IRC log started Sun Dec 26 00:00:01 1999 [msg(TUNES)] permlog 1999.1226 -:- NetSplit: devlin.openprojects.net split from sterling.openprojects.net [12:08am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [devlin.openprojects.net] -:- Netjoined: devlin.openprojects.net sterling.openprojects.net -:- Fufie [stig@tunnel-44-14.vpn.uib.no] has joined #tunes -:- SignOff hcf: #TUNES (Leaving) -:- hcf [nef@me-portland-us222.javanet.com] has joined #tunes -:- _ruiner_ [DIY@ppp068.wi.centurytel.net] has joined #tunes Hey _ruiner_! -:- SignOff hcf: #TUNES (Leaving) 01:30am -:- SignOff _ruiner_: #TUNES (destroy what destroys you) -:- SignOff lar1: #TUNES (Leaving) -:- Guadeloup [Guadeloupe@cr250960-a.poco1.bc.wave.home.com] has joined #tunes -:- Guadeloup [Guadeloupe@cr250960-a.poco1.bc.wave.home.com] has left #tunes [] -:- Fare [fare@s195.paris-119.cybercable.fr] has joined #tunes -:- simoz [sasas@a-mn5-93.tin.it] has joined #tunes -:- SignOff simoz: #TUNES (Leaving) -:- Fare has changed the topic on channel #tunes to: Reflective Computing System -:- SignOff eihrul: #TUNES (Ping timeout for eihrul[usr5-ppp162.lvdi.net]) -:- FareWell [fare@s195.paris-119.cybercable.fr] has joined #tunes -:- eihrul [lee@usr5-ppp162.lvdi.net] has joined #tunes -:- SignOff eihrul: #TUNES ([x]chat) * Fufie/#tunes yanws.. ya lispmas morning Fufmann! -:- FareWell [fare@s195.paris-119.cybercable.fr] has left #tunes [] morning :) 04:30am -:- SignOff Fare: #TUNES (Leaving) -:- SignOff air: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep) -:- SignOff Fufie: #TUNES (Leaving) -:- Fufie [stig@tunnel-44-14.vpn.uib.no] has joined #tunes set: there are several decent tools for that 09:20am -:- hcf [nef@me-portland-us139.javanet.com] has joined #tunes -:- AlonzoTG [Alonzo@216-164-131-150.s150.tnt2.lnhva.md.dialup.rcn.com] has joined #tunes -:- mibin [mibin@62.11.103.94] has joined #tunes -:- Downix [down@d-gnaps-11.ici.net] has joined #tunes -:- SignOff mibin: #TUNES (Ping timeout for mibin[62.11.103.94]) hmm.. mails on tunes@tunes cool wow! =P 11:20am -:- hcf has changed the topic on channel #tunes to: TUNES, Free Reflective Computing System || Towards an Open System hmm -:- eihrul [lee@usr5-ppp155.lvdi.net] has joined #tunes hey eihrul lo 11:40am what's up? nada mucho just playing with ocaml finally Same * Downix/#tunes found out why Clementine didn't work earlier. hehe tis pretty nice why? the new disk was gzipped you got a new disk? :( * eihrul/#tunes spent all last night making an OpenGL/sockets/etc chess hack. :) Same one you did, silly got it from start to buggily-working in a few hours :) Hehehe Ok Well, we'll see how it goes but, it really disgusted me how unstandardized some C++ facilities are that are part of the fscking standard and that all the supposed portability is just imaginary I know 11:50am eih: portability? Fufie: like i said, it doesn't exist what kind of portability do you mean? OS or Compiler portability? or atleast in the context of two different vendors implementing a single standard more like environment portability which is more of a melding of the two as they're both important factors we gots to gets that punk that posted earlier today in here... I want to discuss libertarianism with him. =) uh huh eih: the std was finished last year? eih: remember how long the simpler ansi c took to get working so when will it be standard? 10 years from now? when i won't need the language anymore? it will be a usable standard in about five year I guess c++ as of 92 is pretty much ok to write though * Downix/#tunes nods (ok as in compatible with compilers, and not 'ok' as in something nice :) lol * Downix/#tunes is talking OS's with a bunch of Amiga-freaks That is a good laugh everytime downix: I thought you were the resident amiga-freak ;) Fufie: I am, but the problem is I actually know how HW works, and more about OS's than they do Fufie: it is only okay so far as you stay in the encapsulated world provided by GCC Fufie: venture out, and you face the consequences Fufie: I keep hearing them go "USe QNX, it runs Amiga programs!" or "Amiga IS the OS!" which is a major laugh amiga is dead :) eih: not really.. see how much the c++ part of gcc changes.. I have a lot of code I need to go through to make it 2.96 compatible.. it was last tried with 2.7.2.3 eihrul: Amiga is dead, the concept is alive, but the Amiga itself is dead 12:00pm Fufie: not much Fufie: all they've done is made it a tad more strict lately eih: a lot of the rtti and template stuff and libg++ has changed one or two more errors get spat out, nothing big and C++ rtti is nearly useless.... rtti? Fufie: what would be nice is if you could have type signatures for methods that included the children of polymorphic classes defined in a virtual method abi: rtti is run time type information, a c++ hack to allow access to the vtable ok Fufie: and the C++ compiler could easily generate stubs that dynamic_cast'd for you or threw exceptions but otherwise C++ rtti is useless eih: anything related to c++ compilers are non-trivial and never ever easy * eihrul/#tunes shrugs. (my experience from doing work on c++ front-ends) there's a lot of room for improvement in all features of C++ like constrained parametric types C++ never made complete sence to me to be honest a better language for creating parametric types (which is part of the above) better handling of class namespaces * Downix/#tunes nods perhaps a real module system some day? :) references and pointers are very non-orthogonal concepts etc etc why bother with fixing a dead horse instead of choosing another language? Fufie: i have Fufie: I'm learning about C LISP still, not quite confident in myself for it yet i'm still learning it though :) downix: it takes time.. it's a very different paradigm Fufie: have been spending whole weekend reading through ocaml manual Fufie: Well, so far I'm learning a lot more about it, and I am very impressed eih: hehe.. Fufie: i looked at haskell as well... while being a prettier language, the compilers seemed kinda ad-hoc and old though i really like Hugs Fufie: I am thinking of using C LISP as the standard language on my system, or at least porting it to my system. (Would be very nice to use) downix: when you say 'C LISP', do you mean common lisp or the common lisp implementation CLISP? eih: haskell is nice, it fits on a disk as well :) Fufie: Common LISP ahh Fufie: not the specific implimentation ahh.. it is just so uncommon for lisper's to caps-lock the name :) that's the job of the system ;) OIC C Lisp then? Fufie: yes, but it doesn't seem to have the nice compiler like ocaml :) Fufie: i've been pondering a haskell -> ocaml translater though :) downix: CL is the easiest Fufie: Ok * Downix/#tunes knows he needs to figure out how to make a simulator/emulator tho eih: translators can be quite messy 12:10pm Fufie: yes, but it's the only way i'm going to get Haskell with an OCaml compiler :) its not as messy as a whole compiler or even as messy as writing a front-end for gcc :) there are front-ends for gcc? yes wow and i think haskell and ocaml are similar enough to easily translate from the first to the latter ocaml is just a tad bit more specific and imperative haven't really tried ocaml Fufie: Is there any place I could study up on sound assembly codes? don't know k How the hell do all of these programmers make drivers then, hehe 8) I don't know, I don't make a sound, even if I am in the forest and there is noone listening hehe -:- Kaufmann [Kaufmann@dial149.infolink.com.br] has joined #tunes FOO! Awww, how cute... an EnchantedForest URL 12:20pm hey -:- SignOff Kaufmann: #TUNES (Ping timeout for Kaufmann[dial149.infolink.com.br]) he forgot the someone's lost in there analogy... 12:30pm Fufie: hmm, ocaml is eerily like c++ crossed with haskell in a way * eihrul/#tunes sighs. 12:40pm y0 that guy who posted to tunes@tunes.org earlier today lives in australia! =((((( poor d00d... that's why his page was so strongly anti-government. 01:10pm our government isn't perfect either nope 01:20pm -:- ultima [ultima@user-38lcn61.dialup.mindspring.com] has joined #tunes -:- Fagui [fagui@s195.paris-119.cybercable.fr] has joined #tunes -:- Fagui is now known as Fare Gakuk! hey -:- Fare [fagui@s195.paris-119.cybercable.fr] has left #tunes [] odd that was deep where is fare anyway? fare is, like, sometimes looking at another screen and interruptible with a beep (/msg Fare ^G) or at http://www.tunes.org/~fare/ or connected through a crappy ISP (cybercable.fr) or pronounced Fahree abi: i said where! eihrul keep in mind abi was originally designed as in infobot for storing URL's...it probably translates what into where =) gah .../whois abi she says she's a borg i'll trust her to know what where means! but...abi's an infobot... its just imagery but with a big brain! hmm tril is logged into bespin not idle 01:40pm -:- SignOff hcf: #TUNES (Leaving) -:- hcf [nef@me-portland-us906.javanet.com] has joined #tunes * ultima/#tunes rapidly toggles hcf's bus lines y0ink -:- SignOff ultima: #TUNES (Read error to ultima[user-38lcn61.dialup.mindspring.com]: Connection reset by peer) there is a windstorm in france... 01:50pm -:- SignOff Downix: #TUNES (BitchX: the quilted quicker picker upper) !sterling.openprojects.net!! Remote CONNECT jordan.openprojects.net 8000 from is -:- binEng [Anders@d212-151-56-239.swipnet.se] has joined #tunes -:- SignOff binEng: #TUNES (Ping timeout for binEng[d212-151-56-239.swipnet.se]) -:- mibin [mibin@a-ii2-55.tin.it] has joined #tunes -:- _ruiner_ [DIY@ppp445.wi.centurytel.net] has joined #tunes -:- SignOff _ruiner_: #TUNES (destroy what destroys you) om -:- SignOff abi: #TUNES (Ping timeout for abi[bespin.dhs.org]) 03:10pm -:- abi [nef@bespin.dhs.org] has joined #tunes HELP!!!!! =\ EyE KaHnT KoDe. :) Fufie: sigh, ocaml seems way less clean than haskell the need for pragmatic solution often ruins initial elegance but ocaml is overly pragmatic :) it reminds me of C++ this is not good haskell has a very pragmatic type system yet it is way more elegant go CL then, it has still elegance even though it is practical but does it have good compiler systems? :) yes implementations ranging from free to very expensive well, free would be nice :) being the unemployed student that i am a student could use ACL which is normally the very expensive one, but is gratis for students.. www.franz.com (it's gratis for linux and fbsd) ACL? i think ACL is Allegro Common LISP, free of charge proprietary software at http://www.franz.com 03:20pm Fufie: but i'm looking to do system programming in it :) Fufie: in which case an open implementation is better than trying to hack something closed into a system CMUCL may be what you want, but it is not too good for learning the language.. common lisp systems are however much more open than other language systems as they are pretty reflective and you can change the implementation yourself as it is dynamic sheesh, why does choosing a language have to be so confusing? :( ie you don't need access to source code as desperately as you do in c/c++/java/etc !is:*! anyone know what's up with jordan ? its like trying to decide your preferred method of execution * eihrul/#tunes sighs. execution? being shot would be nice sure, but either way you end up dead but it's less messy and painful than going through a lawnmower yes, but in 1 minute you won't remember any of it :) 03:30pm no matter which way you go you're dead you can take the pretty language you can take the ugly language there's always some little con! but an ugly language as c++ has many ugly parts and cons.. elegant meta languages don't suffer that badly ocaml doesn't look any better in this respect though too many special purpose primitives all cluttered together oh doesn't sound too nice 03:40pm perhaps i haven't look at the language closely enough though... its just nto giving me good vibes HELP!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! * AlonzoTG/#tunes hurls a laser guided flaming woodchuck at eihrul * AlonzoTG/#tunes works Fufie over with a 2x4 * AlonzoTG/#tunes slaps mibin around a bit with a large trout. yes love, what is your problem? :) -:- SignOff mibin: #TUNES (Read error to mibin[a-ii2-55.tin.it]: EOF from client) I need help with this file. change the ';' in the main() argument list to ',' Fufie: how large is the core lisp language? I need to 1 get the lengths of a list of files; (bailing out if they don't exist) Fufie: neglecting standard libraries made *from it* int main(int argc; char *argv[];) should be int main(int argc, char *argv[]) { ... } okay fixed... -:- mibin [mibin@a-ii2-55.tin.it] has joined #tunes The problem I am having is with this: alonzo: as for length of files, which OS do you have? D0S. for Unix-like OSes it isn't too hard.. see in the comp.lang.c for examples on how to use stat DOS.. ugh you don't need stat you can just use fseek () to SEEK_END and ftell () umm yes.. but using stat() is recommended for speed and lesser memory-use sure, but for a one time run program i don't think its an issue fseek/ftell is a butt-ass ugly solution when I don't want to open the files (yet) cuz I hafto open them again later in the prog... al: if it's DOS, check the msdos-programmer faq.. I think there was a simpler solution Fufie: well, what about my lisp question? :) and Wintendo has a system call for it 03:50pm eih: which lisp-question? Fufie: how large is the core lisp language? Fufie: neglecting standard libraries made *from it* eih: the core of common lisp is pretty small what primitives are provided? fuck this. It is T00 UGLY!!!! primitives as in? primitive types? -:- AlTeRnAtIvE [AlTeRnAtIv@ppp67.greco.com.br] has joined #tunes hi all I have to fopen, fseek, ftell, fclose in that order to just getfilelength. =\ alonzo: use a proper un*x-system then AlonzoTG: which compiler? BC++ 4.51 Fufie: things that the language provides that are not expressed in the language itself eih: it is somewhat hard to answer as one could manage with a really small set of operators and types and implement the rest out of it, but the rest wouldn't be fast enough Fufie: okay, so what about for the fast implementation? :) ie the object system can easily be written in simple common lisp, but the object system is usually given a native implementation and really tuned AlTeRnAtIvE: ask the channel eih: as for fast I have managed to get to c++ level roughly but it took work eih: the more experienced lispers talk about c/fortran speed I'm looking for a free C compiler. Can anyone help me ? (i just have de Quick C - that sucks..) alt: which OS? gcc? hmmm... gcc is crap. glibc 1 and 2 are crap. * Fufie/#tunes looks at abi, mildly amused haha 04:00pm -:- SignOff AlTeRnAtIvE: #TUNES (Leaving) *SOB* it looks like /me will have to write his own flaming library if he wants this done reasonably efficiently. =\ Fufie: i suppose i'll just wait for my lisp books to arrive then and hold off on ocaml for now Fufie: its primitives were truly so... they could not be easily expressed with the language if one desired Lets assume I can get file lengths within the next few months of computation, How can I coppy files QUICKLY in C? al: system("cp foo bar"); or use mmap() AlonzoTG: http://www.snippets.org 04:10pm -:- Kaufmann [Kaufmann@dial524.infolink.com.br] has joined #tunes * AlonzoTG/#tunes smashes Kaufmann with an Iron Butterfly * Kaufmann/#tunes blocks Al's strike with a Led Zeppelin Al, why do I want this file? * Kaufmann/#tunes pokes Al Al, is this program supposed to do anything, at all? I need help with it. cuz I can't code. =\ ah what do you need help with? I need to verify the existance of and get the size of a list of files... 04:30pm ah, that's all? do you need to do this in Perl, tho? s/Perl/C/ It's damn easy in Perl al: do it in common lisp ;) FILE-LENGTH and PROBE-FILE :) Fufie: what if the environment doesn't have files? :P open OUTFILE, '>'.shift(@ARGV); for (@ARGV) { print OUTFILE (-e && -s) }; close OUTFILE; exit; eih: if the environment doesn't have files, what point would thwere be in "verify the existance of and get the size of a list of files..." Fufie: al's problem :P obviously LOL my solution, however, is short and elegant can anyone suggest a good hexadecimal magic number? (mapcar #'(lambda (x) (if (probe-file x) (file-length x) nil)) lst) 32 bits in length preferrably? Gemme a lisp interp and I'll do it. that is reasonably unused? 00 and FF are always good choices 04:40pm ah, 32 bits The program needs to create a table of file offsets and then save that table as well as the files into one big file. 00000000 or FFFFFFFF but aren't they common? :) or atleast easily mistakeable om =\ 0xB0B 00000000 is exactly as common as any other 32-bit hex number... maybe less so in format-specific binary data, and more so in ASCII 0xDEADBEEF okay fine i'll go with deadbeef just because i like it and even if lotta people probably use it :) 0xDEADC0P heh that's 28 bits hmm.. P should be included in hex * AlonzoTG/#tunes pondering paying one of you $25 to finish this program... Alonzo, couldn't you be a bit more specific about your program? 0xAFFEC7ED well it has to create a file containing a number of input files similar to copy a+b+c d but it has to generate a list of file offsets at the beginning of the file. good enough :) need sleep.. goodnight.. afk Funny link: http://www.nada.kth.se/~asa/Mage/deadbeef Al, what do you mean by offsets? How far down the file does your next file begin? =\ hmm. I just came up with a grotesque workaround. How do you coppy a set of files into a gnu file? 04:50pm interest in this conversation seems to have drifted away :P Fuck this, I'm beginning to design my file format around my shitty language (C) =\ maybe I'll hire KIETH. LOL I've already stripped it down as much as I can. =\ 05:00pm All I want to do is find the length of a bunch of files and then coppy them into a new file... -:- mibin [mibin@a-ii2-55.tin.it] has left #tunes [] There's probably a Unix shell script for that. =\ ah easy to do huh? !NickServ:*! HyrliK used GETPASS on arpwatch Okay how do I coppy them into the new file easily? AlonzoTG: ur dos shell can probly do it forget DOS... I want it to be minimally portable. dos can't do stuff like that anyway. open OUTFILE, '>'.shift(@ARGV); for (@ARGV) { -e && open FILE, $_; print OUTFILE -s, ; close FILE; }; close OUTFILE; exit; does that save a list of file offsets at the beginning of the file? hmm... I think I have a workable solution... lemme write that now.... -:- SignOff Kaufmann: #TUNES (Ping timeout for Kaufmann[dial524.infolink.com.br]) 05:10pm =\ yelp! Gelp! Kelp! Lelp! Nelp! Jelp! Velp! .elp! oelp! ;elp! zelp! relp! pelp! eelp! -:- Fare [fare@s195.paris-119.cybercable.fr] has joined #tunes Gakuk! * AlonzoTG/#tunes greets "Salutations Fare!" :-) lo howz da weafer? weather Stormy yep, its on the n00ze in the US trees are unrooted 26 dead in France due to the wind. 05:20pm damages to the house. less than 10 years ago, tho. I am batteling C to try to force it to generate one of MY programs rather than just another C program... hum. How that? I need to get it to coppy binaries from .bin files and create a .img file... .img? what I'm trying to do in DOS is: copy a+b+c d but I also need the length information of each file. hehe. Put it in each files' header! get any existing OS, to see how they do it tww.... hmm.... just get NASM to do it.... now that is supprisingly elegant! 100% manual.... fantasmagorically bugphilic, but doable! MaHn ThAt Is UgHlIe.... If this works, I might even write a kernel for Tunes! =) thanks, fare! =) ATG: I once did it w/ as86+m4 (ouch) seen the Tunes 0.0.0.25 stuff? 05:30pm Fare: i read through ocaml manual Fare: seemed as if ocaml was almost C++ all over again... even though it was functional in some respects, it just have a monolithic amount of primitives uh???? eihrul: what have you been smoking? muscle relaxants * eihrul/#tunes shrugs. ocaml, monolithic? It has maybe the most advanced module systemaround. as for "functional in some respects", that's laughable. well, not in all respects it's as functional as can be (if you except Haskell or Clean, perhaps) what do you call mutable variables? functional in some of its respects i should say ok, it has some kind of (somewhat crude) object system (that ain't that much used, since it has crude aspects) eihrul: 'a ref yes, crude is what springs to mind eihrul: most people don't use it eihrul: they use the functional part of the language, without bothering with objects where can I grab a gnu coppy of NASM? there was a nice talk by xleroy at ICFP about modules vs objects Fare: then what's the purpose of using ocaml if not for the objects? over say, caml? eihrul: caml has been discontinued. Work goes on in ocaml. ocaml has the advanced module system the "o" is for buzzword bragging (which doesn't mean that ocaml is bad as far as OO goes) not exactly good either (OO is such a crap; the Object system in ocaml may be crude, it's much better than most OO languages) (or at least, much cleaner) 05:40pm okay, i'll make sure to ignore the object system then i was just wondering if i was the only one disgusted by it eihrul: it's not worse than C++ and it has its uses, too that isn't saying much i don't like C++'s object system too much but it's definitely not the strong point of the language THE FUCKING NASM HOMEPAGE IS DOWN!!! =(((((((((((((((((((((((( (although it's infinitely better than what you have in, say, SML: nothing) -:- hcf_ [nef@me-portland-us121.javanet.com] has joined #tunes h c f! -:- SignOff hcf: #TUNES (Leaving) -:- hcf_ is now known as hcf Is there a mirror NASM site? Fare?\ the One True Nasm site is 100% nonresponsive. =((((( I need the latest version!!!!! * AlonzoTG/#tunes smashes hcf with an Iron Butterfly * AlonzoTG/#tunes hurls a laser guided flaming woodchuck at hcf * AlonzoTG/#tunes slaps hcf around a bit with a large trout. * AlonzoTG/#tunes works hcf over with a 2x4 AlonzoTG: why must u act like such a moron? cuz eye yam. Now tell me where I can get NASM 0.98 . i would tell u, but ur above actions disqualify u =\ If you tell me I'll leave! for good? I'll leave untill I have a loadable/executable version of my OS. =P http://www.kernel.org/pub/software/devel/nasm/binaries/ 05:50pm -:- colorg [r@humbubba.smart.net] has joined #tunes blerf cg Hi 06:00pm Fare, does eforth use the x86 stack as the parameter stack? -:- Downix [down@d-gnaps-44.ici.net] has joined #tunes hey-lo all hi What's up? not much. I've been trying to kill the same bug for 3 days. I'm going to have to give it a name or something. hehehehe What kind of bug? In a forth-like thing, in the equivalent of EKECUTE. Nasty stuff. EXECUTE rather Oh man Call it Herman OK * Downix/#tunes saw the original bug once they photographed it? Stuck in a relay? It's in a computer museum ah, eye to eye a Moth got stuck in ENIAC This one is self-inflicted, I changed a fundamental stack from an array of pointers to void to an array of ints, and am now patching the damage. Hence, why the term is "bug" cg: no, it uses x86 stack as return stack okay I'm fixed up. =) cg: it uses ebp as data stack pointer -:- SignOff AlonzoTG: #TUNES (Have Nice Day :)) Fare oh. So much for my grasp of the code :o) 06:10pm wait been a long time since I looked at the code... well after all, yes, it uses %esp as data stack pointer neat did you see Muench posted to clf? yup also seen Jeff Fox's commentary on hForth hmmm, I missed that I think -:- lambda [lambda@cogeco-88-83.cgocable.net] has joined #tunes 06:20pm -:- AlTeRnAtIvE [AlTeRnAtIv@ppp155.greco.com.br] has joined #tunes i'm looking for a Visual C++ Programmer. Is there anyone here ?? Ewwwww * Downix/#tunes hates that M$ crap why Downix ? Just case it's from M$ ? AlTeRnAtIvE: No, more like I dislike it's stylings and the fact that it's not 100% C++ complient -:- mibin [mibin@62.11.102.196] has joined #tunes oh... AlTeRnAtIvE: try efnet/c++ AlTeRnAtIvE: They make standards for a reason where can i find this efnet/c++ ? uhm, on efnet ahah Alternative try: /server irc.megsinet.net then /join #C++ hehe ok. i'll try later What time is now on USA ? 9:30 for me abi: time? time is 18:35:33 hehe. here it's 00:35 ! abi, time is 18:35:31 ...but time is $time PST... here 3:35 where are u from mibin ? is that important, alternative? =p hummm... just that.... aahh... forget it. nahh =) i am from italy cool why that nick? cause i like diferent things. "Thing Diferent" or "Be Diferent" (MacOS and BeOS) 06:40pm wow, cool * Downix/#tunes likes some serious "Alternatives" Serious? Hey Alternative, give Amiga a try sometime A have an Amiga 500 !! and a MSX too !! Fare: Well, Mac and Be are not as alternative as a system that's been effectively undead for 8 years AlTeRnAtIvE: Cool. * colorg/#tunes couldn't be non-alternative if he tried * Downix/#tunes has an Amiga 1000, Amiga 500, 1200, CD-32 wow ! Negotiating to buy a 2000 I love the Basic language... i used to program on my MSX a lot... * Downix/#tunes has a 3000 a friend of his lets him use * Downix/#tunes shrugs a have a litle problem on my amiga... the disk with the Work Bench is gone... now i can't use it anymore. AlTeRnAtIvE: Then buy a new one. AlTeRnAtIvE: Hell, upgrade your OS, $95 for OS 3.5 but. Where ? Here In Brazil ?? No way... Even in Brazil order through National Amiga out of Canada, they ship worldwide well, i can try. Ok http://www.nationalamiga.com thanks. There are many upgrades to the A500 cool Add memory, HD, faster CPU, etc. i have to buy a new one (A500 is too old...) Nah, the A500 is nice The newer models are slower hehe. realy ? AlTeRnAtIvE: After the A3000, COmmodore fired all of their good engineers The end result is a series of bad designs oh man.. * Downix/#tunes hates his A1200, it is such a POS My A500 and A1000 are much nicer and more expandible not as much as a Compaq Portable Fare: Any COmpaq. hehe 06:50pm -:- hcf has changed the topic on channel #tunes to: TUNES, Free Reflective Computing System || Towards an Open System || Amiga lovefest what did the good engineers become? some went to 3D0 Fare: Dave Haynie (father of the Zorro 3 and AAA) now works for Met@box, the rest ended up in 3DO 3DO actually Except for the one engineer I snagged. 8) Wich OS do u use Downix ? my IRC nick is from colorga_pd in Aminet mus/misc, check it out AlTeRnAtIvE: On my Amiga? AmigaOS 3.0 or 3.1 AlTeRnAtIvE: Otherwise, a mixture of Linux, BSD, and a few smaller ones and on PC ? BSD ? What's that ? AlTeRnAtIvE: BSD == Berkley UNIX AlTeRnAtIvE: University of California, Berkley produced a varient of UNIX, called BSD. Berkeley Software Distribution colorg: *nods* cool. I tried to use Unix-like systens (Linux SlackWare, RH, and even WinLinux) but that's a litle hard for me. colorga_pd will blow your mind AlTeRnAtIvE: You used Linux, admittedly, all UNIX's are difficult for a novice, other than AmigaOS which is also a UNIX-like OS (In fact the first UNIX derivitive to be given the Sys V approval) -:- Fare_ [fare@s195.paris-119.cybercable.fr] has joined #tunes -:- Fare_ is now known as FareWell -:- FareWell [fare@s195.paris-119.cybercable.fr] has left #tunes [] hehe. i'm just a newbie... AlTeRnAtIvE: it's ok. i tried to use linux, but i couldn't conect to the internet ! What is a Linux without internet ?!? AlTeRnAtIvE: Setting up Linux for internet is either hard or easy, depending on your distribution Well... I tried in averyone (: - SlackWare, RH, OpenLinux... AlTeRnAtIvE: Depends on the release as well. RH for instance is innane damn i cant get ppp in win95 instead AlTeRnAtIvE: Try Debian 07:00pm while in debian i can get it easily AlTeRnAtIvE: Debian ships with KDE I believe, use the kppp setup utility, and bam, your all set -:- air [brand@p0wer.qzx.com] has joined #tunes Hey air i personally use pppconfig in debian, pretty easy to set up i tried the kppp, but that shit says that my modem is busy ! AlTeRnAtIvE: it may be Downix: hey AlTeRnAtIvE: DO you have a WinModem or a real modem? air: What's up? know i'm tring under WinLinux, (with the KDE) A real one use dcon to dial up how can i do it ? AlTeRnAtIvE: I mean the kind of modem I know. An Acura 33.6 AlTeRnAtIvE: It may not be a real modem you know. i have another one here. I'll try the another one. AlTeRnAtIvE: Ok it's all about Scheme But KDE sucks... well, better that nothing... Gnome dialer is good too WIMP sucks as is XISP cLIeNUX, my distro, comes with pppsetup, which works OK. * Downix/#tunes nods I set up my ppp manually myself * Downix/#tunes is too used to that bit over my head. Doing a distro myself I try to concentrate on non-net stuff. colorg: Same here cLIeNUX itself is client-use only. No servers. colorg: not a bad idea I do a distro as an idea-sink :o) I try to make cLIeNUX Forth-influenced also, but so far the Forth influence is subtle. 07:10pm now if I could kill off Herman then I'd be off to the races cg: describe Herman? Fare my bug of the week, I decided since he's part of my life I'd name him :o) oh hehehe Herman, right? Fare cLIeNUX is much simpler than larger distros, and much more rigorous about dependancies. There's a 28 meg Core, and it can build itself without C++ ok does it have emacs? it's smaller than emacs :o) I SAID does it have emacs? Downix : Do u know an OS called FreeBeOS ? No no. Nor does it have vi. In true Forth style it has Pico and cLIeNUX ed. i person told my that is better that BeOS... Do u know an OS called FreeVMS ? a person told my that is better than BeOS... Pico + cLIeNUX ed is smaller, and IMO much better, than vi. or Free360 ? Fare: I do about FreeVMS nope colorg: anything remotely similar to emacs? may I do about a 5-line paste? colorg: eh? colorg : go on lambda: GForth. I consider emacs to be a front-end to a Lisp, not an editor. cLIeNUX Core guarantees a Forth. the closest I get to forth is RPL on my HP49 very similar actually a marriage of Forth and Lisp kinda 07:20pm does anyone here works with DirectX ? I'm a lisp hacker by trade HP28 rocks! hmm AlTeRnAtIvE: I prefer using a real system for games, sorry lambda: neat. cLIeNUX currently defines a core. The rest of the world is welcome if it doesn't break the core, and is honest baout it's dependancies. Downix : Well, i was thinking about learn VC++ to use DirectX into games... AlTeRnAtIvE: Use a more generic system for games, like OpenGL AlTeRnAtIvE: That way you can port easier well I'm coding a lisp compiler so I can get away from it all Scheme rather Herman is a bug in my 3-stack Forth AlTeRnAtIvE: ID stuck it out by not using DirectX, and look at their popularity in the "non-Windows" market. Quake and Quake 2 is on every box, from Windows to mac to Linux to BSD to Amiga.... Downnix : In which language i'm able to use OpenGL ? Is this going to run in a computer with out a 3D Video Card ? lambda: hey join us! we need a lisp compiler for a scheme dialect, too AlTeRnAtIvE: OpenGL is a software standard, not a hardware one. okay if I ever get anything done I'm planning on using CPS to virtual registers and then optimizing the bloody hell out of it AlTeRnAtIvE: C is an all-around decent language sure (or equivalently A-forms) and then performing spillage very simple will probably work better on risc machines though Downix : hehe. C rulez.. AlTeRnAtIvE: Nah, C is not perfect, but is is good to work with 07:30pm I think the downfall of C is that it is nothing more than a macro assembler C++ is even worse lambda: Agreed. high-level abstractions in a macro assembler C produces code that performs well one normal CPUs portably. Otherwise it's nasty. a good lisp compiler can beat gcc ian wrote ccg in c++ colorg: *shrugs* in real-world applications lambda: I've found that out if anyone ports it to scheme or cl ... stupid benchmarks mean little C is a *contrived* macro-assembler lambda: benchmarks mean nada I think cl is a relic of the past as well I think the gabriel benchmarks are a good test of some things yes, but it's a real prog language, unlke scheme Scheme's a real programming language Herman is a result of me trying to re-write a C program without pointer declarations lambda: Never used Scheme. Learning LISP (and am VERY impressed) Scheme is more elegant than cl I don't think anyone denies this lambda: I think that's true even cl users * Downix/#tunes has never used Scheme depends on "elegance" Scheme has elegant aspects, but also has braindamage some aspects of CL are more elegant braindamage? Scheme macros are braindamaged Scheme lack of (begin) is braindamaged. Fare: in what way? ahh yea, i have never understood them dood Scheme has (begin) no, it hasn't is that like awk BEGIN {} ? only (begin foo bar ...) what is (begin) used for? colorg: no, like LISP progn http://www.swiss.ai.mit.edu/~jaffer/r5rs_6.html#IDX95 mibin: macro expansion yes it does unless you mean something somewhat different Scheme has something called (begin) at the least in Scheme, (begin) cannot have 0 arguments s/arguments/in-body forms/ that's braindamage what's the point of begin having 0 arguments? in Scheme, standard macros introduce yet another completely alien sublanguage. in Scheme, you have no module system at all etc, etc ah, gotta love those alien sub-languages so that's sugar sugar sugar sugar Forth is the Few-Level Language. cg: alien sub-languages are nice, when you define them from a clean general language; not as part of the primaryspecification Fare: I was being sarcastic lambda: syntactic sugar causes cancer of the semi-colon quite quotable CL is loaded with more crap than anything else and Scheme's macros are quite nice I don't mean that Scheme hasn't good things but it's 1) far from perfect 2) not quite uniformly better than CL 07:40pm I actually usually use a subset of Scheme - do and the standard I/O usually do is ugly -:- AlonzoTG [Alonzo@216-164-135-244.s625.tnt4.lnhva.md.dialup.rcn.com] has joined #tunes om Scheme macros would have been much better done as a standard library over real macros, and using a real pattern matching language that would also be a standard extension of the language hmm I need some ICE to debug my OS. Anyt good macros language>? do is horrible; standard I/O almost non-existent the standard has been purged in the past R3RS->R4RS t, nil, and last-pair were eliminated why eliminate last-pair? beats me it's probably a bit redundant why keep anything at all, then, besides lambda and function application? (see unlambda) abi: unlambda? hmmm... unlambda is a silly language based on call-by-value combinatory logic at http://www.eleves.ens.fr:8080/home/madore/programs/unlambda/ Scheme is really just a core yes and no it doesn't try to provide anything beyond basic I/O it's an incomplete and burdened core. even as a core language, there's a mismatch oh, Fare, : was that your note in the eforth source about how to do ?key ? cg: no cg: where in the source? uh, is there a canonical list of novelty linux distributions? like cLIeNUX not afaik; maybe around linux.org or something (or in dmoz.org) my bad, I was thinking of pForth maybe lambda: there's pages for smaller distros yeah that's what I mean essentially and there are several small ones in metalab.unc.edu/pub/Linux/distributions 07:50pm http://dmoz.org/Computers/Software/Operating_Systems/Linux/Distributions/Tiny_Linux/ freshmeat has a few theres a distro howto that may have some if its uptodate at all the distro HOWTO is "available on CD" only oh fun colorg: wrong metalab docs/HOWTO/Dist* lambda: you know #lambda on ircnet? cg: I don't see clienux there it's well hidden under "umsdos" stuff -:- SignOff mibin: #TUNES (Ping timeout for mibin[62.11.102.196]) 08:00pm nope Fare lambda: yes, you do, now bwhaah so what do you think about my idea of a Scheme compiler? I must have missed part of it. or was it that general? heh well generally I can elaborate on my technique well, in Tunes, we do need a scheme compiler to base our stuff upon the nice thing about CPS is that you don't have return conventions CPS is good (or necessarily calling conventions) well, you have 1) call conventions 2) linearity analysis you can eliminate calling conventions for internal procedures with optimization you need a way to produce detailed description of ad-hoc calling conventions oh you can force calling conventions if you have to 08:10pm but sticking to them will result in ineffcient code there's no real point in obeying them you must *encode* them, so that you know what conventions to use when linking to an optimized extern module hmmm, esr did more changing than he told me he would then esr? no if you're compiling something to be an extern module you can use standard calling conventions that sucks I want separate (and incremental!) compilation standard calling conventions suck Eric Raymond, self-appointed mount of open-source unix mount/mount mouth even cg: what did he change in what piece of code? I'll take up your idea though I'm talking about the Distributions-HOWTO I prefer RMS as a fearless leader hehe he's got the shtick down schtick to be more correct perhaps :P Fare: so if I am dynamic linking to a new version of a library Fare: I have to lookup its calling conventions Fare: and dynamically recompile myself? r0x :P hmm colorg: oh which doesn't have the string "cLIeNUX" in it that I can see lambda: yup or compile stubs, or whatever 08:20pm this whole technique however will suck on x86 compared to MIPS or PPC -:- lar1 [larman@2Cust42.tnt20.sfo3.da.uu.net] has joined #tunes Anyone know the interupt for CMOS setup? lambda: I lean for MIPS over PPC hcf: where did you see cLIeNUX? metalab.unc.edu/docs/HOWTO is 404 colorg: add /pub/Linux to that path s/add/prefix/ why suck on x86? x86 has somewhat fewer registers it's a CISC machine the mere fact that you have stacks/heaps/registers fit to your app help a lot register optimization is futile for it so the spilling code will have to be optimized hmm Registers does not mean performance, but the right use of them can help (spilling is when your bignum number of virtual registers doesn't fit into the real number) of course you optimize so the most-needed registers don't get spilled * Downix/#tunes nods Right Right now I'm looking more at how to optimize a sound system than CPU however this is fucked. I will have mail esr again it seems hmm 08:30pm hcf which HOWTO did you cLIeNUX in? It's not in Distributions colorg: i didnt say it was s/you/you see/ colorg: u said the distro howto was only available on cd lambda: with few registers, the distribution of registers for heaps and stacks is important hcf no, I said it was only about distros available on CD (also use of global variables as registers) colorg: ok, whatever, n/m which, despite the BS about submissions, it is -:- SignOff AlTeRnAtIvE: #TUNES (Leaving) Ok everyone, does Quake 3 Arena need GLIDE to work? -:- water [water@tnt-10-10.tscnet.net] has joined #tunes big crowd tonight Yup who's lambda? lambda is a syntactical constructor for abstraction of variables. See http://tunes.org/Review/VMs.html#lambda heh who's Downix? you are probably doing it right now hehe lol i read irc logs the other night from the ship Oh? noticed ppl making up funny os-name acronyms acronyms? water: hehe, I remember that on the 25th they need to settle ESR vs. RMS with an MC battle lambda: no, a "Celebrity Deathmatch!" Fare: with few registers, you're required to have better spilling algorithsm err algorithms lambda: do i know you? lambda: Right after the "Linus Torvalds vs. Bill Gates" Fare: which speeds up execution on platforms with more registers water: probably not k Downix: nah a DJ battle [RMS] lambda: Hehehe "I'll cons your ass five times before you hit the stack" btw, jim carrey as andy kaufmann is hilarious lol lol galileo vs torquemada heh "then I hit my gat and blow you back" lambda: chosing the heap/stack convention is part of the spilling algorithm lifo implementation is not terribly hard and it's also not hard to determine which registers have to go (actually, I'm joking about using a lifo) 08:40pm hey all, there's a lisp hacker on board my ship (to those i haven't already told) lisp pirate bignum silver! he's trying to make another better lisp it seems no just a better compiler kinda he hasn't given me the full details on it though, so it's hard to say for sure but it is (will be) gnu s/gnu/gpl of course teh weird thing is that he hand't heard of tunes before he met me a month ago (bleh... typos) well TUNES is known as the archetypal vapour (no offence) lambda: Agreed yeah, i stayed well away for about 4 years but then i got sucked in :) * Downix/#tunes is getting sucked in at this rate no! :) well, if someone (hint, hint) would just decide to describe exactly what kind of lisp would be needed, it wouldn't be such a lot of vapour water: Big hint heh unfortunately, i'm just not the lisp/scheme guru that i should be i get lost after seeing about 200 parentheses ;) 08:50pm water: will 200 arrows be any better? * colorg/#tunes is approaching 200 parentheses in C. Not good. hcf: hopefully hopefully, an infinite number of arrows will be much better I'm writing an R5RS compiler that's it anyone else can modify it heh. i've got r5rs on my handheld who's david vennik? Scheme on an HP49 that was a bad idea I had hehe heh :P gnight I'll come here often k I'm going to start writing tomorrow ok #) -:- SignOff lambda: #TUNES ((lambda (x) (exit x)) irc) * Downix/#tunes has been writing up ideas for his system for the past few hours which system? the os for your product? No, my product itself Working on implimentation ideas ah I've concentrated on it's openness now it's time to map out the implimentations we'll offer 8) k * Downix/#tunes so far has a list of 12 i could really use some mentoring on developing my arrow spec I wish I could helpo so do i :) I need a hardware spec to look over 09:00pm my format now is useful only to me * Downix/#tunes needs an industry standard one to show people yeah, that makes sense The only one I have is for the Amiga And I don't know if that's standard or not 15 page summary of it -:- SignOff lar1: #TUNES (Ping timeout for lar1[2Cust42.tnt20.sfo3.da.uu.net]) I think I just got Herman's number. whose? my bug of 3 days * Downix/#tunes needs spec sheets evil little fucker hehe cg's bug is named Herman oh Once something is named, it's easier to isolate we can only hope yup nope, didn't get him. dang Herman, get out of there! what if Herman turns out to be a family of ethereal ghost-bugs? hehe water: he's got kin, no doubt heh hehe GOT HIM! well, a piece of him... cool 09:10pm I changed a pointer declaration to an int, and changed all the uses of it to explicit casts, except in a comparison. ok * colorg/#tunes celebrates with 3 aspirin lol * water/#tunes tries to imagine this sort of error occurring in smalltalk or lisp nope. :) nope om :b or assembly, or BASIC. This is "Fun With Writing a Machine Spec in C" hehe colorg: Done that. not fun no idding -:- SignOff eihrul: #TUNES (Read error to eihrul[usr5-ppp155.lvdi.net]: EOF from client) kidding, even this is already legend for ugliness of code -:- eihrul [lee@usr5-ppp155.lvdi.net] has joined #tunes * colorg/#tunes goes looking for Herman's relatives * eihrul/#tunes thinks he has Herman's grand father. hehe 09:20pm /* contentsof.. |<-&HNC->| + |<--- offset to CFA --->| */ ip = (int *) (ps[psi] + ((*(byte *)ps[psi] + 3) & 0xfc) + 8) ; * Downix/#tunes is installing a new distro of Linux what flavor? * Fare/#tunes is scoring 564650 at xmris Enoch * Fare/#tunes is going 3053 -:- SignOff Fare: #TUNES (:Connection reset by pear) Downix: and that's about what? what's 3053? Sleep mode on The Lisp Machine ? colorg: RT performance for things like audio/video work ah, comes with an rtlinux kernel? colorg: yes, plus ReiserFS and some SMP enhancements k00l B and C I can do without, but I ship an rtlinux kernel myself * Downix/#tunes nods rtlinux is a good kernel it's a great hack, best thing for Linux in a while I think reiserFS includes journaling, right? Yes k. didn't remember for sure ok water I think there's 3 journalling filesystems for Linux now in various stages. ext3 works, Reiser, and XFS I think We're looking at adding ext3 yeah, i read about it on /. a few days ago * Downix/#tunes nods there, networked opengl chess lives... 09:30pm eihrul: cool. Downix: as an option I hope * water/#tunes recalls Fare not getting the hint about Tunes earlier AGAIN. Yes Right now I'm studying the various patches out there Looking at adding in USB for the system om 09:40pm hey Alonzo Downix: are you an Enoch developer? yep, * AlonzoTG/#tunes is about to Leap into bed. colorg: yes, i designed the kernel * Downix/#tunes leaps into Al's bed with him. j/k "designed" meaning what? colorg: Well, modified the kernel a bit for better RT performance, added in the patches, etc. I've mentioned my DSFH before, eh? * Downix/#tunes nods that's my big kernel hack :o) Now it looks like our head is thinking of going with BSD, which throws a monkey wrench into my plans 8) why would s/he do that? * Downix/#tunes shrugs, don't ask me I'd like to have a "B" kernel for cLIeNUX, but Linux is it for main * Downix/#tunes nods the problem is that kernels just don't cut it many times HURD's MACH kernel is not impressive IMHO yeah, just the idea of a microkernel is not a sell, by itself om * Downix/#tunes nods the kernel is irrelevant. Its how the SYSTEM is put togeather that counts. true My system will look like a sphere. Downix: hurd is not a microkernel lol anyway... a sphere? a sphere is Jim Little's Sphere project, at http://www.teleport.com/~sphere 09:50pm eihrul: Last I heard they were going microkernel Downix: and they're trying to get rid of mach anywho eihrul: good I took over the sphere project. hurd is merely the OS on top of whatever microkernel they choose on a whim oh -:- SignOff AlonzoTG: #TUNES (Have Nice Day :)) Took over? Um... the Sphere page was updated only a few days ago heh hmm, this is a lot more detailed about it than Alonzo's explination darn it, i had a paper that jdl/atg could have used it's the only phd thesis i've ever seen with a formal idea of "metamodelling" ok don't worry about it you know, i still haven't figured out why jdl's project is called "sphere" Me either or Alonzo's 10:00pm as far as i'm concerned, atg doesn't have it water: I hate to say it, but I feel the same unfortunately, it seems the only "active" participant in tunes right now is me * Downix/#tunes nods I'm being drawn in slowly it seems and i've been wrestling with arrow ideas for too long * Downix/#tunes is not an OS-guy tho well, it's your own doing, then ;) TO me an OS is a way to harness hardware i'm not either. a language guy * Downix/#tunes nods To you the OS is a way to run your software er... i'm in the smalltalk camp * Downix/#tunes nods, ok I'm in the "Amiga-camp" I like my hardware to be well designed hehe yeah, that's cool Of course I like total performance and giving the programmer as much control as possible, well, compiler more to the point sure, i just don't want to have to deal with c/asm to get that control * Downix/#tunes nods between a rock and a hard place, of course I know but make the hardware well enough, and the compiler can turn code into something extraordinary well, good night yep night -:- water [water@tnt-10-10.tscnet.net] has left #tunes [] hmmm, I wonder if that's one of Herman's nephews lol aaaaaaaahhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh 10:10pm 3 4 5 + + .x 0000000c hehe in the interpreter baby! :;cp 0.9.m4 attic/sans_Herman -:- _ruiner_ [DIY@ppp426.wi.centurytel.net] has joined #tunes hoy _ruiner_ <_ruiner_> greetings hoy Ruiner <_ruiner_> salutations downix <_ruiner_> ' 10:20pm So what's up? <_ruiner_> just got home <_ruiner_> oops signal:noise ok 10:30pm damnit, i just got opposing force yesterday and i already beat it oh fun Quake 2 was no challenge to me, then I learned about quakematches 8) And player mods * Downix/#tunes now Quakematches as a Spider Dalek 8) or as Duke Nukem whats a quake match? air: Deathmatches in Quake or Quake 2 (and Quake 3 now) uhh every game has deathmatch Not every But a good percentage do i like the bot matches in unreal I had never done it before Quake 2 tho Never played Unreal <_ruiner_> quake3 is kinda tough in dm unreal is kewl <_ruiner_> very * Downix/#tunes wants to play Unreal Unreal for Linux is supposed to be out soon if not already I want Duke Nukem Forever Unreal engine w/ a few "special" mods 8) unreal single player sucks just as much as quake single player * Downix/#tunes nods Multi-player, that is the keu key but unreal bot/death match is very kewl Well, DUke Nukem Forever will rock single player hl and hl:of are good single player games hl impressed me * Downix/#tunes tried it yesterday 10:40pm DN:F will be nice, a real 3D engine where the game has, ya know, a PLOT! And most of all, the Duke! half/fully naked women won't hurt either. 8) :) Shake it Baby! Hey, he's running around Vegas, tons o' dancing girls They don't make game that that everyday 10:50pm * Downix/#tunes talks to a compiler designer about weither or not HLL's can get the performance of C or ASM He's saying that C is better it's a matter of what machine model is implemented. Forth runs like shit on "C-machines" and vice-versa hmm there's a fundamental "impedance mis-match between one-stackers and 2-stackers. They are different machines. "impedance mis-match" hmm 11:00pm What would my machine be classified as, a "Fourth" or "C" type? so unfortunately people think Forth is a fast interpreter. It's a slow emulator. * Downix/#tunes nods most CPUs have one stack supported by the instruction set. Forth engines have two, and the whole stack is usually on-chip Hmm x86 has push and pop for a return stack in memory * Downix/#tunes nods push and pop are implicit to other instructions on a Forth engine, and don't hit memory so Forth on a Forth engine crushes C but nobody is spending 200 mil on a Forth engine Well, what kind of engine would I have then? in what? On Eddas Oh wait, did I tell you about that? no http://www.cs.cmu.edu/%7Ekoopman/stack_computers http://www.cs.cmu.edu/~koopman/stack_computers/contents.html Eddas is my computer design. Similar concept to the Amiga, just re-implimented for modern computing needs. (Namely keeping as much off of the CPU as possible through accelerated co-processors, from everything to memory functions to video to I/O calls, thereby leaving the CPU with less overhead) -:- SignOff hcf: #TUNES (Leaving) -:- hcf [nef@me-portland-us121.javanet.com] has joined #tunes Downix: my hardware ideas are at http://linux01.gwdg.de/~rhohen/fire8.htm colorg: Well, how far along are you? hmm Downix: 0 on the hardware. The program I'm writing is a demo of the CPU arch * Downix/#tunes nods We're not quite the same thing 11:10pm * Downix/#tunes focused on memory control and acceleration over compiler systems and such Using the 68020 tho is a good choice, cost effective We're using the MIPS arch. ourselves Downix: look at the PPC QUICC, 7 IO ports and copocessor colorg: Not quite the same kind of system as Eddas it's the same IO as the 68360 with a PPC * Downix/#tunes nods Very impressive But, does not answer my q, would something akin to Eddas be a C or Fourth-style system? I focused on giving the compiler a key strength in the system a MIPS is a one-stacker, all the big stuff is * Downix/#tunes nods, alright So a C-esque style system you are saying? in FIRE the idea was to use a standard device like the 68360 as the scheduler CPU for a bunch of Forth-like "task-processors" Uh huh not a dissimilar idea there, but in our case with dedicated co-processors for the majority of the work not SMP, task-switched CPUs on one main was of real SRAM * Downix/#tunes nods not SMP, task-switched CPUs on one main wad of real SRAM not SMP in our situation either I'm not a fan of SMP, but "DWA-tasking" has multi-media written all over it fuck, DMA-tasking * Downix/#tunes nods multi-media is what we're doing that Koopman book is real lucid about what is a one-stack machine, 2-stack, etc. 11:20pm Which one? well, there's probably FPGA's now that you can get 16 or 64 Forth engines onto one device. http://www.cs.cmu.edu/~koopman/stack_computers/contents.html control that FPGA with a regular CPU, and you have a whole CPU for each main medium your're processing. colorg: FPGA's are not quite capable yet. and you task-switch in one memory cycle too expensive yet Downix: and kinda slow, yes? colorg: Agreed. Sticking to hardware that's smartly planned will gain you better performance and many of the benefits Downix: well, an FPGA might be good for demo purposes colorg: We're using FPGA's for our initial prototypes, to debug but for final production, hardware nod FPGA's will never be capable of keeping up with hardware a modern FPGA may be as fast as hardware from 20-yeard or even 10-years ago but the advancement in hardware will always keep up and surpass it that's the nature of things, specific will always outperform generic sure. But it would be the same as anything as far as task-switch latency. And you can program the hardware :oP * Downix/#tunes nods 11:30pm Fastest FPGA out there is 200Mhz, costs $1800 dollars to do the same thing in hardware, $16 The other thing I'd look real close at if I had any hardware toys is a 2048 bit data bus <_ruiner_> a 2k data bus? H3sm has 8 to 2048 bit ints _ruiner_: sure. hehe <_ruiner_> what frequency? * Downix/#tunes is going with a 128-bit data bus heh, SRAM speed 133Mhz yeah <_ruiner_> you gotta gimme a complimentary demo piece to play with hehe <_ruiner_> ;-p soon as someone drops 10 mil on me <_ruiner_> lemme get out my checkbook.... do <_ruiner_> if I make it big I'll be sure to look into your project <_ruiner_> which means "don't count on me" well, the non-vapor part is the distro and H3sm man, noone ever offers to help Eddas like that a bit less pie-in-the-sky, but not much :o) <_ruiner_> I can make a solid offer for you.... <_ruiner_> but its dependant on me getting off my butt and doing something * Downix/#tunes sighs <_ruiner_> if I have my os written already for the x86 architecture, I'll code whatever you want for your project -:- SignOff air: #TUNES (Ping timeout for air[p0wer.qzx.com]) 11:40pm * colorg/#tunes breaks his ip after 3 days of trying to fix his pointer stack pointer * Downix/#tunes yawns I feel like I"m an outcast here because I'm designing a C-type rather than a FOURTH-type computer * Downix/#tunes pouts :;H3sm Illegal instruction never managed that one in C before lol <_ruiner_> downix: I thought you were an outcast here because you were designing a computer..regardless of the type <_ruiner_> ;-p _ruiner_: True 11:50pm Am I the only person here who is not designing an OS? <_ruiner_> I believe so....unless you count the tunes team.....they don't seem to be doing anything at all hhehe cool, now it's a mere segfault PROGRESS! _ruiner_: There's a tunes team? <_ruiner_> yes Where? * Downix/#tunes has seen Water, but that's about it <_ruiner_> fare, water....and some other jabrones I can't think of right now Ok <_ruiner_> it should be up on the tunes site somewhere Well, seen and known Water and I have talked about Arrow before * Downix/#tunes is looking forward to Tunes, even if it is not even close to the OS type he wants for his system <_ruiner_> I don't think it has much promise Arrow or Tunes? <_ruiner_> both <_ruiner_> what type do you want for your system? _ruiner_: Something concentrating on performance, expandibility and being a part of the system <_ruiner_> heh....have you tried to con air into porting brix to it? <_ruiner_> sounds like his would be a nice fit air's already said he'd be willing to <_ruiner_> excellent so has Core with his Clementine And Alonzo, but we won't go there <_ruiner_> like I said before....if I have mine written for the x86 architecture, I'll port it * Downix/#tunes nods <_ruiner_> how about lar? lar? <_ruiner_> lar1 is writing RavenOS [msg(TUNES)] newlog 1999.1227 IRC log ended Mon Dec 27 00:00:01 1999