IRC log started Thu Sep 23 00:00:01 1999
[msg(TUNES)] permlog 1999.0923
-:- SignOff Zhivago: #TUNES (Leaving)
-:- Zhivago [brian@th.merddin.com.au] has joined #tunes
<Zhivago> hello there
<eihrul> lo
<Zhivago> that slab allocator looks interesting
<eihrul> get that :)
<eihrul> i had to search high and low for a free copy
12:10am
<Zhivago> ahh, lovely
<eihrul> and low and behold...
<eihrul> it was sitting, on of all people, Miguel de Icaza's page
<Zhivago> wow $80 for membership
<eihrul> 'xactly
<Zhivago> print print print
12:20am
<eihrul> mmm... slab allocator == good
<Zhivago> oh, its just a graveyard system
<Zhivago> was hoping for somthing more exciting
12:30am
<eihrul> it's a bit more than that though
<eihrul> the key part being caches are groups of slabs
<eihrul> and that slabs can easily be returned for other caches to use
<Zhivago> well, that's just a block-wise allocator
<Zhivago> I'm looking at a similar system for a persistent-virtual-memory implementation
<Zhivago> although the graveyard is handled above that level
<Zhivago> by vendors
<Zhivago> the main advantage I have here is that there's no difference ebtween kernel allocation and user allocatio
<Zhivago> they're all in the same system, so vendors and the memory system are in intimate contact
<Zhivago> the main concern I have now is localsiation of data, which I think I can do via a variation on a generation idea
<Zhivago> so basically its a pool allocator by witdth
<Zhivago> with graveyarding
12:40am
<Zhivago> hmm, an interesting paper
<Zhivago> although I have a feeling that they're doing it backward because of Unix design
<Zhivago> hmm, liar was saying some odd things about where tunes was headed
<Zhivago> could you clarify them?
12:50am
<eihrul> what things were those?
<Zhivago> well, he was talking about instead of context-switching dynamically constructing a stream of execution with the various tasks interwoven
<Zhivago> but I have a feeling that he might have been confused
<eihrul> the idea of tunes is to dynamically recompile the system
<eihrul> to complete the overall goal
<eihrul> or atleast, as he was saying
<Zhivago> yes, but taht's completley different to this
<eihrul> that if you knew all the individual tasks
<Zhivago> that's just profile driven compilation
<eihrul> you could optimize them into one globulous giant task
<eihrul> that did everything optimally
<Zhivago> ok, he seemed to think that this is how tunes would work normally
<eihrul> that's the idea :)
<Zhivago> which would kind of rule out use by humans
<eihrul> how so?
<Zhivago> given the highly variable latency issues associated
<eihrul> gah, minor hindrance!
<Zhivago> well, lets consider a case, we have a system which has three program son it
<Zhivago> it has a mail program, and an mp3 player, and a screensaver
-:- demoncrat [darius@shell.accesscom.com] has joined #tunes
<Zhivago> I, a human, have left this system running overnight
<Zhivago> and it has optimise itself perfecly as a screensaver
<Zhivago> now I start up my mail reader
<Zhivago> what happens?
<eihrul> exactly waht you expect to happen
<Zhivago> the system recompiles itself?
<eihrul> it's already recompiled itself long ago
<Zhivago> sooo, how are we avoiding task switching?
<eihrul> that's not my job to figure out :)
<Zhivago> that's not an answer.
<eihrul> i'm just a lurker man...
<Zhivago> I mean look at this, at night, its obviously not running optimally if its set up to run a mail-server
<Zhivago> er program
<eihrul> i cannot really speak for the tunes project :)
<Zhivago> so, either its suboptimal at night, or it rebuilds itself as a perfect saver
<Zhivago> oh, fair enough
<eihrul> the members don't speak for the tunes project often enough
<Zhivago> in any case, I don't see how this approach can work in practice
01:00am
<demoncrat> what is the approach?
<Zhivago> well, I'm not sure, the claim was dynamic recompilation in order to avoid task-switching
<demoncrat> hm.  sounds expensive.
<Zhivago> this could work in a homogenous environment
<Zhivago> yes
<Zhivago> if you were always doing the same thing, then global static program analysis profiler driven would work
<Zhivago> but you'd get _incredible_ latencies on doing anything hetrogenous
<Zhivago> which would pretty much rule out human usage
<demoncrat> I wonder if those new multithreaded architectures get you essentially the same benefit but in hardware
<Zhivago> demon: well, task switches are prett cheap these days in any case, especially if you're not rezapping the mmu
<eihrul> but tunes could anticipate a threaded architecture :)
<Zhivago> just a register save/restore really
<Zhivago> ei: it would get the same problem
<Zhivago> ei: just over more cpus
<demoncrat> this could be more useful with stuff like ia64 and it's incredibly huge contexts.
<demoncrat> er, its
<Zhivago> ei: optimal function for one task is suboptimal in general
<Zhivago> demon: yeah, but you have a huge cost in code production
<Zhivago> demon: which I suspect outweighs this cost for general purpose
<Zhivago> demon: if you're just running a mono-purpose server on the other hand...
<demoncrat> I can see it happening this way: you have profiling and profile-driven dynamic optimization going on all the time, and threads gets coalesced only when you're pretty sure it'll pay for itself.
<Zhivago> sounds reasonable
<Zhivago> but that's not a replacement
<demoncrat> but usually that infrastructure is going for other optimizations.
<Zhivago> have to see what fare has written lately
<eihrul> but i think that was supposed to be a side-effect of tunes
<eihrul> not it's actual purpose
<Zhivago> assuming there's any progress on that side
<Zhivago> heh, I should see what gclist is on about
<demoncrat> anyway, hi.  I don't think I've been here before.
* demoncrat/#tunes waves.
<Zhivago> oh, looks like a pretty pithed list
<Zhivago> july actually got 200k of mail :)
<demoncrat> any names?
<Zhivago> wilson
<demoncrat> that could be worthwhile.
<Zhivago> heh, guiseppi artadi
<Zhivago> attardi
<Zhivago> although I'm not sure impressed with him, having worked on ecl
01:10am
<Zhivago> boehm, presumebly of boehm's fame
<eihrul> there, finished that slab allocator paper
<Zhivago> shapiro I recognise, but I don't know where from
* eihrul/#tunes is appeased for the night.
<Zhivago> and of course fare
<demoncrat> what is a slab?
<eihrul> that explains it
<demoncrat> thanks
<eihrul> tis a nice way of managing memory allocating
<eihrul> er allocation
<Zhivago> demon: looks like same-width pooling with a graveyard, and refcounting
<eihrul> in other words, slab allocation :)
<eihrul> by any other name...
<Zhivago> or a backward way of doing packed generational memory allocation
<Zhivago> (with graveyarding)
-:- SignOff eihrul: #TUNES (sleep)
<Zhivago> well, this channel is dead again :)
<demoncrat> at least it's not full of s0ren and aliensex
<Zhivago> this is true
<Zhivago> but since people won't stand up
<Zhivago> what can I do?
<demoncrat> I've thought of simply kicking them both out until they show some interesting in lisp.
<demoncrat> maybe we should take the discussion private, though.
<Zhivago> I have no problem with that
<Zhivago> probably :) best not to disturb the dead
<demoncrat> that is not dead which can eternal lie
01:20am
-:- SignOff demoncrat: #TUNES (Leaving)
<Zhivago> hmm, bfe is hideously ugly code :|
<Zhivago> and the release is broken
02:40am
-:- AlonzoTG [Alonzo@216-164-128-73.s73.tnt1.lnhva.md.dialup.rcn.com] has joined #tunes
<AlonzoTG> tpp
-:- iefbr14 [user4808@gw.capgemini.fr] has joined #Tunes
<iefbr14> hi there.watabout round here 
06:00am
<Zhivago> what about it?
<iefbr14> it's oh, so quiet
<iefbr14> it's all so stil
<iefbr14> u'r all alone
<iefbr14> and so peacefull until.....
<Fare> hum
<Fare> salut
<Fare> on se connait?
<iefbr14> pas que je sache.
<Fare> comment connais-tu #tunes?
<iefbr14> slash list 
<Fare> slash list?
<iefbr14> irc.linux.com
<iefbr14> slash list
<iefbr14> ca donne les channels servis par ce serveur
<Fare> /list ?
<iefbr14> dont tunes.
<Fare> oh!
<iefbr14> c'est french speacking ici ?
<Fare> seen www.tunes.org?
* iefbr14/#Tunes goes to ....
<Fare> no, it's english speaking, although you can /msg in french...
<iefbr14> k. 
<Zhivago> wow, the screen filled up :)
<Fare> time to /clear
<iefbr14> <Zhivago> wow, the screen filled up :)
<iefbr14> <Zhivago> wow, the screen filled up :)
<iefbr14> <Zhivago> wow, the screen filled up :)
<iefbr14> easy..
<Zhivago> ie: well, its just that 5 minutes ago it hadn't moved for hours :)
06:10am
<Zhivago> fare: so, how's tunes progressing?
<iefbr14> 
<iefbr14> We don't have a well-defined language specification, even less a compiler, so Don't hold your
<Zhivago> ie: I haven't been :0
<Zhivago> ie: the retro project has been very useful to me though
* iefbr14/#Tunes picked it from the FAQ, and is not part of the 'WE'
<Zhivago> ie: I can now compile lisp to bootable images :)
<AlonzoTG> awesome!!!!
<Zhivago> after I rewrite the rest of the retro drivers in lisp, and we get the compiler self-hosting, it should actually be useful.
* AlonzoTG/#tunes needs a compiler expert to help with his OS
<AlonzoTG> =P
<Zhivago> alonzo: what os is that?
-:- iefbr14 [user4808@gw.capgemini.fr] has left #Tunes []
<AlonzoTG> u know my OS!!!
<Zhivago> alonzo: nope
06:20am
<Fare> Zhivago: is it progressing?
<Fare> ATG: what particular problem do you face?
* Fare/#Tunes is not an expert, but tries to be a meta-expert
<Zhivago> fare: is which progressing?
<Fare> Zhivago: Tunes :(
<Zhivago> fare: oh, I haven't noticed any progress :|
<Zhivago> fare: but we might have a project here that you might be interested in in a few months
<Zhivago> fare: see how it goes
<Zhivago> fare: although adopting retro seems to have been a step forward
<AlonzoTG>  How to ensure my LISP, C, and ASM code can talk to each other.
<AlonzoTG> =\
<Zhivago> alonzo: our solution w3as to remove the C and asm
<AlonzoTG> =\
<AlonzoTG> How do you get LISHP to cover for your ASM?
<AlonzoTG> How do you get third party programs (probably in C) to call your LISHP funcktions?
<Zhivago> alonzo: ahh, well, we extend our lisp dialect down to the assembly level
<Zhivago> alonzo: initially writing an assembler in lisp, and then reflecting it back in
<AlonzoTG> =\
<Zhivago> alonzo: wrt C we don't, we ditch it completely
<AlonzoTG> huh 
<Zhivago> C is too insecure to be useful in a single-space os
<AlonzoTG> oh single space OS...
<AlonzoTG> ./mine will be heavily polly space... 
<Zhivago> well, I don't see much point in multiple spaces ig you have a secure language
<Zhivago> one of the major advantages is that IPC becomes a non-issue
<AlonzoTG> yes...
<AlonzoTG> I have been worrying about IPC reciently...
<AlonzoTG> If you can get it streamlined so have a protocol by which prog A calls prog B like a library but has the permissions, in only that section of code, to make a trigger for a prog B thread.
<AlonzoTG> =\
<AlonzoTG> that way kernal calls can be almost completely avoided.
06:30am
<Zhivago> oh, I avoid that problem as well
<Zhivago> I ditch the kernel
-:- hcf [nef@me-portland-us225.javanet.com] has joined #tunes
<Zhivago> alonzo: the problem with IPC is the processes
<Zhivago> alonzo: why do we need processes? only for conservative security
<Zhivago> alonzo: with security at the code level we don't need this mechanism
<Zhivago> alonzo: which is why I'm only interested in a single address space
<AlonzoTG> hmmzerz...
* AlonzoTG/#tunes doesn't want to limit the programmer at all...
<Zhivago> there are always limiations
<Zhivago> limitations give power, at the expense of flexibility
<AlonzoTG> doesn't see how he can enforce sekuritie in softwarez without taking a performance hit.
<Zhivago> alonzo: well, we do it in the compiler
<Zhivago> alonzo: we need a fairly intelligent compiler to help us elide checks
<Zhivago> but consider this.
<Zhivago> int i = rand()%20;
<Zhivago> this returns an integer between 0 and 19 - necessarily, since % is well defined
<Zhivago> now if I have a char a[20];
<Zhivago> then I can safely say a[i]
<Fare> ATG: Asm can, by definition, always communicate with anything
<Zhivago> it is impossible for me to violate the bounds of the array's uinterface
<Fare> ATG: so the difficulty is with LISP&C
<Zhivago> so we need no check
<Fare> ATG: the next question is 1) what lisp, and what C?
<Zhivago> if we extend this concept thoughout the rest of the primitives, then we end up with very few run-time checks, and secure code.
<Fare> (see http://www.elwoodcorp.com/ for a LISP->C compiler)
<AlonzoTG> dunno
* AlonzoTG/#tunes is an idiot.
<AlonzoTG> hmmzers...
<Fare> the deeper question is: what programming model for your system?
<Zhivago> lisp is essentially a very simple functional machine
<Fare> you don't need two systems programming language
<Fare> settle for one
<AlonzoTG> perhaps i am just trying to get my prototype to do too much
<AlonzoTG> =\
<Fare> or else, settle for two, and define an interface (something like CORBA?)
06:40am
<AlonzoTG> I'm not sure even how to get ASM and C to agree on the size and shape of data structures.
<Zhivago> alonzo: you make the asm do it the C way
<Zhivago> alonzo: which is a nuisance
<Zhivago> alonzo: C isn't a very good language for this level imho
<AlonzoTG> =\
<Zhivago> it lacks access to registers, and op-codes
<AlonzoTG> I had a C compiler that did.
<Zhivago> which is why I've chosen to use a lisp which has been extended to the op-code/register level
<AlonzoTG> hmmzers
<AlonzoTG> I am not good enuff to program LISHP yet...
<AlonzoTG> I really need a compiler expert to suggest a good development language for me.
<Zhivago> alonzo: ok, then the question is, why are you developing an os?
<AlonzoTG> I use windoze 3.11
<AlonzoTG> I need something better.
<Zhivago> there are better pre-packaged solutions
<Zhivago> fbsd, linux, nt, etc.
<AlonzoTG> Its an integral part of my world domination plans >= )
<Zhivago> do you have any useful os development skills?
<AlonzoTG> they all suck
<AlonzoTG> I guess not, but I have been studying hard for years and years and years
<Zhivago> well, you need to have an idea of what doesn't suck
<Zhivago> can you give some features that you want.
<AlonzoTG> I want it to feel like DOS.
<Zhivago> have you looked at DrDOS?
<AlonzoTG> A little.
<AlonzoTG> Unfortunately, it sucks too.
<Zhivago> well, I think you don't have a sufficiently clear idea of what you want to do to actually do anything
<Zhivago> you need to decide on what you want to achieve
06:50am
<AlonzoTG> users.erols.com/alangrimes/
07:00am
-:- SignOff Zhivago: #TUNES (Ping timeout for Zhivago[th.merddin.com.au])
-:- Zhivago [brian@203.26.11.253] has joined #tunes
* AlonzoTG/#tunes left clicks on Zhivago
<AlonzoTG> see mi website?
-:- SignOff Zhivago: #TUNES (Ping timeout for Zhivago[203.26.11.253])
07:10am
-:- Zhivago [brian@th.merddin.com.au] has joined #tunes
-:- Piles [helpdesk@krypton.wi.leidenuniv.nl] has joined #Tunes
-:- Piles [helpdesk@krypton.wi.leidenuniv.nl] has left #Tunes []
-:- SignOff hcf: #TUNES (Leaving)
-:- Crimson [crimson@chaosdev.org] has joined #tunes
<Crimson> Hello.
<Crimson> Currently working hard on chaos, and the GUI is starting to become something.
08:00am
<Zhivago> crim: what is chaos about?
08:10am
<Crimson> It is my operating system. Look at http://www.chaosdev.org
08:30am
<AlonzoTG> om
<Zhivago> oh, microkernel
<Zhivago> how are you for drivers?
08:40am
-:- SignOff AlonzoTG: #TUNES (Have Nice Day :))
-:- hcf [nef@me-portland-us117.javanet.com] has joined #tunes
<Fare> h c f
<hcf> f a r e
<Fare> h o w   a r e   y o u   ?
10:00am
<hcf> f i n e, a n d u ?
* Fare/#Tunes is back from TPHOLs99, where I found other people somehow interested in a proof system based on category theory...
* hcf/#tunes is away: (afk)
10:10am
-:- AlonzoTG [Alonzo@216-164-130-113.s621.tnt1.lnhva.md.dialup.rcn.com] has joined #tunes
<Fare> abi: SAC?
<abi> SAC is Single-Assignment C, a revamp on SISAL, meant to have C-like syntax while preserving pure functional semantics, efficiently supporting *multidimensional* arrays, for parallel optimized number crunching or see sacbase
<Fare> abi: sacbase?
<abi> sacbase is at http://www.informatik.uni-kiel.de/~sacbase/index.html
10:30am
-:- tmf [s720@barkebille.ii.uib.no] has joined #tunes
<tmf> hi
10:40am
<Zhivago> hmm, single assign is a goo thing
* AlonzoTG/#tunes left clicks on tmf
<tmf> alonzoTG, does it crush me or just display a menu?
<Zhivago> or possibly bloth
<tmf> anyone here have any experience with cgi or plugin programming?
<AlonzoTG> sorry..
<AlonzoTG> but I wouldn't expect it to be too difficult.
<Zhivago> tmf: some
<tmf> great!
<tmf> I have this particular thing I awnt to do...
<tmf> Basically I want to be able to drag and drop files in a form.
<Zhivago> this is not cgi
<tmf> ( The files I'm using are too be to write into a text window )
<Zhivago> this is javasscript evil
<Zhivago> you can do this, however it requires all sorts of security sillinesses
<tmf> how?
<Zhivago> imho it is not worrth it
<Zhivago> developer.netscape.com
<tmf> imo it's neccasary for what I want to do
<tmf> thanks, I'll check out the site
<Zhivago> capture the d&d evens, but you'll need complete access to ths machine, and a few other security clearences first
<Zhivago> which is a real nisance, otherwise you won't be able to tel what was dropped
<Zhivago> and then to upload it, you'll need to build a forwith an upload widget, fill it, i and press the button magically
10:50am
<tmf> I'm going to use this to send parameters to cgi scripts ( using forms ), so the thing collecting the data is on the user side
<tmf> so there is no need for a security bundle
<Zhivago> that's what you think
<tmf> the users will have the software on their side.
<Zhivago> you need the security clearances at the brower end to talk to the machine its runing on
<Zhivago> to do d&d
<tmf> communication will only use cgi scrits
<Zhivago> tmf: that's all irrelevent :)
<Zhivago> tmf: never mind, you'l find out the hard way
<tmf> what's d&d?
<Zhivago> drag and drop
<tmf> ok:)
<tmf> so this isn't really possible - it's not a turn on/off thing?
<Zhivago> turn on/off?
<tmf> it's not an option - the idiots who came up with netscape/cgi/html/java didn't think about it so it's no f***ing possible.
<Zhivago> its possible to do as I've said
<Zhivago> you just need to get a vast number of security authorisations for the machine
<Zhivago> if you have certificates its not too hard
<tmf> if it poses security problems beyond a normal form then it's not possible.
<Zhivago> well, that's your choice, effectively no
<Zhivago> use the file upload widgets instead
<tmf> where do I want to go on the netscape page?
<Zhivago> probably into te javascript referenc or something
<tmf> jesus, the now their taling about DHTML - i just gotta get it !
11:00am
<Zhivago> dhtml means html + javascript etc
<Zhivago> which is what your'e talking about anyhow
<tmf> ok :) I'm just being negative!
11:10am
-:- SignOff hcf: #TUNES (Leaving)
-:- CRASH [user4875@modem-2.tinactin.dialup.pol.co.uk] has joined #Tunes
-:- SignOff CRASH: #TUNES (Leaving)
-:- SignOff AlonzoTG: #TUNES (Have Nice Day :))
-:- NetSplit: clarke.openprojects.net split from sterling.openprojects.net [02:07pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [clarke.openprojects.net]
-:- Netjoined: clarke.openprojects.net sterling.openprojects.net
-:- Tril [dem@bespin.dhs.org] has joined #Tunes
-:- ServerMode/#Tunes [+o Tril] by clarke.openprojects.net
-:- eihrul [lee@usr5-ppp82.lvdi.net] has joined #tunes
-:- SignOff tmf: #TUNES (Leaving)
-:- demoncrat [darius@shell.accesscom.com] has joined #tunes
-:- pyro [tcn@cci-209150250122.clarityconnect.net] has joined #tunes
-:- NetSplit: irc.linux.com split from devlin.openprojects.net [05:36pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [irc.linux.com]
-:- Netjoined: irc.linux.com devlin.openprojects.net
-:- Tril [dem@bespin.dhs.org] has joined #Tunes
-:- Crimson [crimson@chaosdev.org] has joined #Tunes
-:- demoncrat [darius@shell.accesscom.com] has joined #Tunes
-:- eihrul [lee@usr5-ppp82.lvdi.net] has joined #Tunes
-:- ServerMode/#Tunes [+o Tril] by sterling.openprojects.net
-:- abi [nef@bespin.dhs.org] has joined #Tunes
-:- pyro [tcn@cci-209150250122.clarityconnect.net] has joined #Tunes
!irc.linux.com!! Received :sterling.openprojects.net SERVER jordan.openprojects.net from asimov.openprojects.net !?!
!irc.linux.com!! Received :sterling.openprojects.net SERVER varley.openprojects.net from asimov.openprojects.net !?!
-:- NetSplit: varley.openprojects.net split from sterling.openprojects.net [05:38pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [varley.openprojects.net]
-:- Fare [fare@quatramaran.ens.fr] has joined #Tunes
-:- Netjoined: varley.openprojects.net sterling.openprojects.net
-:- pyro [tcn@cci-209150250122.clarityconnect.net] has joined #tunes
-:- abi [nef@bespin.dhs.org] has joined #Tunes
-:- NetSplit: irc.linux.com split from asimov.openprojects.net [05:42pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [irc.linux.com]
-:- Netjoined: irc.linux.com asimov.openprojects.net
-:- abi [nef@bespin.dhs.org] has joined #Tunes
-:- NetSplit: varley.openprojects.net split from irc.linux.com [05:46pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [varley.openprojects.net]
-:- Netjoined: varley.openprojects.net irc.linux.com
-:- abi [nef@bespin.dhs.org] has joined #Tunes
-:- iLLf8d [moe@206.117.14.93] has joined #Tunes
<iLLf8d> hmm
<pyro> howdy
<iLLf8d> i'm looking for a cool openprojects channel to hangout in
<iLLf8d> so far everyone is so uptight or just plain a$$holes
<pyro> hehe
<pyro> this channel is unusually dead
!doogie:*! yes, we know irc.linux.com is having issues.  please be paitient.
<pyro> where have ya been so far?
!wichert:*! you mean it's suffering from VA's infamous disappearing network?
<iLLf8d> how come linux users only come in 2 types
<iLLf8d> cool and dicks
<iLLf8d> well i usually sit in debian
<iLLf8d> although alot are ok there are a lot that aren't
<pyro> I went back to slackware :)
<iLLf8d> i've never tried it
<pyro> anyway most Linux users I see are obsessed with it
!netgod:*! this wont hurt a bit
-:- NetSplit: irc.linux.com split from sterling.openprojects.net [05:54pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [irc.linux.com]
<iLLf8d> started with rh then suse then debian and installed caldera openlinux (but haven't used it 16 colors yuck)
-:- Netjoined: irc.linux.com sterling.openprojects.net
-:- abi [nef@bespin.dhs.org] has joined #Tunes
<iLLf8d> woa
<iLLf8d> netsplit bigtime
<iLLf8d> =)
<pyro> heh
<pyro> "this won't hurt a bit" my ass ;)
<iLLf8d> i've been in stampede in and out alot and have almost ever seen anyone talking
<pyro> ok, it only hurt abi that time and abi's a stupid bot
-:- liar [brand@p0wer.qzx.com] has joined #tunes
<pyro> hey
<iLLf8d> hi
<liar> hey
<iLLf8d> liar
-:- NetSplit: irc.linux.com split from asimov.openprojects.net [05:59pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [irc.linux.com]
<liar> if u guys see core tell him that kevin fixed his e9 hack
<pyro> hehe
<iLLf8d> most of debian splits from my server
-:- Netjoined: irc.linux.com asimov.openprojects.net
-:- abi [nef@bespin.dhs.org] has joined #Tunes
06:00pm
<pyro> try sterling.openprojects.net
<pyro> only irc.linux.com seems to be splitting
<iLLf8d> i don't care really
<iLLf8d> i left that channel anway
<pyro> tried #f-cpu yet?\
<iLLf8d> i've gotta read my samba book so i can get stuff running the way i'd like
<iLLf8d> nope
<iLLf8d> i joined 
<iLLf8d> only 2 in there and they look liuke ghosts
<pyro> oh well.. I wanna go work on my OS stuff
<iLLf8d> well seeya 
<iLLf8d> gotta read
-:- SignOff iLLf8d: #TUNES ([x]chat)
-:- NetSplit: irc.linux.com split from sterling.openprojects.net [06:03pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [irc.linux.com]
<pyro> later, liar
-:- SignOff pyro: #TUNES (.)
-:- Netjoined: irc.linux.com sterling.openprojects.net
-:- abi [nef@bespin.dhs.org] has joined #Tunes
-:- NetSplit: irc.linux.com split from asimov.openprojects.net [06:08pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [irc.linux.com]
-:- Netjoined: irc.linux.com asimov.openprojects.net
-:- abi [nef@bespin.dhs.org] has joined #Tunes
-:- NetSplit: varley.openprojects.net split from irc.linux.com [06:09pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [varley.openprojects.net]
!sterling.openprojects.net!! Remote CONNECT varley.openprojects.net 8005 from netgod
-:- Netjoined: varley.openprojects.net irc.linux.com
-:- abi [nef@bespin.dhs.org] has joined #Tunes
06:10pm
!doogie:*! we know irc.linux.com is having issues.  we just can't connect to it because of these issues, so we can't do much. :|
-:- SignOff demoncrat: #TUNES (Leaving)
-:- SignOff eihrul: #TUNES (Leaving)
-:- eihrul [eihrul@usr5-ppp82.lvdi.net] has joined #tunes
<liar> eihrul: kevin told me howto make bochs compile
<liar> configure --enable-debugger --enable-disasm
<liar> manually fix fpu/Makefile
<liar> make
-:- marisconr [RJ@ABD0E340.ipt.aol.com] has joined #Tunes
06:50pm
-:- marisconr [RJ@ABD0E340.ipt.aol.com] has left #Tunes []
<liar> my god did half the ram making plants in the world blow up or something?
07:40pm
<eihrul> pricing?
<liar> 32meg dimm $59
<liar> they were down to like $10
08:00pm
-:- iLLf8d [moe@PPP16M.NWC.Net] has joined #tunes
-:- SignOff iLLf8d: #TUNES ([x]chat)
-:- eihrul [eihrul@usr5-ppp82.lvdi.net] has left #tunes []
-:- ElGato` [no@209-68-229-80.dialup.cust.tfb.com] has joined #tunes
<ElGato`> HI :D
09:30pm
<ElGato`> wow zhivago has become a regular
<ElGato`> AND IT'S ALL BECAUSE OF ME :d
<liar> yup
<ElGato`> :D RATHER
<liar> eh?
<ElGato`> i invited him here from #c on efnet
<ElGato`> so um
<ElGato`> how's brix?
<liar> its not
<liar> i fscked up bfe last night adding some error checking code and need to fix that
<liar> plus i upgraded to the newest version of bochs and then found out that kevin has changed the output format for the debugger
<ElGato`> i haven't done anything on lengua since i started school and im still waiting for retro to be multitasking so my code can work :\
<liar> so i get to rewrite chunks of bfe to make it work with newer versions of bochs
<ElGato`> bummer
<ElGato`> my c coding skills suck so i gotta test each function i write individually
<ElGato`> for some reason i cannot bring myself to an interactive programming environment
<liar> hey u use winsuck98
09:40pm
<ElGato`> i have a winmodem
<ElGato`> i have $0
<ElGato`> ph33r m3
<liar> i took my asus p2bd out of my win98 box and put the p2b in it and ofcourse it had to reinstall all the drivers cuz it sucks, but now it wont power down. it just sits on the now safe to turn off screen
<ElGato`> uh huh
<ElGato`> and you expect me to know what to do?
<liar> hahah, i just shutdown and that "now safe to turn off" screen had a gpf
<liar> win98 gpf's even after it has stopped running
<ElGato`> heh
-:- hcf [nef@me-portland-us305.javanet.com] has joined #tunes
<ElGato`> they are in cahoots with intel
<liar> eh?
<ElGato`> cause i don't see how you'd get a gp when nothing is accessing any memory
<ElGato`> hi hcf
<ElGato`> you know the arrow paper talks about a homoiconic system...
<ElGato`> but aren't all programming languages homoiconic really
<ElGato`> they all have a set of primatives
<hcf> hi
<Zhivago> re
<ElGato`> HI ZHIVAGO :D
<ElGato`> HOW ARE YOU?!
<Zhivago> that version of bfe I dl'd last had some impossible cod ein it
<Zhivago> elgato: I'm being shouted at
<ElGato`> :D
<ElGato`> seriously though
<ElGato`> aren't all programming systems homoiconic?
<Zhivago> no
<ElGato`> don't they all give you a set of primatives to abstract?
<Zhivago> el: not all languages hav a set of primitives
<ElGato`> example?
<Zhivago> well, let me clarify
<Zhivago> there are langauges which allow the reflective construction and destruction of primitives
<Zhivago> in which case there isn't 'a set of primitives' as such
<Zhivago> just the current primitives
<Zhivago> and in this case the distinction between a primitive and a non-primitive is eroded to almost nothing
<ElGato`> yes but non the less they are primatives that exist in the language
<Zhivago> what does it mean to be a primitive?
09:50pm
<ElGato`> how does one construct a primative without another primative though?
<Zhivago> this is the key of it :)
<ElGato`> something that cannot be defined with other objects
<Zhivago> so, if we have primitives which can be constructed and replaced from the language, are they primitives?
<ElGato`> no
<Zhivago> this is what I mean by erosion
<Zhivago> ok, then some lisp dialects might be considered primitiveless in this conception
<ElGato`> but what are they defined from?
<Zhivago> we usually define them underneath the language from within the language
<ElGato`> but the language still needs to start from something
<Zhivago> in a reflective system the starting point isn't very important
<ElGato`> like could you have a language who has a primative '+' that you could redefine?
<ElGato`> would that be what you're talking about?
<Zhivago> yes
<Zhivago> or the function primitive
<ElGato`> that's what i want
<Zhivago> this language is called lisp :)
<Zhivago> but there are other reflective languages around
<ElGato`> um
<Zhivago> lisp is just the simplest/best-known
<ElGato`> i don't want to use another language
<ElGato`> i want my own :D
<Zhivago> well, again I'd look towrad lisp, which is a family of languages :0
<Zhivago> the base of which is very very simple
<Zhivago> which is what I've done
<Zhivago> and the regallocator is getting debugged :)
<ElGato`> but non the less when you first get one of these compilers it starts out with a small set of primatives, no?
<Zhivago> soon I'll be able to port all that nasty retro asm to lisp
<ElGato`> heh
<Zhivago> sure, but you can replace them afterward
<ElGato`> there
<ElGato`> so it starts out homoiconic then
<Zhivago> does 'homoiconic' and 'start out' go together?
<ElGato`> :P
<Zhivago> everything in the universe is 'homoiconic' if you take a temporal snapshot
<ElGato`> hmmm
<ElGato`> then homoiconic would mean a fixed set of primatives?
<Zhivago> I assume so
<ElGato`> ok
<Zhivago> from context, although I haven't seen the term before
<ElGato`> i lean towards the stack based forth and poplog like languages
<Zhivago> why the stack?
<ElGato`> minimizes syntax
<ElGato`> it's powerful
<Zhivago> do you know why forth compilers are rare?
<Zhivago> its the stack's fault
<ElGato`> eh?
<ElGato`> funny i never have actually seen a forth compiler hehehe
<Zhivago> the problem with forth is that it is bound to a particular implementation
<Zhivago> right, and you won't see forth compilers really.
<Zhivago> not enough abstraction
<ElGato`> but why?
<Zhivago> becuase of the silly stack fetish
<ElGato`> i like the stack concept
<Zhivago> all this collapsing of stacked values
<Zhivago> stacks are ok under certain hardware
<Zhivago> but why cut off your knees?
10:00pm
<ElGato`> damn you zhivago you're convinxing me to change my ideas :(
<Zhivago> that's my job
<Zhivago> my goal is to destroy the world
<ElGato`> you haven't seen how many times my ideas have changed
<Zhivago> heh, I cvan guess
<ElGato`> hmmmmm
<Zhivago> been down that path before
<ElGato`> i don't really like the rstack idea
<Zhivago> what idea do you like then?
<ElGato`> but i sure do like the perameter stack
<ElGato`> (bad spelling)
<ElGato`> it minimizes syntax
<ElGato`> syntactical sugar es muy malo
<Zhivago> syntax is easy
<Zhivago> (to minimise)
<ElGato`> bah
<ElGato`> oh
<Zhivago> to do forth, write lisp code, but bacwkard and without the ()'s
<ElGato`> so?
<Zhivago> so its easy to translate from forth to lisp :)
<ElGato`> i like in forth how each word is in context but still independent
<Zhivago> not sure what that means
<ElGato`> think about it some
<ElGato`> it is nearing the time for me to go to bed
<Zhivago> well, its largely meaningless
<ElGato`> nah
<ElGato`> its a thesis and antithesis :D
<Zhivago> or a blind assertion without defined terms
<ElGato`> :\
<ElGato`> no
<ElGato`> well it has been fun chatting 
<Zhivago> what does it mean for a word to be independent?
<ElGato`> bye :D
-:- SignOff ElGato`: #TUNES (Leaving)
10:10pm
-:- SignOff hcf: #TUNES (Leaving)
<liar> Zhivago: what problem did u have with bfe?
<Zhivago> well, there was a completley bogus fscanf() in there
<Zhivago> updates.c: 290
<Zhivago> fscanf(readpipe, "%*s %*s 0x%x", &stack_value)
<Zhivago> I mean, come on :)
<Zhivago> you shouldn't use using fscanf() either
<Zhivago> it has lousy error recovery
10:30pm
<liar> uhh that fscanf reads the stack value
<liar> its not bogus
<Zhivago> that's true
<Zhivago> its still bad code, and crashes though
<liar> what version of bochs?
<Zhivago> it doesn't matter :) bfe is the crasher, but its the september release b
<Zhivago> 990907b
<liar> bfe doesnt work with new versions of bochs
<liar> kevin changed the output format
<Zhivago> heh
<liar> i just upgraded to 990923a and im fixing the problems
<liar> he was oh so kind by removing the physical address from the output so now i get to generate it myself
<Zhivago> in any case, I would advise using fgets() to get a line
<Zhivago> and using sscanf() after that point
<Zhivago> scanf() and fscanf() aren't wise choices, they have lousy error handling
<liar> does fgets have the same syntax as fscanf?
<Zhivago> no :)
<Zhivago> fgets() reads in a line from a FILE *
<Zhivago> which limits errors to that line, with the scanf's you'll slip
<Zhivago> fgets(buf, sizeof(buf), fp);
<liar> yer welcome to make those changes :)
<Zhivago> if you know the max line len, then this is safe
<Zhivago> its not worth it to me
<liar> its not worth it to me either
<Zhivago> yes, stability rarely is :)
<liar> im not writing this thing to make money or to make others happy
<liar> im writing it so i can have a halfass ok tool to help debug brix
<Zhivago> ok, then it suits that purpose :)
<Zhivago> I might submit a patch to bochs
<Zhivago> get them to change the output format, and the way the debugger goes
<liar> heh
<Zhivago> its all they need really
10:40pm
<Zhivago> they just need to have a sensible query format
<liar> ya
10:50pm
<liar> Zhivago: u here?
-:- SignOff liar: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep)
<Zhivago> well, yes, but you aren't.
11:30pm
-:- BlackPhoenix [black-phoe@adm.univd.kharkov.ua] has joined #tunes
-:- SignOff BlackPhoenix: #TUNES (Leaving)
[msg(TUNES)] newlog 1999.0924
IRC log ended Fri Sep 24 00:00:01 1999