cust.tfb.com] has joined #tunes
<Wandrr> basically, opening up the compiler's goings on so they can be modified as a part of the program you write 
<core> wanddr: what do they use it for? dynamic compilation and code specialisation?
<HickServ> uh i was really lagged
<core> wandrr: ohh.. reflectivity then
<Wandrr> Well, Simonoyi says you can use it for anything 
<HickServ> could you tell me that file name one more time?
<Wandrr> he's saying to try to - yes, reflectivity, which is how I found this place - he's trying to eliminate syntax difficulties 
<core> wandrr: if you can modify the syntax tree inside the program that the syntax tree describes, it is reflectivity
<core> _qz: nothing so far.
<core> _qz: what does the image you sent me do?
<Wandrr> Rather than sticking to an artificial syntax and expecting programmers to know a language *and* the problem domain, make the problem domain the program 
<Wandrr> xl, core 
<core> wandrr: xl? :)
<Wandrr> So I've been trying to figure out the minimum necessary (xl=exactly) native code to start a reflective system 
<core> wandrr: well that sounds logical.. reflectivity should solve a lot of problems
07:50pm
<core> wanddr: you don't need native code
<core> wanddr: you need to pretranslate some of your code into native code for bootstrap though
<Wandrr> It has to run, though, right? 
<Wandrr> yes - bootstrapping. I want a five-byte bootstrap, and everything else to be coded in intentions 
<Wandrr> (ok, not really five bytes) 
<core> wanddr: that can probably be done (and yeah i know what you mean)
<core> wanddr: all you need is a native stupid loader code to read your pretranslated (to native code) code that handles syntax trees
<core> wandrr: then the rest doesn't even have to be directly translated i guess
<core> wandrr: clementine needs a loader, memory management and syntax tree management before the rest can be loaded.
<core> wandrr: there's no dynamic compilation though, only interface matchmaking
<core> (well, there is some, but only to connect interfaces)
<Wandrr> right - I'm thinking (i'm not a compiler expert, by any means) of letting it only complile what it needs to run something, in a segmented way, as needed 
<core> you can do that too
<Wandrr> interface matchmaking? 
<core> you can even heuristically decide to inline some code during that process
<core> yeah.. components of clementine describe what they provide (export) and require (import) in an interface
<Wandrr> how detailed is the description? 
<core> it's a language
<core> it's not like "i am bobby. i need tim and tom."
<core> :)
<Wandrr> [I've been thinking that source code, as an AST, should be self-documenting too.] 
<Wandrr> =) shoot, I was hoping for just that :) 
<core> heheh
<core> nah, it's more like corba's IDL but it doesn't compile any stubs !!!! and it's fast.
<core> and it doesn't use message passing when it doesn't need to, like in-process calls.
<Wandrr> I have recurrent daydreams of an interactive programming environment, where you get natural language responses to mis-syntaxed or expressed constructs 
<core> all components can have multiple entry points
<HickServ> core: im sorry my dad had to use the comp, what's that file your name is in?
<core> hickserv: what file name?
<core> hickserv: oh
<HickServ> in linux
<core> hickserv: /usr/src/linux/drivers/video/fbcon.c
<HickServ> k
<HickServ> hmmmmmm
<HickServ> that directory doesn't exist
<core> what kernel?
<HickServ> 2.0.something
<core> ah
<core> well, this file only exists on 2.1.107 and above (that means 2.2.x)
<core> wandrr: girlfriends do that :)
<Wandrr> lol core :) 
08:00pm
<core> i'm the person responsible for the fact that the linux console is black from 2.1.107 to 2.1.109 *duck*
<Wandrr> Yes, but they don't help you figure out what you meant.. they just assume. 
<core> wandrr: true
<core> wandrr: the day a computing system does that, it'll be scary :)
<Wandrr> Too m$ helpful/useless 
-:- SignOff Nyet: #TUNES (Leaving)
<Wandrr> Scare me! 
<HickServ> i wanna be able to say "i have my name in the tunes source that's good enough"
<Wandrr> Hey, me too! what can I help code? 
<core> hickserv: well, i'm not really running around saying i have my name in the linux kernel, that'd be ridiculous :)
<HickServ> um
<HickServ> doesn't mean i won't :)
<core> wandrr: write a reflective syntax tree management framework and we'll be your friends for life ;)
<Wandrr> Great - that's just what I'm trying to do - and it's why I came here to ask questions =-) 
<core> wandrr: i guess you'll need a language to compile into it as well :P
<core> wandrr: although something as simple as forth or lisp are good candidates for reflexivity
<Wandrr> I've already picked the language - it just hasn't been developed yet 
<core> wandrr: great!
<Wandrr> I love Forth 
<core> wandrr: what language will you use?
<Wandrr> my father wrote a forth compiler in fourth back in '78-'80 
<HickServ> forth is good
<core> did he? awesome :)
<HickServ> heh
<Wandrr> sped it up like 10 times on our TRS 80 
<core> yeah, forth is neat. infinitely expandable, minmum start dictionary
<HickServ> yeah my forth compiler will have a pretty minimal dictonaly too
<Wandrr> Actually I keep thinking that my 'native code' should be assembly, but then I want to reify assembly into something smaller 
<HickServ> i can't really ca it a forth compiler though
<HickServ> it won't exactly be forth
<core> wandrr: tom novelli is writing a prototype called retro in assembler
<core> wandrr: if you look for a C->compiler->syntax tree->runtime->binary code system, there is something decent called ANDF
<core> but of course, not reflective
<HickServ> uh
<Wandrr> the lower I get towards a non-kernel kernel, the more i realize I have to get my high-level abstractions worked out in extensive detail.. 
<HickServ> damn reflexivity is a lot more than i understand
<core> wandrr: that's very true
<core> wandrr: i work on my system most of the day, but still, i spent that much time for about 2 months thinking about the high-level details, and still it's perfectible.
<Wandrr> ..but that defeats the purpose of an abstracted system. A 'new' programmer shouldn't have to have a million concepts about how the internals of a 'language' operate just to code. 
<core> hickserv: reflexivity is basically the property of a language to describe itself
<Wandrr> You were talking earlier (in the logs) about eliminating the difference between the user and the programmer, right? 
<core> wandrr: it's very hard to think in a generic way for making that happen, though
<Wandrr> Natural language. 
<core> wandrr: yeah, in my rant about girls not liking current ways of programming i think :)
<_QZ> hmm, netscape is having spasms
<core> wandrr: yeah, of course.. 
<core> i suppose you can make a natural language parser if..
<core> 1) you use a reflective syntax tree
<core> 2) it is context sensitive
<Wandrr> =) oh - maybe I caught you at a bad spot *lol 
<core> 3) your computer is networked and can socialise ;)
<HickServ> core: well that doesn't tell me much
<Wandrr> that's key! :) 
<core> wandrr: no.. i was just commenting on something :)
<core> wandrr: what is? :)
<Wandrr> networking & socialization 
08:10pm
<core> wandrr: oh. yeah it is
<Wandrr> Havn't the foggiest about how it would happen, though. 
<core> wandrr: me neither. i have been discussing that a lot with a friend who wants to create intelligent agents (real ones) tho
<core> wandrr: those would badly need to socialize if they want to be useful
<Wandrr> Maybe we make a natural language system, feed it full of relations, then let it work out similarities with what it knows with other pc' 
* HickServ/#tunes is creating awn intelligent agent
<HickServ> an
<Wandrr> s on the net.. (gettin' crazy, now :) 
<Wandrr> how so, HickServ? 
<core> wandrr: that was actually the topic of my discussion with him. getting to a point where those agents can work out similarities
<Wandrr> really? by developing a coding method that highlights them? 
<HickServ> well it's basically a combination of context free grammars, fuzzy logic, and a little bit of ann's
<core> wandrr: for example, if you show a child a green and a red cube, then a green pyramid, and ask him to complete, it will pick the red pyramid
<Wandrr> and he has that working? 
<core> wandrr: key word was 'discussing' :)
<core> wandrr: but he thinks he can do it with valued graphs
<Wandrr> gotcha :) 
<Wandrr> I spend most of my thoughts on neural networks, actually. since HS - 10 years 
<core> wandrr: a reflective language is key to the system though
<core> i don't know much how those work, besides the theory
<Wandrr> pc's are missing a critical function : there's no "like" comparison 
<HickServ> neural networks are limited by current technology
<core> i have a friend who wrote a neural network system for image compression that works like crazy tho :)
<Wandrr> they're great at that... 
<HickServ> that's why people like me are looking for shortcuts
<Wandrr> I wonder how a nn-based state machine could be, maybe, taught how to emulate a CPU 
<core> wandrr: yeah, they just taught the neural net how to go from picture 1 to picture 2 and the other way around i think
<core> wandrr: i don't know.. but for your 'like' comparison they just need tons of information
<Wandrr> if you can make a nn asssociate A with B, then with C and D in  a chain... 
<core> wandrr: the human brain has it easy for that, our five senses give it millions of informations a second
<core> wandrr: and then it sorts and associates them, and we can reflectively operate on ourselves
<Wandrr> ... then let it 'naturally' (through its learning function) tend to associate A with D, you have automatic optimization 
<core> wandrr: right
<Wandrr> well - it's still association, not true reflectivity 
<Wandrr> Reflectivity comes from being able to collect data about *yourself* 
<core> true
<core> but no
<HickServ> i can't see ann's being implemented in an os
<core> you can only do that by collecting data about others
<Wandrr> no - bnn's 
<core> others that are like you
<core> i don't think you can learn much about yourself if you're alone
<Wandrr> hmmmm. that's interesting.. 
<core> you can't know what's unique to yourself and what's common to other people, in this case
<Wandrr> the way we think is available as an 'extra' data source to each of us.. 
<core> only external input from beings similar to you allows you to isolate what is your own
<Wandrr> .. but what we think is often, whether we know it or not, a product of what's outside of us, 
<core> right
<core> i really think networking and socialisation are the key to such a system
<Wandrr> Gee, I don't think anything of mine is 'my own'.. (reflecting) 
08:20pm
<core> communication is what allowed humans to distinguish from animals anyway
<core> our rich communication i mean
<core> wandrr: sure you do
<core> wandrr: you have your own personality
<Wandrr> See, if Micro... had made their system reflective, we'd already have the critical mass we need for all these PC's to start getting really smart 
<Wandrr> :) 
<core> wandrr: probably so :)
<HickServ> well we ahve sensor input and then interaction which when we understand their relationship we are sentient
<core> wandrr: although with such a crazy thing as the internet today, a few thousands boxes would probably be enough; it would grow exponentially
<Wandrr> I do - sure, but I mostly just resonate the ideas of others - sometimes the ideas they can't quite say yet, but others nontheless 
<Wandrr> I keep thinking of making a big webcam analyzer... 
<core> wandrr: but you only know yourself because you can communicate with others :)
<Wandrr> I only think about who I am when I communicate - that brings it into contrast 
<Wandrr> which is what you said earlier.. 
<core> that's exactly my point :)
<core> yep, but i'm glad you agree )
<core> :)
<Wandrr> Glad to know I'm not the only one who thinks too much :) 
<Wandrr> can I ask you a semi-personal ? 
<core> heheh.. i don't think that much, but i like to give myself headaches :)
<core> sure :)
<Wandrr> are you a scientologist? 
<HickServ> hehehe
<core> no.. i'm french, they aren't that big here (yet) ;)
<Wandrr> lol 
<core> why? is that one of their big ideas?
<HickServ> i heard it's outlawed in germany
<core> i'm not up to date with their program :)
<Wandrr> Some of your stuff seemed along their lines... had to ask :) 
<Wandrr> really Hick? 
<Wandrr> I heard that too 
<HickServ> yeah 
<core> wandrr: i don't mind you asking :) i don't put a bad connotation on them, as i don't know them
<core> wandrr: i talk a lot, so i have to persuade myself that communication is good :)
<Wandrr> I used to work for them - in a private company.. had a bad impression for a while 
<HickServ> that's good of you core
<Wandrr> Ah, a rationalization of your personality 
<core> hehe.. but of course
<Wandrr> I've stumbled into something of the opposite - always *disproving* what I see as a character 'trait' 
<core> wandrr: what do you mean?
<Wandrr> I somehow don't like the idea that I have any sort of a 'fixed systemicy' to myself 
<core> putting yourself down all the time?
<Wandrr> You know, like when someone sees something about you - your 'main loop' if you will - that you don't see 
<core> yeah
<core> my girlfriend is good at that
<Wandrr> I tend to lately - trying to debug this the past few years :) 
<Wandrr> Girls get this like they've got an organ for it 
<core> not saying that just because she's my girlfriend.. she is good at spotting things about myself that i'm not aware of
<HickServ> i must be off for another while 
<HickServ> bbl
<core> okay hick :)
<Wandrr> I think that's a crucial resource to have 
<HickServ> :)
<core> wandrr: same. been working on it for a few years too
<Wandrr> Bye Hick 
<core> wandrr: yeah, she's a crucial resource to me anyway :)
<Wandrr> They get right into the kernel, eh? 
<core> yeah, they sneak in by standard syscalls and then blow the entire thing up to replace it with something better ;)
<Wandrr> I was reading an article by Ken Thompson about how he hacked the C compiler code with a Uni* logon bypass 
08:30pm
<Wandrr> talking about recompiling the compiler code to detect the login routine code, then 
<Wandrr> insert his bypass, 
<core> hehe
<Wandrr> then recompile it again with a modification that would not require his original 'hack' to show up in the source 
<core> that's really sneaky
<core> and probably bad because someday or other another program would use the same sequence and brutally crash for no reason when the 'patch' is applied :)
<Wandrr> Quite. He used it as a way to explain how the source isn't always the truth. 
<core> well, i check my own code at object level sometimes to see if it didn't compile too stupidly, so i'd agree :)
<Wandrr> sounds like your girlfriend did the same thing Tunes wants to do with OS's & programming :) 
<Wandrr> I have a bad habit of checking the assembly when I refuse to believe I made a simple C-code mistake:) 
<Wandrr> .. which is usually what I did 
<core> hehe
<core> well i trust my compiler, but i still check the assembly for bad optimization
<Wandrr> Wow, this was fun! I'll have to come back and bug you more! 
<core> depends for what, of course i don't care much about init code
<Wandrr> Same here- 
<core> sure, or you can bug me at core@suntech.fr :)
<Wandrr> I started on that for one of my bnn's 
<Wandrr> next thing you know my whole C code was asm statements 
<Wandrr> lol 
<core> heheh
<Wandrr> i'm bryan_price@hotmail.com 
<core> okay :)
<core> almost homonym to brian rice :)
<core> my low-level C code has a few gotos. i don't get why c compilers haven't figured out that branches are expensive, yet.
<Wandrr> talk to you soon ... yes, I noticed the similiarity - but that's my real name :)O 
<core> okay, you will :) have a good evening
<Wandrr> night! 
-:- SignOff Wandrr: #TUNES (http://ppproject.com/product/iamc)
<core> yup!
<core> well, dear log, it's 5:41 AM here, before i am cut into little cubes by a female entity, i should head to bed.
-:- SignOff core: #TUNES (nanosleep();)
08:40pm
-:- _QZ [brand@p0wer.qzx.com] has left #tunes []
-:- _QZ [brand@p0wer.qzx.com] has joined #tunes
-:- SignOff _QZ: #TUNES (kernel panic)
-:- _QZ [brand@p0wer.qzx.com] has joined #tunes
<HickServ> i be back
<HickServ> kernel panic?
<HickServ> are's ya ok?
<_QZ> heh
<_QZ> thats a joke quit message
<HickServ> i see
<HickServ> :(
* HickServ/#tunes was scared
<HickServ> j/k
<_QZ> i gotta figure out howto tunr off these STUPID tootips
<_QZ> tooltips
<HickServ> are you on linux?
<HickServ> Rand?
<_QZ> yes
<_QZ> rand?
11:20pm
<HickServ> someone says he knows you and your name is rand
<_QZ> brand
<HickServ> dcc me /usr/src/linux/drivers/video/fbcom.c
<HickServ> please
<_QZ> no such file
<HickServ> :(
<HickServ> anything mabye i spelled wrong
<_QZ> what is it suppose to be?
<HickServ> look in that dir
<_QZ> i got some fbcon files
<HickServ> core says his name is in it
<_QZ>  *  Hardware cursor support added by Emmanuel Marty (core@ggi-project.org)
<_QZ> fbcon.c
<HickServ> mucho gracias
<HickServ> hmmmmm
<HickServ> im not getting anything
<_QZ> im new to this client
<_QZ> and i dont plan on keeping it
-:- Myrmidon [ubiquity@dm1-14.slc.aros.net] has joined #tunes
<_QZ> when i get bitchx reinstalled its gone
<Myrmidon> Greetings fellow esteemed colleagues and coal mine workers.
<Myrmidon> Err, I meant hi HickServ.
<_QZ> hello myr
<HickServ> buenas noches
<Myrmidon> Hello to you too.
<HickServ> hey myr you running linux?
<Myrmidon> Not currently, but it resides on another partition. :(
<HickServ> oh
<Myrmidon> I needed to use Windows momentarily.
<HickServ> well qz it doesn't look like this file is coming in
<Myrmidon> You need that header file, don't you??
<HickServ> no
<HickServ> a driver
<Myrmidon> Oh.
<HickServ> get it to me when you get bitchx qz
11:30pm
<HickServ> do you know someone with the nick of glimli?
-:- SignOff Myrmidon: #TUNES (Read error to Myrmidon[dm1-14.slc.aros.net]: No route to host)
<_QZ> wtf
<_QZ> he was using aros.net?????????
<_QZ> myrmidon lives in my city
<HickServ> i'll be damned
<_QZ> we both use aros.net
<HickServ> maneesh on the mailing list lives in my county
-:- Myrmidon [ubiquity@dm1-32.slc.aros.net] has joined #tunes
<HickServ> * Myrmidon begins singing "It's a Small World".
<HickServ> hmmmmm
<HickServ> im becoming suspicious
<_QZ> hey myr
<Myrmidon> Ok.
<_QZ> guess what
<Myrmidon> Hello again.
<_QZ> u use aros.net
<Myrmidon> I hate guessing. :(
<Myrmidon> Yes.
<_QZ> so do i
<Myrmidon> I know.
<Myrmidon> I knew you from #ASM on Efnet.
<Myrmidon> :(
<_QZ> hmm
<HickServ> myrmidon is part of the asm l33t
<_QZ> ugh, i cant get my samba server working
<HickServ> shah
<HickServ> pshh
11:40pm
-:- SignOff Myrmidon: #TUNES (Ping timeout for Myrmidon[dm1-32.slc.aros.net])
<_QZ> that boy has got problems
-:- Myrmidon [ubiquity@dm1-11.slc.aros.net] has joined #tunes
<_QZ> Myrmidon: having problems?
<Myrmidon> I must be masochistic to come on here for a third time.
<HickServ> should i give every smaller thread a page of memory?
<_QZ> huh
<HickServ> uh
<HickServ> for protection
<HickServ> like a thread can have 4k
<_QZ> 4k for what
<Myrmidon> I think my brother is mad at me for not using Linux and sharing the connection.
<HickServ> i feel like im being closed in
<Myrmidon> So he must be fooling with the phone line.
<HickServ> nevermind qz
<Myrmidon> Or using Windows is leading to my demise.
<_QZ> windows sucks
<Myrmidon> I need it for a certain program, though. :(
<HickServ> yes
<Myrmidon> Which WINE or VMWare can't remedy.
<Myrmidon> HickServ: you were using Windows to play me Ice Hockey, so bleh.
<HickServ> windows is a 32 bit patch for a 16 bit graphical user interface of a 8 bit operating system originally created for a 4 bit processor made by a company that can't stand one bit of competition
<HickServ> it isn't my comp
<Myrmidon> HickServ: tell me something I don't know.
<HickServ> im a poor mexican so i only have linux on my 486
<_QZ> windows?
<abi> windows is a 32bit patch for a 16bit shell on an 8bit OS originally made for 4bit machines by a 2bit company that can't stand 1bit of competition
<HickServ> YES!
<HickServ> well obviously that is well known
<HickServ> crack?
<HickServ> oh i thought abi would know that for sure
<HickServ> fare?
<abi> fare is sometimes looking at another screen and interruptible with a beep (/msg Fare ^G) or at http://www.tunes.org/~fare/
<HickServ> linux?
<abi> i heard linux was bloated
<HickServ> abi?
<abi> HickServ?
<HickServ> you are a sly one
<HickServ> tril?
<abi> tril is a security person. for tunes, though ,not unix ;)
<HickServ> tunes?
<abi> rumour has it tunes is in my head or tunes is a useful nevertheless expedient system. or at http://tunes.org/ or a free reflective computing system or to programming languages what the internet is to computer networks
11:50pm
<HickServ> c?
<abi> i heard c was the category whose objects are valid observable types through which we look at programs, and whose arrows are interactions we are allowed to do on programs
<Myrmidon> _QZ: do you use an ISDN terminal adapter to connect to Aros?
<HickServ> lisp?
<abi> i think lisp is easy, once you learn it :)  That could take awhile
<Myrmidon> Or does Magna now have DSL connectivity??
<_QZ> hah i wish
<HickServ> forth?
<abi> forth is at http://www.forth.org or a nice user interface
<HickServ> c++?
<HickServ> blarg
<HickServ> java?
<abi> java is ok
<_QZ> a 33.6 is overkill for magna phone lines
<HickServ> hcf?
<abi> hcf is mailto:nef@tunes.org, http://www.tunes.org/~nef/ or halt and catch fire (see: http://wombat.doc.ic.ac.uk/foldoc/foldoc.cgi?Halt+and+Catch+Fire)
<_QZ> abi: kick hickserv
<HickServ> no!!!
<HickServ> ?
<Myrmidon> _QZ: so you have that multiple analog to digital conversion crap in your phone lines too?
<HickServ> hehehe
<_QZ> ya
<HickServ> qz?
<abi> qz is the two least used letters of the alphabet
<_QZ> my lines are full of load coils
<_QZ> plus i live 9010 feet from the CO
<Myrmidon> Hahaha.
<HickServ> HickServ?
<abi> you are, like, the same
<HickServ> Myrmidon?
<abi> Myrmidon is part of the asm l33t
<HickServ> cool
<_QZ> dsl will be available in magna in the year 2050
<HickServ> BuenOS will own you
<_QZ> by that time 99% of america will have terabit lines in their homes
<HickServ> im getting adsl soon
<HickServ> BuenOS?
<HickServ> bastard bot
<_QZ> buenos?
<abi> buenos is muy mal
<HickServ> no
<HickServ> no!!!
<_QZ> abi: good girl
-:- BuenOS [thrustit@209-68-229-88.dialup.cust.tfb.com] has joined #tunes
<HickServ> BuenOS is cool
<HickServ> BuenOS?
<abi> rumour has it BuenOS is muy mal
<HickServ> no!!!
<_QZ> hahah
<_QZ> u dont know howto sweet talk abi
-:- SignOff BuenOS: #TUNES (:()
<HickServ> later
<HickServ> i gotta sleepx0r
<_QZ> no u fdont
-:- SignOff HickServ: #TUNES (night)
[msg(TUNES)] newlog 1999.0530
IRC log ended Sun May 30 00:00:01 1999