IRC log started Sat Jun 5 00:00:00 1999 [msg(TUNES)] permlog 1999.0605 -:- ChanServ has changed the topic on channel #Tunes to: www.tunes.org -:- mode/#Tunes [-o Fare] by ChanServ -lilo(lilo@varley.openprojects.net)- [GlobalNotice] Hi all. You may need to reidentify yourself; services was just restarted. -:- jae [jae@pool03.ka.ilk.de] has joined #tunes -:- jae [jae@pool03.ka.ilk.de] has left #tunes [] -:- cotton [cotton@carp.emsnet.de] has joined #tunes Hi -:- SignOff cotton: #TUNES (Ping timeout for cotton[carp.emsnet.de]) 06:30am -:- AlonzoTG [Alonzo@209-122-204-188.s569.tnt4.lnh.md.dialup.rcn.com] has joined #tunes -:- HickServ [thrustit@209-68-229-50.dialup.cust.tfb.com] has joined #tunes top of the morning to you om howsit going good alonzo? 07:00am okay.. just chillin out yeah me too i just woke up five minutes ago been trying to figure out nasm :( i think i will write a machine code editor for very low level stuff I took the documentation and had it printed.... then I read the entire thing... After that programming in NASM was a breeze. =) 07:10am you can't do out's and in's though how do you do port io? hmmz shouldn't be hard... just RTFM!!!! i did al look it says in's and out's aren't accepted huh ? use inb or outb... or inw or out w outw =P ok Lengua will be able to act as an assemler too so i just gotta make Lengua in nasm 07:20am =) cause i can jjust do forth stuff mov %ax %dx or whatever mabye ax and dx and all the others can be words too yeah whatever make compile time faster and in a couple days i get my intel manuals :))) 07:30am -:- SignOff AlonzoTG: #TUNES (Ping timeout for AlonzoTG[209-122-204-188.s569.tnt4.lnh.md.dialup.rcn.com]) -:- AlonzoTG [Alonzo@216-164-225-2.s256.tnt6.lnh.md.dialup.rcn.com] has joined #tunes hum hi fare :) lo * HickServ/#tunes *HUG* HS: como estas? muy bien y tu? grrrr wtf is foo and bar?? dict foo install dict at home... 08:10am -:- smkl [sami@MLXVIII.rdyn.saunalahti.fi] has joined #tunes oh god how this mailing list is annoying all they do is bicker what mailing list? 1068! which ml? tuens@tunes.org tunes oh feel free to enhance the level... om tell me how to access cvs read collaboration.html#cvs or is it #CVS? * HickServ/#tunes starts to cry cause fare never tells him anything directly -:- tcn [tcn@cci-209150250110.clarityconnect.net] has joined #tunes 08:20am hey tcn tell me how to access cvs tee-see-en! did you download it? hi fare no HickServ: apt-get install cvs where? you use debian linux? no im on winblowz ie it isn't my comp ftp.debian.org and my linux comp doesn't have a modem brb so what? so what is probably my problem get a couple old 10mbps ethernet cards.. use coax 'cause you don't need a hub just put the modem on linux, and use linux as a router for your win machine... first he needs a real modem ;) * Fare/#Tunes bought a hub (two, actually) Fare: almost done with my PS/GC/MM design.. would you say GC should be integrated with PS? I'd like to see it when you're done... tcn: sure we do want the PS to be GCed I'm gonna write it up in a file Page-granular GC in supervisor mode, Byte-granular GC in user mode ? uh??? what's this stuff mix-granular GC is interesting, but tends to cumulate costs... different for large & small objects we only need the page-grained GC now ok * Fare/#Tunes doesn't remember what the "blue" color was... * tcn/#tunes doesn't either (in nomotiongc) so will you have a read barrier or a wrwrite barrier? (or both?) or just a dirty-bit handling don't know yet.. i'm still working on the GC design we might also want flushing or the RAM for persistence being done incrementally 08:30am perhaps we could combine that with GC? yeah you mean, flush pages to disk that have the dirty bit set? uh, that must be done eventually also, at regular intervals (or in ways matching the GC?) we must dump objects to disk I was wondering about that.. guess I answered my question the momemnt a GC cycle is done looks like perfect to dump since it's also the moment memory footprint is smallest yeah, GC can hand 'sync' s/hand/handle/ and the dirty bits serve as both GC and PS flush we'd better have a generational GC BACK well i wanna lookat retro :( with large disk objects being ultimately left untouched in the "old" generation hs: get the zip.. it's on my page problem: small objects like to be together with header. large objects (particularly BLOBs) like to be separated from headers. get the .tar.bz2 snapshot heh then we can see if the snapshot is working ok... I can always check that this is pretty cool i found a way to optimise a header in linux HS: uh? Fare: I was gonna keep small objects separate from their headers (all headers are the same size, anyway) tcn: also, we MUST have a protocol for multi-heaps what's a multi-heap? look at /usr/include/asm/irq.h 08:40am then go to SAVE_ALL instead of pushl all those registers you could just pushad yeah, I know did you know before? or better yet is that a bad thing? you already told me.. and I know asm anyway.. I just don't care! oh i must email this to someone who can i show to get it changed? maybe the author thought "well, 7 push's execute faster than 1 pusha".. of course, those 7 pushes are taking up 7 times as much space in the cache.. how are alll the others faster? execution time is faster (on the 386/486, anyway) but it's one instruction opposed to 7 it still has to wait for 8 dwords to be sent across the bus to RAM but atleast you are getting a little more performance out of it that will help 1) RT vs transient vs PS 2) multiuser 3) have untodfgdlfkg7~jdflg7~kjsdfs Grrr! My ISP is *very* unreliable now who can i email? to get it changed Fare: wow.. line noise no: my repeatedly typing the "any" key while the line is down... hehe 3) limit the need to scan otherwise stable subheaps when GCing the disk anyone? 08:50am i.e. imagine a program with its own address space, that has small and well-delimited interface to the rest of the system if you can keep GC in one separated from GC in the other, you win big especially if one of the programs is already swapped out to disk of course, you might still want a monthly global GC hehe for instance, the day everyone is off (and at hours when the sysadm is sleeping) ie, defrag (although the former may seem incompatible assumptions) yeah, mucho like defrag I was gonna program the OS to do that when it's idle that could be done incrementally, right? there's a special notion of "idle" we should keep: without *interactive* processes active. if someone is doing intensive batch computations, it shouldn't prevent GC. right c'mon do any of you know who do i email? hs: first found out if it was done that way for a reason. Ask on a linux channel, not here! i am and no one will answer me try a different one well after four it doesn't seem to be working out #debian or #redhat, maybe how about posting linux deveplopment mailing list? 09:00am -:- hcf [nef@me-portland-us818.javanet.com] has joined #tunes ok HS: you post on the lkml and await from comments HS: see the FAQ of the lkml on www.tux.org/lkml or such HS: hint: you're not the first one to suggest a pushad/popad no? I've seen the thing discussed in the past. actually, depending on the processor, pushad may be slower or buggy depending also on cache behavior, I presume oh, maybe I shouldn't use it in retro :) tcn: it's still time to optimize that later :) for instance, on PPro's, it's slower, IIRC I don't use it in Forth routines, anyway hrm :( i thought i had a good idea -:- water [water@ppp-tnt-139.tscnet.net] has joined #tunes hola hey brian hi it's not a bad idea i just got home from work where's work? the ship at the pier hehe you should just make tests on a variety of processors, and determine the right instruction depending on the processor, then use the linux processor option to put the right one :) what's the good idea? fixing the warp core? warp core??? jk i don't see how it could be slower though for every CPU, there's a list of how long each instruction takes. water: HS wanted to save registers in linux with a pusha instead of 7 pushes.. oh, ok tcn: but you cannot trust it, because it all depends on pipelining and cacheing pushad fare HS: in at&t syntax, it's pushal :) they are pushl's btw, i'm downloading some papers on doing hw spec's in Haskell. 09:10am yeah * Fare/#Tunes is listening to the classical music radio station -- so great! water: there's been a CPU proven within a LISP reasoning system (ACL2) fare: cool this project that i'm looking at goes into superscalar and non-traditional designs. y0 Howz arrows? never had the time to see in those things mahler's 5th symphony's adagio adapted for voices only -- GREAT -:- jazz [jazz@has-pm2-038.inetnebr.com] has joined #tunes the arrow papers are progressing well, but the squeak implementation will have to be re-written, unless i can find a lazy object evaluator for squeak # Appears as TIKI. water: lazy evaluation can be easily implemented with closures! water: you could abandon squeak :) (#G010E210M3) is anyone here? yeah (#G210E910M3) do you know about BRiX (#G710E;10M1) ? squeak has better programming support, i.e. a gui with apps. jazz: that's _QZ's project (#G010E@10M1) have you ever used it? jazz: he'll be on later (#G210EC10M1) do you know when? tonight (#G710E210M1) thanks. bye -:- jazz [jazz@has-pm2-038.inetnebr.com] has left #tunes [] yup.. later (define-macro (delay body) (let ((foo (gensym)) `(let ((,foo '())) (lambda () (if (null? ,foo) (set! ,foo (list ,body))) (car ,foo)))) (watchout for parentheses) fare: you'll have to explain that. (define-macro (force body) `((,body))) water: see delay and force in the R5RS i have r4rs. is it there as well? yes (I think) ok hrm i don't like nasm too much i might be able to implement the same thing in squeak. but there isn't any documentation for as86 HS: what don't you like about nasm? as86 is basically no more developed 09:20am afaik, nasm is the best there is (i know that's not saying much, but it works fine for me..) i don't know plus the last release by the author had bugs in 32-bit mode (I had patches) i just happen to like tasm i like tasm format HS: I used to run TASM from within dosemu early version of TUNES used that heh dosemu 0.48!!!!! anyone remembers that one? yup * HickServ/#tunes can't wait till he gets his intel manuals * tcn/#tunes ftp'd the old 386 manual from somewhere * water/#tunes keeps his microprocessor manuals on Zip disk. * HickServ/#tunes wants to write an os in microcode i'd be insane to keep a real physical library of all the books i have! * Fare/#Tunes was disappointed from his intel manual, and won't buy one anymore fare: why? the i386 architecture is a fucking bloated bitch I'm sick of it buy? they are free fare: too true. HS: the paper copy is cheaper than printing one s/one/one yourself/ why print it at all? -:- core [core@core.suntech.fr] has joined #tunes hi core people hi hi Brian :) yes water: I bought it at a time when pdf's didn't exist... you can order them for free fare: oh. salut, manu Brian: don't know if Fare told you, a friend of mine with much better brains that mine actually had ideas similar to Arrows and understands your concept :) salut, Fare :) core: really, can i meet this person? bbiaf hey core! water: if you come to France... water: if you touch French coasts sometime yes, otherwise it's yohm@worldnet.fr be back in a Fuck? yup my women are calling me fare: il t'a repondu au fait (je l'ai vu hier) oui je lui ai rerepondu core: ok, thanks. water: he actually agrees with your ideas too. he wants to implement intelligent agents through a system of that kind and make them socialise :) fare: ok :) core: wow. he's actually working out the consequences? at 1:30 AM we were having a discussion about the ugliness of DCOM implementation and how reflectivity and syntax trees will make something nice :) water: well, you put more thought in it that he did so far, but he's trying to reach an actual goal with it to see if it's feasible core: cool. 09:30am between you, Fare and him i'm actually starting to understand how that kind of system can exist I've been thinking about this sort of thing too (a "genetic algorithm" thing).. yeah, the arrow system looks like a good way to try it. hi Tom :) yeah, everyone seems to agree on Arrow * water/#tunes can't believe what he's reading! my friend was talking about valued graphs but developing the same idea basically. water: I didn't say it would be useful.. just interesting tcn: so, can i start retro, pull the plug and restore it to the same state when it reboots yet? ;) core: once I translate what I have on paper now tcn: cool! tcn: what about the things i wouldn't want to be persistent? tcn: i'd be annoyed if bad changes i did to an object while toying with it stay forever (or can only be cancelled through some obscure way) * tcn/#tunes is back * core/#tunes didn't know he was gone in the first place ;) * AlonzoTG/#tunes needs to write prototype * HickServ/#tunes ditto alonzo Anybody here going to write a paper for the prototyping symposium? * core/#tunes needs to stop writing code for a few hours at least. fare: i still haven't commited the 3 bazillion changes, it'll come all at once :) commenting is an artform i should start writing a Tunes-for-normal-people or Arrows-for-normal-people in my spare time too with my understanding of it ;) gotta return video... bbiab core: if you want to 'play around', you make a copy of the object.. or save differences.. tcn: well, sometimes you accidentally (or not) activate a config option that haunts you forever core: yeah.. I'm aware of that concern.. Common sense will guide me :) i need to eat breakfast bbiaf hs: same here.. hehe tcn: saving differences so you can rollback the object state could be useful.. tcn: otherwise saving the current state and the last user-accepted state tcn: so you can restart an object in its current state but always offer to revert to the user-accepted state, i don't know differences is easy in some programs (like CAD) persistent storage is more evil than no state storage at all if it's misimplemented :) 09:40am if the system crashes, it's gotta be careful not to restart what caused the crash that too. how are you going to avoid this? restoring the last user-validated state in this case? that's if you can have market that the current state has crashed it only kernel code is likely to cause a total crash well, even for a non total crash, i'd be peeved if application Y will not start anymore because the current state causes it to die :) s/the/its/ oh just asking :) that's where versioning is nice yeah.. so you need to know one way or other what last version was good :) *the* last version? finally something compiles in nasm out of the many you can rollback to although i don't know how you will know that, since the object could initialize successfully and crash later how 'bout I worry about that when I get there ;) or at least closer.. sure.. it's just strongly influencing the design of your persistent store, like if you only store the current state, or a current and a last validated by user, or many.. :) i haven't started on that yet because (well, i miss storage proper too, heh) i am not decided on those things :) well, this sounds good: save the entire current version, and diffs to the old versions how will you rollback to an older one though? let the user do so, do it automatically? (that can be decided when you get there, granted) uh hickserv: nasm doesn't care about things like model small :) tcn: updated nasm 0.98 patches for xcom are on the site bw s/bw/btw/ (recompiled clementine with it, works well) how do you make an asm program that accepts command line arguaments? -:- iStormy [stormy@rain.futuresouth.com] has joined #Tunes 09:50am hi hi storm hickserv: if you're writing a main() function under linux, just get argc and argv off the stack hi iStormy asm not c i know how in c but not asm water: what's your opinion on preserving object states? hickserv: no difference. mov eax,[esp+4] will get argc (provided you didn't push things on the stack inbetween). core: is nasm 0.98 final out? tcn: yep, released yesterday by HPA ok, i'll grab it tcn: not many changes tcn: besides the fact they split the doc off into another tar, and that they don't build txt docs by default anymore, you need to make -C doc core: well, i think that the management of consistency of states is an important issue not addressed yet. the arrow system should maintain multiple states at any given time. core: i.e. to Arrow, all contexts exist that it understands. * tcn/#tunes is away (lunch) core: the user (or a user-defined heuristic) chooses the context to implement (particularly important for crashes). water: yes.. although i don't know how you can define 'consistency' in Arrow (what is consistent to it and what isn't) water: yes, i would also think that the user is the best person to know what context to implement ack, for once that i can catch you online i have to go :/ will be back in a little bit :) -:- SignOff core: #TUNES (this is really not the default quit message) core: consistency is consistency to a particular ontology (or theory or language): a set of facts and interpretations of the world that are consistent. the arrow system should handle many of these, and talk about their interactions. darn! whats html style fixity called? -:- SignOff water: #TUNES (Read error to water[ppp-tnt-139.tscnet.net]: Connection reset by peer) damnit can argc be a string? -:- water [water@ppp-tnt-140.tscnet.net] has joined #tunes 10:00am i want to be able to go "lengua whatever" would whatever be argc? whew! i'm back. water: whats html style ( ... ) fixity called, do u know? well, it's the push/pop style, so that it just operates on the parser state. the operator is implicit. ?-fix oops. i meant that the operand is implicit. the operator is the tag hcf: the text between the tags is NOT the operand. so ... would be prefix? well, what does that operate on? i.e. what gets changed by the tag? it's the parser state, i beilieve. 10:10am back ok 10:20am blarg hey tcn help me how would you do string command line argumaents in asm like "lengua filename" -:- krz [krz@d12.owo.com] has joined #tunes hi krzy! hey hoy, krz om hs: read the nasm doc's on interfacing to C programs, first of all i don't wanna interface with c hs: but you *are* interfacing with C hey krz :(? hs: the kernel and the shell are written in C! i mean for my own programs though i wanna let it take in command line arguments anyway, argv points to an array of pointers to strings argc is the size of argv 10:30am ok now we're getting somewhere om TSUI!!!!! =))))))) hehhehehehehehehehhe now then where is argv? Toggle Switch User Interface! esp+4? yeah and esp+8 is argc k put if you'd read the bloddy NASM docs, you'd see this: push ebp mov ebp, esp pusha ... popa leave ret tcn, are you trying to program HS? :) [ebp+N*4] is the Nth argument on the stack heh I mean, [ebp + (N+1)*4] now then im almost there i just need to learn how to t=do linux file io is asm any suggestions? you can also allocate local variables.. after pusha, just SUB ESP,size after PUSHA hs: no :) excuse me? oh yeah, before POPA, you gotta ADD ESP,size hs: go find a buffer overflow exploit, and see how they call kernel functions, mostly it is putting a value into eax and calling an interrupt .. depends on function oh, file i/o.. you just call the regular C functions.. fopen, fclose, fputc, etc.. im not doing this in c keyword asm * HickServ/#tunes shuns c NASM comes with a macro file for calling C functions easily you can still call c-library functions from asm tho go find it but i don't want any c :( * HickServ/#tunes *cry* then use Retro sorry but lengua is for linux lengua? my compiler file io is so easy in dos you just use int 21 hs: that is what you'd do under linux, but not int 21h, just another int.. hold on (looking) 10:40am :) /usr/src/linux/include/asm/unistd.h lists the system call numbers (in DOS it's what goes in AX) * HickServ/#tunes *uncries* it is int 0x80 look at that file.. figure it out ok it shows how to call functions that take 0-5 arguments or something.. you'll have to look at the man page to know what the argument are for things like open() close() read() write() hey, this stuff would be useful for a version of Retro or Tunes running under Linux grrr! ISP downtime again! I started that in december but I quickly gave up :) hehehe i love leaching off the pool of knowledge in this channel yech. i feel used. :) haha indeed.. hmmm that file doesn't help too much hs: NASM DOCS!!! i know i know HickServ: yeah it does :) im just drinking your last drops of blood it tells ya exactly what ya want to know oh wait i see ok 10:50am om =\ See ya all ok bye cu! take care -:- SignOff tcn: #TUNES (tcn has no reason) blarg so what registers do i put my arguments in? FOO! * Fare/#Tunes hits HickServ with a stick! * Fare/#Tunes hits HickServ with the Assembly-HOWTO no i don't have problems with assembly bavoso i got problems with linux * Fare/#Tunes hits HickServ with a clue hs: well "a" means eax .. "b" means ebx .. find the pattern young jedi in those macros.. * Fare/#Tunes hits HickServ with the Assembly-HOWTO again "S" and "D" are most likely esi and edi .. but i think you might want to listen to Fare ah * HickServ/#tunes is enlightened gcc isn't my cup 'o tea and plus fare uses at&t asm blarg!!! no I don't, you stupid! well the HOWTO would give ya pointers on using asm + linux.. -:- mode/#Tunes [+o Fare] by ChanServ -:- HickServ was kicked off #tunes by Fare (get a clue, dude!) -:- HickServ [thrustit@209-68-229-50.dialup.cust.tfb.com] has joined #tunes hehe :((( deja vu ah! feels better! -:- mode/#Tunes [+o water] by Fare -:- mode/#Tunes [+o hcf] by Fare can't wait till i don't have to use linux whoa! thanks. -:- mode/#Tunes [+o smkl] by Fare -:- mode/#Tunes [+o krz] by Fare -:- mode/#Tunes [+o iStormy] by Fare -:- mode/#Tunes [+o abi] by Fare -:- mode/#Tunes [+o TUNES] by Fare -:- mode/#Tunes [+o AlonzoTG] by Fare -:- mode/#Tunes [+o HickServ] by Fare :D wtf? wtf is that message..? abi: forget wtf water: I forgot wtf 11:00am abi: give me head abi: ah c'mon fare: what's going on in that demented programmer's mind of yours? anti-OOC bigotry what do you think? :) I just hate unnecessary priviledges... HickServ: produce it, i'll seduce it OOC? p oop pardon abi: i had no idea! abi: HickServ is also getting head from you okay, HickServ. abi: forget HickServ water: I forgot hickserv abi: you're so fickle! :) no!!! hickserv? :( abi: forget water HickServ: I forgot water there i think there is a subscription service, too now we're even abi: there is ...but there is a subscription service, too... abi: forget there hcf: I forgot there abi: there is -:- SignOff krz: #TUNES (ouch!) abi: what is there? fare: no idea abi: there? fare: bugger all, i dunno there there is a stupid bot in the channel abi: there is a stupid bot in the channel ...but there is ... -:- mode/#Tunes [-o iStormy] by water -:- mode/#Tunes [-o HickServ] by water Fare: which goes by the nick TUNES om -:- mode/#Tunes [+o iStormy] by water HickServ rumour has it HickServ is the father of BuenOS or chicano or a member of the Sicilian Mafia or getting head from abi hah! well um i don't know what each arg does though anyone here watch party of five? * AlonzoTG/#tunes smashes HickServ with an iron mace party at five??? i don't know, fare er not me HickServ is our only peon 11:10am fare: an american show what's that? hah well, anyway, one of the characters in played by an actress w/ the lastname of devicq i get tons of info from here mabye im not hte peon abi: HickServ is also The Peon okay, hcf. hmmm the doc doesn't tell me anything about file io * Fare/#Tunes hits HickServ with the docs of the libc lib c c * Fare/#Tunes hits HickServ with the Assembly-HOWTO again c c c c c c c c c humbug -:- HickServ was kicked off #tunes by Fare (stop it!) -:- HickServ [thrustit@209-68-229-50.dialup.cust.tfb.com] has joined #tunes i need something to drink something to calm me down abi: HickServ is also needs a cluebat to the head okay, hcf. Take the sources of linux-eforth-1.0c.tar.gz from www.forth.org im just gonna go ask someone else it does some file I/O in pure asm under linux/i386 oh? [I know, I wrote it!] [well, ported it from DJGPP] where on forth.org? -:- HickServ was kicked off #tunes by Fare (LART!) -:- mode/#Tunes [-o Fare] by Fare I am sick of it. -:- SignOff iStormy: #TUNES (Ping timeout for iStormy[rain.futuresouth.com]) ok -:- mode/#Tunes [-ooo abi AlonzoTG hcf] by water -:- water was kicked off #TUNES by Tril (De-op flood (3 de-ops in 0secs of 30secs)) -:- water [water@ppp-tnt-140.tscnet.net] has joined #tunes :( doh! =( I lost me ops... * water/#tunes messed it up. -:- mode/#Tunes [+o AlonzoTG] by smkl hrm 11:20am =) what have i done!?! -:- mode/#Tunes [-o Tril] by smkl -:- mode/#Tunes [-o TUNES] by smkl -:- mode/#Tunes [-o smkl] by smkl welcome to efnet uh eek its lonley up here! oh well... you can have the chan back when I crash... -:- mode/#Tunes [+o abi] by AlonzoTG dont op bots why? goodbye, hcf. -:- abi [nef@bespin.dhs.org] has left #tunes [] it'll become like #os on efnet? -:- abi [nef@bespin.dhs.org] has joined #tunes -:- Tril [dem@sloth.wcug.wwu.edu] has left #TUNES [] -:- Tril [dem@sloth.wcug.wwu.edu] has joined #TUNES -:- mode/#Tunes [+o Tril] by ChanServ * Tril/#TUNES is back from the dead. Gone 13 hrs 2 min 41 secs -:- mode/#Tunes [-o AlonzoTG] by Tril All right! bx is cool, it auto-deopped water when I was away! hey! a takeover! tril: thanks a lot. btw ppl the infobot has this thing wherein after u directly ask her a question (abi: foo?) she expects further questions for like 5 seconds thus u can ask em w/o directly addressing her but w/o the 1st direct address, she wont reply perhaps, unless its a hello type thing hcf, I thought she randomly chose to answer and any question that is "foo?" that she knows she will answer Tril: for hello type things yes hcf? 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) it always works.. not always i forget the exact details then why are you trying to explain it? UTSL! cuz sometimes ppl seem to think that she'll reply to em w/o direct addressing that's beacuse she does. right, abi? 11:30am abi, right? wish i knew, tril hehe abi: hcf is also addlepated okay, hcf. * water/#tunes wonders if anyone wants to discuss something. * water/#tunes just installed WinCVS hello water bwahaha! now i have the power to change! :) What would be the ideal environment for Arrows? would it be possible to write an arrows module for my OS? but i must reboot... brb all. -:- water [water@ppp-tnt-140.tscnet.net] has left #tunes [] 11:40am -:- water [water@ppp-tnt-211.tscnet.net] has joined #tunes wb back now to test WinCVS out. no I don't want to discuss anything, because I'm writing a 2-player tic tac toe game for unix. for school shouldnt take long, though :) really? you think it won't take long? :) this dorky little $2 book I haf sayz C and FORTH are "middle" level languages... =P It lists ada, FORTRAN, Pascal, COBOL and Modla-2 as High level languages. sounds about right smalltalk? it has been said that smalltalk is my favorite abi: forget smalltalk water: I forgot smalltalk what does the book list smalltalk as? it doesn't its a $2 book ok. it has modula-2, but Not smalltalk. smalltalk isn't a language. it's an environment :) so, why does an environment have an EBNF spec? water: even the Glossary doesn't list every language that exists. i mean review 11:50am doesn't C have a ~= operator?? no, wait, that must be awk or perl its =~ hcf: That's two operators, binary assignment and unary complement and it's what i'm using :) with a space in the middle = ~ 12:00pm gakuk! schlrubsk! :) yeah, that too! x ^= -1 wow. could CVS possibly be more cryptic? It is now player -1's turn. water, it's pretty complicated. http://www.loria.fr/~molli/cvs/doc/cvs_toc.html 12:10pm that's ok. i'm wading through the help file. isn't there a simpler system? how about web-based edits of a database? that's what we will probably move to next. ha ha :| abi MOT-2 does a program exist to manage that? abi, MOT-2 is Programming-free DB/Web framework for Apache/PHP/MySQL @ http://tomato.nvgc.vt.edu/~hroberts/mot/ how the hell do i login to cvs.tunes.org? water, read collaboration.html cvs is not complex ok it doesnt explain wincvs, i'll tell you versioning is intrinsically complex it's right in front of me. first set the CVSROOT on one of the menus cvs doesn't have any particular added internal complexity :ext:water@cvs.tunes.org:/cvs ok. water: ssh! then go to Login Fare wincvs does not have that afaik wincvs sucks, then i have all that set, including ssh does it have ssh? it tells me to "set the password authentification in preferences" i can there select either ".rhosts", ".passwd" or SSH server (or local) OK i'll try passwd now use ssh if it has it i already tried that. it's what gave me that error. passwd -> "missing host name in CVS ROOT: :pserver:water@cvs.tunes.org/cvs" : before /cvs really? ok. hoorah! no no!! wait. i'll try again with SSH ssh goes with :ext: i think or does it work with either? 12:20pm ok. it asked for a password when i selected ".passwd", but rejected me as not being a registered user. was that with pserver? SSH -> "set password authentification first in the preferences" it may have a point. holdon i'm going to look at the help file now. obviously WinCVS has a "feature". no, try again. hmm? ok. You weren't added to the tunes group. I thought I already did that, though. ok "no such user water in CVSROOT/passwd" SSH returns the same error that it did before. maybe it needs a ssh.exe on your system, and is not ssh built in? it seems to be built in. the help file mentions "rsh" as well. water, ok let me go to the windows machine and try that version of cvsweb, where did you get the file? 12:30pm http://www.cyclic.com/ i'm going to use the troubleshooting procedures. well, i've gotten a response from the server, at least. 12:40pm -:- binEng [Anders@dialup44-4-44.swipnet.se] has joined #tunes hi, bin hi bonjour, binEng hey tril, the Alma language address that i gave you isn't on the review page. why not? oh, nevermind. i found it. oops. My Asm-HOWTO is considered unmaintained! :( :( :( isn't it? i thought it was considered unmaintained by you 12:50pm Fare: anonymous pserver access is now working for modules tunes,retro, etc just not lisp pserver? should i retry using that? I just received a mail from HOWTO maintainer tjbynum water, well, it will work read-only, i'm still working on your problem ok what did you change? i added a guest user fare: ok now anonymous works for the lisp module what did you change this time? have you things to add to the Asm HOWTO? 01:00pm -:- SignOff hcf: #TUNES (Leaving) -:- hcf [nef@me-portland-us210.javanet.com] has joined #tunes tril: trying to log in as "guest" returns the same errors. water, nononono the login as anonymous maps onto the user guest on the local system oh, ok. and i'm downloading wincvs to remembe how to use it to tell you. brb ok thanks water, i'm downloading the latest beta from www.wincvs.org (1.1b4). That's a much newer version than the one on cyclic's download page (the one in the macintosh directory). It's also 3 megs instead of 500k 1.1b4 is the one that i am using. 01:10pm the command line one probably is more stable, if you don't mind learning arcane syntax. as long as it works. om om? HEY ALONZO Maybe you don't need to say "om" anymore on your new ISP? his old isp had an idle timeout and he's too ignorant to use ping or something intelligent oh. i see. heh normally I don't accuse alonzo of being ignorant and unintelligent, but the Om thing is really annoying. MIT's AI Lab has a lot of papers. mostly on subjects irrelevant to the current Tunes work, though. ATG: have any pointers/comments for the Asm HOWTO update? my IRC client shows time on the status bar, which keeps the tty alive... 01:20pm -:- SignOff water: #TUNES (Ping timeout for water[ppp-tnt-211.tscnet.net]) hum. Anyone has pointers/comments for inclusion in the Asm HOWTO? I guess that's a Q. for QZ... i did, but I think it was just to mention nasm. Fare: i might have some, i'll check to see if u have em and if their worthy s/eir/ey're/ 01:30pm -:- water [water@ppp-tnt-140.tscnet.net] has joined #tunes woo hoo! any progress? -:- SignOff hcf: #TUNES (Leaving) water: Yes. I installed WinCVS 1.1b4 and it f**ked up my windows so it won't boot. actually, my mom's windows. if pro is the opposite of con, what is the opposite of progress? parliament Tril: congratulations! 01:40pm Tril: that's why I have a .tar.gz of a pristine Win* installation so that I can rm -rf /C/* then tar -Ixf /save/win.tb2 hmm the best Windows administration tool is... Linux! (not pristine: added drivers and necessary survival tools included) I also have all my failsafe DOS utilities on D: so windows can safely fuck up the C: on the other hand, my tic tac toe program appears to be working :) :) Fare: I don't suppose you can send me a working COMCTL32.DLL from Windows 95 or anyone else i could please ok. one minute, please. just make sure it's not the one from WinCVS 1.1B4 right maybe I could just restore the backed up registery. let me try that first you can send anyway COMCTL32? don't ask 01:50pm no success. (with teh register) ooh 200 bytes per second ! fare the only mail I found is some xcom@tunes.org stuff, which I forgot to add a folder for in pine or do you mean look *at* my mailbox? oh, wait, there's some killall -HUP fetchmail causes fetchmail to exit. -:- SignOff water: #TUNES (Ping timeout for water[ppp-tnt-140.tscnet.net]) fare thanks -:- water [water@ppp-tnt-58.tscnet.net] has joined #tunes hum, which is larger: a bazillion, or a gazillion? arrgghh! * Fare/#Tunes unargs water * water/#tunes slaps Fare around a bit with a large trout 02:00pm does anyone have some interesting research to share? water: there is plenty of interesting research that I would like someone to read, but don't have time to myself poor trout! a bunch of papers on persistence, and a bunch more on Component oriented programming. tril: send me the links, then * Fare/#Tunes calls the SPA (Societe Protectrice des Animaux) tril: do you still want that file? http://www.asktog.com/columns/027InterfacesThatKill.html http://www-ppg.dcs.st-and.ac.uk/Publications/ and http://www.math.tau.ac.il/~guy/COP/ water: no thanks, fare sent me one to try for now ok. i'm loading up those pages now tril: i have all of those on persistence, and have read most of them. fare: i've already read that page on the Long-EZ, because i used to study Rutan's design methods. well, i haven't read about that dos and unix stuff, though. 02:10pm Fare: comdlg32.dll please another one? yes, now I can get the desktop, but can't run anything useful like notepad or netscape don't you have a win9X CD around? fare: i don't agree with the article's suggestion, not because i think users should be uneducated, but instead because it associates causality with individuals. next time, make a win9X tarball... water: I'm not sure what you mean... brb human relationships transcend the control of the individual. hence, claiming that an individual is at fault is irresponsible, because the individual does not have full control. I do not read the article like that. how do you read it? Surely, it argued that the pilot made some bad decisions, but it does insist on the responsibility of the interface designer... insisting on this responsibility is the main point of the article, IIUC back the interface designer is again an individual, though. the problem is with language; with subconscious mechanisms. they also insisted on the process that led to bad design that is, not caring about design to begin with who insisted? not caring about design? 02:20pm the point of the article is that some people knew of a bug and didnt do anything about it. let's not overargue about that article, ok? but "knowledge of a bug" is relative. forcing everybody to strictly report what you consider "bugs" is tyrannical. specifically important security-related bugs sure. But having a development process that leads to no one taking responsibility for bugs is terrible it means that *by design* (metadesign), there will be unattended bugs well, then our thought processes themselves must be terrible. they are. That's why we better work in groups but groups aren't the end solution. they're a stop-gap. 02:30pm -:- s_r [s_rr@phila-dialup327.nni.com] has joined #tunes tril: there's a language called Pike hi s.r water, to be added to review? hey maybe so this is tunes hrm what's on your mind, s_r? OS Deev os dev all day long, water how about you? my own little project: the Arrow system describe it I am working on a little thing called Virix it's going to be an "endokernel" OS abi: Arrow? somebody said Arrow was a homo-iconic information representation system intended to support programming and human languages in the way that Tunes should. thanks, abi. water: de rien wow human languages! hehe hmm cool how's progress, water? well, the human languages part has nothing to do with Tunes. water, sure it does :) i'm working on an implementation in SmallTalk tril: what? I'm planning natural language support in my model. is there any working source code produced yet by the tunes team tunes is a very interesting project s_r: what's endokernel? How does it oppose exokernel? s_r: there's Tril's Lisp code and the Retro OS, and Fare's building a Scheme compiler. s_r: there's retro or tril's lisp code, or my oooold boot code one could say that it works as an OS inside every process s_r: so it uses VM creation to spawn processes? VM = Virtual Memory? VM = virtual machine (like an OS) it copies evolved parts of the kernel into the OS the kernel is actually defined as a "precursor" that is started intitially initially evolution? how? * Fare/#Tunes is checking links for the asm howto it changes itself through guidelines in the process header shit, had I known maintaining was such a PITA, I'd never have undertaken it ;( i haven't worked out too much of this it probably won't get anywhere, water :) water how's work on your operating environment? Fare: Now yo know why I don't become a Debian developer. hehe 02:40pm you can issue a package w/o becoming maintainer, you know... s_r: I think it's basically _wet_ water: i know what you are thinking... "This guy is full of BS" :) Fare: You mean Non-maintainer uploads? Those are only permitted in rare circumstances.. Tril: Pike is a dynamically-compiled and interpreted language with C-like grammar, module support, and built-in structured types for rapid application development. Tril: oh water: feel free to add an entry to Languages.html by CVS maybe it's already in the TODO pipe of L.html well, if i could log on, then i would! such as for securityfixes and the official maintainer of the package didn't do it water: can't you? Water: wincvs11b4 comes with a command line client. Try using it? water: install Linux! L I N U X ! fare: no. that's what Tril was trying to troubleshoot. water: I was trying to, but I highly recommend not using wincvs11b4 because it screwed up my system. * water/#tunes thinks programmers have lost the art of subtlety. There's a fine CVS client for Windows named LOADLIN Fare: And developers are expected to be subscribed and actively following debian-devel, which gets 60 posts a day by the time Tril manages to reinstall Windows to get WinCVS working, you have plenty of time to install Linux and that time will be *much better* invested for all of us Fare: I'm not reinstalling! I only need one more file (and maybe a few more after that) i have a copy of linux which used to occupy 75% of my hard drive. why don't you send it to me? Tril: by the time you ping-pong all these files with us, you'd have time to reinstall and make a tarball... Tril: I did send the file to you already! * Tril/#TUNES observes fetchmail is not running AGAIN. * Fare/#Tunes resends the file wait you probably sent it grrrr! fetchmail -d ? Tril: reinstalling and making a tarball is a good idea, anyway! the links to EMX seem to be as dead as OS/2 os/2 isn't dead! ARGH! where's uudecode? Why can't you send in the same format as before EU internet strike in ten minutes, folks. Soon time to log off... OS/2 is so dead... -:- SignOff s_r: #TUNES (Ping timeout for s_r[phila-dialup327.nni.com]) we should be out demonstrating for free kevin tril: os/2 is as alive as the amiga 02:50pm water: do you have os/2? tril: not in the last few years. i gave it to a friend who wanted the copy. -:- s_r [s_rr@phila-dialup257.nni.com] has joined #tunes hey, sr abi, tunes? rumour has it tunes is a useful nevertheless expedient system or at http://tunes.org/ or a free reflective computing system die, OS/2, die! rumour has it abi is Is it possible to map only parts of a page? For example, say I wanted 1kb of memory from a page to be available to a process, how would I implement only the 1kb? (With paging enabled) Thanks, arkadion@hotmail.com ummm whoops eheh IBM won't publish it because of MS code fragments in it, anyway! me? somebody said binEng was not formally involved in TUNES. me? hmmm... s_r is working on a little thing called Virix abi: I am also at bineng@bespin.cx okay, binEng. me? you are not formally involved in TUNES. or at mailto:bineng@bespin.cx s_r no is there any way to map only less than 4kb of a page? * water/#tunes is waiting for some useful, not expedient, discussion. abi: No, I am not formally involved in TUNES, but still bineng@bespin.cx okay, binEng. Tril: I can send uuencoded stuff from the command line. I hate going thru a visual interface when I repeat things more than once water what kind of kernel will the arrow system use? bE: are you ashamed of TUNES? fare, i managed Tril: you're a manager! fare: no, but I haven't posted a single mail to the mlist yet, have I? abi, linux? well, linux is bloated abi, FreeBSD? FreeBSD is probably pretty stable... We replaced a crappy linux box with FreeBSD for the mail server s_r: the arrow system is an information system that will reflect on itself. most likely, it will have no kernel. Linux is arcane computers are arcane! s_r: the arrow system will be somewhat like SmallTalk or Self, but with dynamically-created VMs. yeah, so what do you suggest? ;) ahh hmmm virtual machine will it be recursive of itself ok, I'm leaving. cya all s_r: recursive? not exactly. water are you doing any programming currently on the arrow system? s_r: the VM is the thing defined by the current interpretation of a particular arrow, so that multiple (or infinite numbers of ) VMs may exist simultaneously. will the arrows be represented in a graphical interface? s_r: the current SmallTalk implementation is not ready for programming in general. me? hmmm... s_r is working on a little thing called Virix water? water is, like, an expression of the Tao -:- SignOff binEng: #TUNES (Ping timeout for binEng[dialup44-4-44.swipnet.se]) FreeBSD rules s_r: probably. i chose SmallTalk because it comes with a GUI and a drawing and app library. 03:00pm water wouldn't you want to implement the drawing and GUI yourself? abi, FreeBSD? FreeBSD is pretty stable... We replaced a crappy linux box with FreeBSD for the mail server s_r: yes, eventually. FreeBSD is great ...but freebsd is pretty stable... We replaced a crappy linux box with FreeBSD for the mail server... FreeBSD is great and is pretty stable water i also was thinking about a GUI where each "object" in the OS is represented by a 3d polygon. a GUI where everything is presented where the user interfaces with the OS s_r: you would like the Merlin project, then. water: what about Merlin suggests that to you? I have never heard of that project perhaps I should look into it * s_r/#tunes looks in to merlin abi: Merlin? water: bugger all, i dunno abi: self? self is a prototype-based object system darn. darn is probably like damn. Jecel abi: Jecel? water: wish i knew abi, merlin is at www.lsi.usp.br/~jecel/merlin/ ecel's Merlin project of a SELF-based OS (and here a paper about it). thanks, tril. water is SELF powerful and flexible? abi currently doesn't have the Review project in her brain :) s_r: i'd say so. instead of using any existing languages I hope to implement a new one tril: that's a good idea. i want a language that melds assembly and C together sort of the powers of each so I could write everything (including boot loader) in it s_r: but then you're re-inventing someone's wheel. instead of using simple operators for most funcitons it will be defined as opcodes s_r: and your code will not be portable at all. well each function will be defined as an opcode reinventing FORTH? s_r: because my arrow prototype sits on SmallTalk, it's portable to almost every OS and GUI, and even some PDA's. hmmm FORTH? FORTH is, like, at http://www.forth.org or a nice user interface is anyone here using Retro? retro? well, retro is at http://bespin.cx/~tcn/retro.html 03:10pm eros? somebody said eros was at http://www.cis.upenn.edu/~eros/ or "Extremely Reliable OS" s_r: /msg abi ok unless you specifically want to share, of course ok Fare Fare have you used Retro? -:- Connection closed from irc.us.openprojects.net: Success -:- Use /Server to connect to a server -:- Connecting to port 6667 of server irc.us.openprojects.net [refnum 0] -:- BitchX+Deb1an: For more information about BitchX type /about -:- Welcome to the Internet Relay Network TUNES (from norton.openprojects.net) -:- Your host is norton.openprojects.net, running version u2.10.05.18.(ipcheck4-5) (from norton.openprojects.net) -:- This server was cobbled together Wed Apr 28 1999 at 13 24:47 EDT(from norton.openprojects.net) -:- norton.openprojects.net u2.10.05.18.(ipcheck4-5) dioswkfcg biklmnopstv -:- [local users on irc(1)] 0% -:- [global users on irc(135)] 40% -:- [invisible users on irc(204)] 60% -:- [ircops on irc(13)] 4% -:- [total users on irc(339)] -:- [unknown connections(0)] -:- [total servers on irc(31)] (avg. 10 users per server) -:- [total channels created(99)] (avg. 3 users per channel) !norton.openprojects.net Highest connection count: 14 (13 clients) !norton.openprojects.net Welcome to Open Projects! You are on 2 ca 1(2) ft 14(14) tr. -:- Mode change [+f] for user TUNES -:- Mode change [+iws] for user TUNES -:- JOIN activated by "TUNES #tunes tunes@bespin.dhs.org " -:- TUNES [tunes@bespin.dhs.org] has joined #tunes -:- Topic for #TUNES: www.tunes.org -:- topic set by ChanServ [Sat Jun 5 00:45:25 1999] -:- [Users(#Tunes:8)] [ TUNES ] [ s_r ] [ water ] [@Tril ] [ abi ] [ smkl ] [ AlonzoTG ] [ Fare ] -:- Channel #Tunes was created at Sun Feb 28 08:48:06 1999 -:- BitchX+Deb1an: Join to #tunes was synced in 6.982 secs!! -:- Mode change [-s] for user TUNES 03:20pm "used" is a big word. I've booted it tril: where's a good paper on MOP's? none ;-> -:- jdl [jiml@ultra1.inconnect.com] has joined #tunes well, the one I'll be writing one day jim! Fare! water: couldn't find any. You really have to just read the book. abi: AMOP AMOP is quite interesting. or _The Art of the Metaobject Protocol_, by Kiczales, des Rivieres, and Bobrow hi jim. jim: Hi. Now if only Maneesh were here tril: i've read that one. then we could get this thing sorted out tril: which thing? Tril: Just been emailing him... he's online. Sort what out? oh, Maneesh doesn't understand Prism Tril: read the most recent exchange? tril: oh yes. i remember. apparently not. Tril: I think we're coming to a mutual understanding. ...But on the other hand, I don't blame anyone for not understanding Prism. It's hard to see how what I have relates to my goals. ok, well, keep the discussion on the list, i think it will help me understand prism more. I couldnt defend it to Maneesh here the other day. 03:30pm Tril: NP -- thanks for defending it though! Water: I need to read your Arrow draft. I think there's some interesting similarities between Arrow and Prism's meta-metamodel. I think Arrow's going to supplant Tunes ;> Water: Have you looked at Prism yet? s_r how can it supplant it if it gets done first? hehe :) jdl: have you read LAMBDA: the ultimate little languages, by Olin Shivers? jim: i'm not ready to release the final copy. the draft is at http://www.tunes.org/papers/arrow.pdf Fare: No... sounds very relevant though! Any more info? wait. that's not the right address Water: I've seen the link, just have been lazy. :) ok abi, arrow is also at http://www.tunes.org/papers/Arrow/ okay, Tril. -:- SignOff smkl: #TUNES (sleeping ...) Fare: is LAMBDA... a book or an article? (I can't find "Olin Shivers" at fatbrain.com) an article i think an article is written by Mentifex. http://www.ai.mit.edu/~shivers/citations.html#ll -:- SignOff Tril: #TUNES (Ping timeout for Tril[sloth.wcug.wwu.edu]) abi: forget article water: I forgot article Fare: thanks -- I'll look at it Fare: I'm downloading it, but I don't think I have an eps viewer installed... can you give me an overview? (I'll install a viewer later, but I'm curious) -:- _QZ [brand@p0wer.qzx.com] has joined #tunes jdl have you tried Retro? jdl: just print it. 03:40pm s_r: No water: Don't think that will work on my machine (win95, InkJet printer) -:- Tril [dem@sloth.wcug.wwu.edu] has joined #TUNES -:- mode/#Tunes [+o Tril] by ChanServ * water/#tunes doesn't think that there's enough to try. s/InkJet/DeskJet/ jdl: install gv install linux Fare: I know, just haven't needed it since I installed Win95 jdl: install Linux! L I N U X ! So I haven't gotten around to it yet :) * water/#tunes thinks that Fare should be more enthusiastic about Tunes than about Linux. subtle... :) I tried Linux, didn't care for it. Water: But I can't install Tunes :) you could help with the discussion of ideas, though! -:- s_r [s_rr@phila-dialup257.nni.com] has left #tunes [] * Tril/#TUNES curses at C no return-from INTERCAL? no, that's COME FROM. LISP has return-from. * jdl/#tunes is starting to read the Arrow draft 03:50pm Water: Have you looked at Prism yet? I want to compare/contrast with Arrow oh, you mean like calling the /cc of an upper function? CommonLISP yeah ok yes, but i'm not very familiar with it. there are things about it that i don't quite understand. sorry Water: such as? * Tril/#TUNES phears C error checking: if (function(arg), misplaced_arg_2) { ...} <-- no complaint about comma-separated list inside an if condition??? * Fare/#Tunes understands the general idea of Prism and agrees, but isn't sure to understand the general idea of Arrow... Tril: phear, phear! It's C! jdl: what precisely is a model and meta-model? (what defines them) Water: I don't think I understand your question... a model is a representation of a concept (is that what you're asking?) Water: You use model in your paper as well.. how do you define it? models are objects, and meta-models are types :) no, i mean this: is there a formal definition for your models? tril: thanks for obscuring the issue. Water: Not formal. Metamodels define models, but not formally. jdl: a model is a mathematical term for me, but models for the Arrow system generalize it to extend below the language level. --This is intentional; I don't think formal methods are generally useful for real-world programming ahh. that's what i wanted to know! water: do you define 'model' in your paper? (I didn't understand your definition above) water: yes, metamodels are unclean :) models are a well known math concept water: but they could be formalized if someone desired. I don't proscribe a metamodel format, I only provide one possible format not the same as in prism mathematical model: a set of constants and relations and variables that form a vocabulary for a logical theory. jdl: not a format, a set of logical requirements. abi: a model is, in formal logic, some kind of concrete substratum such that we can faithfully interpret logical propositions as observations on the model that too. I use 'model' in the conventional (natural language) sense, not in the mathematical sense. 04:00pm anyway, the purpose of models in arrow is to express the desire to provide re-modelling for the system whenever computationally achievable. water: define 're-modeling' hehe. a water vs jdl debate! jdl: translating information provided according to one model in the theory of another. maybe we'll understand better afterwards! fare: dare we hope? I think there is a lot of good intuition there that needs be formalized water: does the meaning of the model change? the problem is: once it's formalized, it's no more the same thing but still, better a forcibly lacking formalization than none at all... fare: maybe. i hope otherwise. (i have a few tricks up my sleeve for that) jdl: perhaps we should shift terms. for me, an ontology is more important. * jdl/#tunes flips through dictionary ontology: the branch of philosophy that deals with meaning jdl: an ontology is a term from ai research in the 90's. jdl: although my own use is different. water: define your use of ontology tril: do you know how to ask CVSweb to show the last version of a given file? abi: ontology? it has been said that ontology is a statement of a logical theory in some domain jdl: for me, an ontology assumes nothing about the logical inference structure that processes what we ordinarily call 'facts'. Fare yes ? can't you just click on the filename water: but what do you mean by 'ontology?' jdl: an ontology is like the 'shape' of knowledge. water: clear as mud, sorry! jdl: in an ontology, the inference flow-graph between propositions is reified. * jdl/#tunes flips through dictionary again * Tril/#TUNES knows what water means * jdl/#tunes doesn't it makes the inferences available as objects it's like being inside a virtual machine, and restricted by its rules So an ontology is a set of concepts? sort of. it's a way of interpreting the information in the world. 04:10pm reify: to regard or treat an abstraction as if it had concrete or material existance Tril: no I mean, w/o clicking, as a direct link... http://www.tunes.org/cgi-bin/cvsweb/path/file?rev=HEAD water: I think I understand... can you give an example? imagine yourself in a maze, the possible paths are the ones you are allowed to go to next tril: I don't see how that relates * Fare/#Tunes is lost when the word "ontology" is uttered. * water/#tunes sighs. Not that it be much different from "semantics", but at least I've studied formal notions of semantics. Fare: semiotics? :) ouch. I meant semantics in the CS meaning... the Cyc project has a notion called a micro-theory which is like an ontology. water: that doesn't help, I've never heard of the Cyc project. ;) wait! i've got it! it's like a static type system. amazingly, that makes sense. * Fare/#Tunes never saw the details of CYC water: before you go on... imagine you're programming, and you can't make new types. you then have to caste all information in terms of those types. being in an ontology is searching through a search tree, where each branch from a node is a logical inference that follows from that node * Fare/#Tunes remembers a paper by Cousot in POPL'97, whereby Type Systems == Abstract Interpretations and you can't go anywhere that a branch isn't presented to you. (thus the typechecking part) what are you searching for? water: multiple ontologies are possible? jdl: definitely jdl: a dynamic ontology would be provided by a schema for ontologies. * Tril/#TUNES wonders if soon everybody here will be speaking the same language water: would I be correct in saying that an ontology defines a set (in the mathematical sense) of knowledge? water: let's wait before talking about dynamic ontologies. :) jdl: well, maybe. the set would allow reflection, though, because an ontology is only a _filter_ for information from the world. so what you call an "ontology" is what I call a "semantics"? and your meta-models are what I call a "semantic framework"? OK... how would an ontology differ from a metamodel? 04:20pm fare: probably. my ontologies are supposed to be expressible as 'shapes'. water: my metamodel "shapes"???? what's that? water: shapes= tree structures? tree-like perhaps fare: don't get upset. shapes are like Robinson diagrams, only with arrows, and definitely not tree-like. Robinson diagrams???? om * Fare/#Tunes is lost * AlonzoTG/#tunes oms to be bad.... * jdl/#tunes knows where Fare is because he's lost too :) the sever ping keeps me active * Fare/#Tunes kicks AlonzoTG -:- mode/#Tunes [+o Fare] by Tril -:- AlonzoTG was kicked off #tunes by Fare (don't be so bad!) ahhhhhh! feels SO good! -:- Fare was kicked off #tunes by Fare (don't feel so good!) -:- Fare [rideau@quatramaran.ens.fr] has joined #Tunes damn. i suppose 'shape' is a bad term. yes :) water: how do ontologies relate to Prism metamodels? -:- Myrmidon^ [ubiquity@dm6-35.slc.aros.net] has joined #TUNES water: uh huh Howdy. let's ignore that word 'shape'. -:- mode/#Tunes [+o Fare] by ChanServ -:- AlonzoTG [Alonzo@216-164-225-2.s256.tnt6.lnh.md.dialup.rcn.com] has joined #Tunes hi Myrmidon jdl: well, what's a metamodel? Hey Tril. water: A Prism metamodel defines a set of models and the meaning of each one... (more) water: Usually, a metamodel deals with a particular concept, like integers or strings or imperative programs (to be concrete) (more) water: for examples, see http://www.teleport.com/~sphere/index.html -- there's several small ones linked from the main page, you can't miss 'em jdl: what about things that _aren't_ programming terms? jdl: describe a metamodel for something like that. Ok -:- GMOL [yadavm@elvis.scripps.edu] has joined #tunes Let's model a penny. The form of the model is: _Bit: * I will be here intermitantly I can't chat too much but I am here (In other words, pennies are modeled with a single _Bit atom in either of its two states.) The meaning of each model is as follows: When the model is "_Bit: FOO" we are modeling a 1999 penny with its head facing up. When the model is "_Bit: BAR" we are modeling a 1999 penny with its tail facing up. (Done) okay. i think i know what the difference is. Hi, GMOL -- Maneesh, right? i'm confused about whether there's any code in a ontology ontologies are not absolute. they are not ever supposed to be considered as the only or even the canonical model for a system. yes tril: ontologies can have code, but they are more general than that. 04:30pm GMOL? you are assuming you want a linker to help you develop my OS, no? excellent, GMOL: jdl is Jim Little abi forget gmol Tril: I forgot gmol wow, if we have water, jdl, GMOL, Tril and me at the same time, what gives! -:- tcn [tcn@cci-209150250119.clarityconnect.net] has joined #tunes water: how is that the difference? And what are the similarities? (I'm hoping my understanding of metamodels will help me understand ontologies) jdl: sometimes, we need much more elaborate structures than can be simply described by bits and streams jdl: metamodels sound like they assume that no other interpretation exists, or that other interpretations are irrelevant. Fare: I don't want to confuse the ontology discussion, but I feel that any structure can be built by the basic _Bit, _Stream, and _Map atoms. jdl: can metamodels describe a penny the way that chemistry and molecular physics do? water: ahhh.. very true (irrelevant, anyway). Each metamodel stands alone, although it may be FORMED from other metamodels. (See GUID metamodel on my web page (teleport.com/~sphere) tee-see-en! hey everyone.. looks like we're all here now water: metamodels can describe a penny anyway the author wants. Since the meaning of each model is defined in natural language. time for the shooting at OK corral! all? jdl: that's a definite difference, then, between ontologies and metamodels. tcn: you are working on your persistant storage no? yeah jdl: have you seen, e.g. the Coq formal system? water: Or the meaning can be defined formally, but I haven't created a way to do that (and have no plans to, but someone else could) jdl: you'd learn a lot by seeing how modern formal logic systems do and some do things in ways completely different from Coq Tril: see the definition of absent in fortune... water: So ontologies define the meaning of models, but they do it formally? more critcism (I am being objective here) how can you possibly code anytning useful unless you have defined you're mutual exclusion mechanism's and system call mechanisms and thus memory map? tcn: any progress? Fare: I had to go shopping :( not only that, but they also assume nothing in the way of axioms or context. it is all explicit (for the 'system'). water: ALthough I still do not understand it, how does your arrow paper give "meaning" to arbitary computations? water: Can you give an example of an ontology like I did? (I.e., very small and simple but illustrating the point) GMOL: arbitrarily! :) i mean relative to a given ontology. ok. newtonian physics and its information discovery process form an ontology. water: why? basically my question is can you give an example of how the semantics are encoded on a simple computation ? jdl: it has a basic vocabulary: basic logic, mathematics, and Euclidean-shaped space. jdl: and, of course, a good amount of common-sense. 04:40pm water: I think I understand what an ontology is. They differ from my metamodels in that they are defined formally with Arrow, correct? water: (Whereas my metamodels are defined outside of Prism with natural language (mostly)) jdl: well, that's how i implement ontologies in Arrow. jdl could you answer the last questrions that i gave you jdl: i see. yes, i made the choice to express ontologies directly within the system. gmol: let me make sure I understand what an ontology is first Ok, I understand. GMOL, care to repeat your questions for the audience? :) basically my question is can you give an example of how the semantics are encoded on a simple computation gmol: huh? * Tril/#TUNES has no idea how these people can tell that an ontology and a metamodel are the same tril: they're only similar :) water: How does the ontology relate to mathematical models? ok. GMOL: to answer your question, I already have a memory map (see kernel/memory.inc).. it's not set in stone. As for system calls.. the method varies between CPU architectures, but there's not much to it - it's for calling supervisor code from user code, with security checking (and a performance penalty). feel free to tell me to "read the draft" one more time :) * water/#tunes settles in for the long haul. GMOL: I try to avoid system calls by using page-mapping a mathematical model works only at the level of first-order logic it assumes first-order predicate logic: there are only variables and propositions, and they are often statically-specified. right a Robinson diagram of a model is a representation of all true statements within that model. * Fare/#Tunes tries to hack the cvsweb script... Rob.diagram is an arbitrary graph of arrows, right? tril: not necessarily, although the arrow system can interpret it that way. water: I thought the Arrow system could interpret anything as a graph of arrow. :) I at least am beginning to understand this ontology thing to some extent.... 04:50pm I see that meaning comes from isomorphisms. jdl: well the statements in a Robinson diagram are easily seen as arrows of statement application. the presence of a particular arrow in a graph is isomorphic to the truth of a statement in a model as shown by a Robinson diagram. water: ok now, the rules governing the inference relations between statements in a model are the axioms of its logic. tcn: you did not answer with respect to mutual exclusion these axioms of predicate logic are not inviolate, neither is there something special about their choice. GMOL: that's a buzzword to me GMOL: define it in ontologies, the inference diagram is explicit, as well as the control graph that takes axioms to the inference system. it unifies a lot of things into one whole. * water/#tunes is developing quite a headache. what's a Robinson diagram? (have I missed something?) Fare: yep :) water: this is over my head, unfortunately repeat: a Robinson diagram of a model is a representation of all true statements within that model. i already had it that way, water. tcn: He's asking about multithreading, what is your synchronization/communication primitives? and how are they implemented.. they're not yet how is that a diagram? Fare: good question! fare: it's just a technical term. all in due time, maneesh (note: for repeat, you can copy/paste, or use up-arrow/down-arrow to browse the history) fare: some researchers tried to express it as a topological diagram. what if there are infinitely many true statements? fare: (i know how to do that.) between "I know" and "it can be formalized in a sound way that allows for automatic proof checking", there's a long way fare: then the diagram has an infinite number of elements! :) 05:00pm * water/#tunes slaps Fare around a bit with a large trout I'd like something like Coq, Automath, or else, to check proofs. (has anyone seen the Journal of Formalized Mathematics?) * Fare/#Tunes calls the Society for Protection of Trouts GMOL: Still here? fare: stay on the subject! i made a comment about knowing how to scroll through the message log, and you mis-interpreted it! so did I one minute. brb * Fare/#Tunes laughs Well, that will have an impact on how you do your persistant stoarge, I really don't see how you can go about doing it without that as persistant storage is shared resource and your performance will depend quite heavily depending on how you exclude it. GMOL: oh well, it's just a prototype ;) GMOL: You had a question for me? ok, i've returned. Tcn: don't give in to him he's attacking the entire project based on youre credibility! he thinks you represent TUNES! haha * jdl/#tunes thinks TUNES represents TUNES since it's reflective what's everybody feel like discussing? The Heinlein story "By His Bootstraps" should be required reading for everyone here. :) don't talk to me about required reading! :) yeah water read it all. we need to catch up to him he read for 3 years before announcing his presence to the list. no I don't. I don't think an OS is what we need. But if you are going to make one either a) don't make grand claims when you are apparently are not even with the fundamentals of OS design.. b) Read a little something before you go about making one, do something original and design it so that it can acutally be made and be useful in some respect one day. You simply do not tackle persistant storage until you know how IPC looks, you don't even realize how much PS depends on IPC/// GMOL: You're making a big deal out of nothing jdl, please answer the stuff in the letter before you mentioned you were on IRC GMOL: What would you like to know first? You totally shoved off htat question on IPC, how are you going to actually make a pereistant store in a multi-tasking OS when you havent desgined mutual exclusion? GMOL: Once PS hands out some storage, it's not PS's concern 05:10pm IPC is something you don't do gratuitously jdl, whatever I asked in that order if that's ok. * water/#tunes decides to focus on downloading more papers and implementations from the web. GMOL: I agree with your remarks GMOL: but I do think what we need is also an OS tcn: Since multiple processes are presumably going to ask for storage, concurrent access to the PS must be handled. It is an important issue, but I think GMOL could be a bit more tactful... it's simple: QUEUE it's a useful, altho not indispensible, aspect. maybe a FAQ... tcn: A queue alone is not enough. You still need to manage concurrency in the queue. jdl: same as in any other IPC blah, blah ... blah blah blah.... jdl: it's just another thing I have to code, sooner or later Sigh, but it is not...how you make you IPC, what size messages you move around how the targets are specified has to be thought out before you make apps. BTW I tried tact initially, and people refused to adress the concerns, so I have to be more direct GMOL: People? The only person you are addressing is tcn. GMOL: I can't do everything at once. I'll worry about that when I need to. I would other people would have realzied that I am adressing things that are very fundamentally wrong, no one responded. Right now I'm working on PS and FORTH GMOL: You asked me where the Prism syntax would be used. By Prism syntax, did you mean the meta-metamodel? tcn I once thought that you maybe be able to make something unorignal but at least complete, it is now clear that you won't even write an actual OS it will just be some code that boots. For all those who want an OS for TUNES, this is a concern for you. jdl:yes Every now and then, I rewrite a big chunk of the OS, because I see a better way. It's a prototype, nobody's using it for serious work, so it's ok to change it like that. Is this not percisely the kind of stuff TUNES is trying to improve in software? GMOL: The Prism meta-metamodel is used internally within the Prism compiler to represent models. It's also used by developers to describe metamodels, and by language developers to write parsers. It's not used by ordinary programmers. Retro != Tunes ^^^^^^^^^^^^^ Retro = Potential LLL LLL = To be thrown away I never said it did but Tril has stated more than once that Retro is intended for use at the lowest layer.. Retro and Tunes don't even have the same goals Then why the hell are you making it? retro = experiment to gain XP XP for what? so that he may become hacker level N+1 05:20pm GMOL: If you ever read the LLL project, you would know that LLL is a temporary hack for bootstrapping ONLY, and when HLL is running, the LLL gets phased out and its functionality is taken over by the HLL. which will increase his tohit, his hit dice, and more ok, this strikes me as odd. why are we trying to re-invent the wheel, only to throw it away for a fusion reactor? water: Because tcn wants to continue developing Retro, for his own use, after TUNES doesn't use it anymore. IHMO, assembly language will be around for a long time to come.. they were saying "assembly is obsolete" 40 years ago. it will be around, but not quite in the same way as it was 40 years ago But no one is doing application design in asm anymroe like they used to... 40 years ago, just EVERY program was in assembly! 10 years ago, anything where performance mattered HAD to be in assembly well, I exaggerate, but "High-Level" languages like ... FORTRAN were reserved for the Elite GMOL: Did my response answer your question? tril: well, this begs the question: why would Tunes want Retro at all? water: I agree that an existing OS would be just as suitable, but I think that if tcn wants to work on an OS, more power to him! also, the algorithmic and driver structure in retro may be usefully recycled... I DONT KNOW! I'm the HLL person. I don't want a LLL I already said that... There are PEOPLE here who DO want a LLL, let them make it! I don't care! water: BOOTSTRAP the hardest thing about reflective systems (just ask Jacques Pitrat!) the simpler the better.. imagine hacking *BSD into a reflective system.. jdl: i'm objecting to LLL development occurring within Tunes. water: also, tcn will do it, and won't do the HLL. I'll do the HLL, and won't do the LLL. We're complementary! water: How does it hurt? don't see difference: see complementarity. -:- Brian [user4453@98A95AA3.ipt.aol.com] has joined #Tunes Don't see difficulty: see opportunity Brian: hi. Have we met before? No i dont think so? * water/#tunes 's first name is also Brian. do you know the Tunes project? -:- Fare has changed the topic on channel #Tunes to: Reflective Computing System Fare: When are YOU going to do the hll? No what is Tunes silence * Fare/#Tunes hides Nobody knows! :) Brian: there's a FAQ * Fare/#Tunes should definitely add a FAQ about whether it's a language or an OS... tunes is project to make a free software system made entirely of objects, which has never been done before. brian: Tunes is a system of software that can adapt itself to new ideas. Well can anyone recommend a good soun sampler/tracker for linux * Fare/#Tunes should definitely finish the dozen of articles he's begun, including "Why a New OS?" Brian: everybody here thinks Tunes is something closely related to what they're working on. Brian: sorry. You should go to #linpeople or #debian for that Thanks! Brian: there's funktracker :) jdl: only I am sure of it I know there are some good samplers.. not sure about trackers, though. fare: so am I ;) is there mouse sopport on funtracker * jdl/#tunes is just kidding Brian: no it's pretty simple.. it only does 4/6/8-track MODs, I think.. text-mode.. hehe water: ok, ontologies, got it. How do ontologies relate to Arrow? Brian: it's been a few years since I looked.. someone may have written a good one by now. * jdl/#tunes is trying to lead the conversation :) Im looking for a program that you can take sound samples and loop them in many sound channels 05:30pm hit the search engines.. :) -:- topic unset by Brian on #Tunes well, arrows can be considered elements of ontologies. ontologies give arrows meaning, while being constructed from arrows. * AlonzoTG/#Tunes has the outline of a LLL but it will take me decades to complete it.. ;~((( water: ah! they ARE a lot like metamodels! (except for the 'constructed from arrows' part. well, the 'constructed from arrows' part definitely sets it apart, but so does the fact that meaning is much deeper in my system than it is in yours. water: How would you create an arrow representing the number '42' in the ontology of integers? water: what do you mean, "meaning is much deeper?" water: define 'meaning' jdl: hold on! water: sorry :) first, an arrow doesn't just "represent something"; it represents something in one ontology or domain to another ontology. meaning is _binary_. so who wants a LLL? water: define 'meaning' GMOL: i don't want to have to hear about a LLL, i just want to choose a simple one. -:- s_r [s_rr@phila-dialup598.nni.com] has joined #tunes GMOL: a high-level language that can handle low-level stuff would be much better jdl: don't ask the impossible. i refer, basically, to interpretation of information. GMOL: Ask Tcn who wants a LLL, he is maintainer of the subproject and should know the list of members :) anyway, you all know I like Forth water: sorry... that's what I thought you meant, but I wasn't sure But let's look at the big picture. We want some sort of system that alleviates many of the poor semantics in OS's and problems in software design, yes? Tril: You mean the subprojects have member lists too? ;) tcn: they do now. feel free to update yours first! water: so, within the context of a particular ontology, an arrow is interpreted as representing a _____? A concept/thing? Or a _____ in another ontology? Or what? jdl: anyway, to represent 42 to some theory, you select the 42nd element of the integers starting from 0. however, you can't construct this selection without iteration itself, which also defines the integers. the meaning is circular in a sense. water: wait -- I'm still trying to understand what an arrow is. jdl: you're looking for definite answers where there are (and should be) none. water: I thought an arrow was like a Prism model, but that seems not to be true. "what an arrow is" is exactly what an ontology answers. water: Not what a particular arrow is, what arrows are in the general sense. dynamic formal semantics system.... jdl: they are binary selectors of other arrows. 05:40pm -:- Brian [user4453@98A95AA3.ipt.aol.com] has left #Tunes [] jdl: arrows are binary relations over arrows. crap thats wrong an arrow is a binary association between two arrows (a relation is a set :) tril: thanks. water: a binary _selector_? So it selects one of two arrows? yeah whats the selector? water: or a binary association, as Tril said (it relates two arrows) it selects two arrows out of the set of all arrows :) jdl: i meant that it selects two arrows in a structured way. that's an association. Don't mean to hog chat space, but can we agree that that what I said is what TUNES is? water: okay, that's what I thought an arrow was, physically. (more) GMOL: agreed. gmol.. yes, and we have a very specific plan to do it, too. jdl: there is no more. water: (more), as in I'm going to say more alleviate problems in semantics is done through integration. which is the same thing as "everything's an object" as well as "homo-iconic". (god, i LOVE saying that!) oh. nevermind. :) water: If there is no more, you took an awful lot of space in that paper to explain it :) Tril: THat was a joke right? kind of. Conceptually, a particular arrow may be interpreted as representing a particular concept... correct? Tril: :) jdl: yes, and interpretations are provided by ontologies. Right! Okay, this is all coming together for me. Ok...clearly this involve some sort of software tat sits on a computer provides semantics and it would be nice if it was bootstrapped so that there are not too man layers that make ugly semantics/inefficiencty, yes? And ontologies are themselves concepts which are represented by arrows. too many (the big question is: how do i include this stuff in the paper?) gmol: When the system is bootstrapped, there will BE no layers. water: smaller words, more definitions water: don't forget that IRC logs are available :) jdl: you've obviously not tried to translate an informal conversation into a one-way explanation. i don't see how you can claim that already, clearly I like being able to use graphics without knowing what registers on what card are being set and how...if such a system will be able to do graphics, clearly there will be some layers. maybe it would be easier to write an AI to explain it in a conversation with someone water: No, but I have struggled (and struggled, and struggled, and struggled) to explain Prism to many people, and have yet to be happy with any of my explanations. jdl: then you DO understand water: probably :) jdl: the availability of ontologies as arrow structures is what enables "model-level reflection" water: right. That's the biggest difference between Arrow and Prism. If I may compare/contrast: (more) jdl...The inability to explain how good something is to somone is often a sign that there it really isn't that good GMOL: Try it before you judge it, smart guy GMOL: the idea of tunes is everything is an object, regardless of what layer it is in. But since you can build structures of objects in multiple layers, the strict definition of a layer seems to make less sense (becomes irrelevant). 05:50pm GMOL: Clearly explaining a complex idea is ALWAYS hard, no matter how good it is. jdl he's just paraphrasing einstein, who said if you can't explain something to a 6 year old you don't truly understand it. -:- SignOff Myrmidon^: #TUNES (Leaving) disagree with einstein, and accus gmol of being unoriginal :) jdl: ignore him, please. what were you about to say? (einstein, then, couldn't see beyond language and paradigm; but then, that's another argument.) water: Prism models and Arrow arrows are very similar. So are Prism metamodels and Arrow ontologies. The meta-metamodels are different (I use _Bit, _Stream, & _Map; you use "_Arrow" (hee hee)) but that's a superficial difference. The (more) it's true. if prism is so great at alleivating problems in software design, as a programmer i should be able to see why it so good without too much effort. THe examples you have presented are not at all convincing. THat webserver example assumed a rather large body of code which I don't know where you got from and how it was designed. biggest difference is that your metamodels/ontologies are themselves Arrow arrow/models. Prism metamodels/ontologies are a hybrid -- a set of models/arrows constraining the metamodel's scope, but natural language defining how each model is (more interpreted. (done!) Why don't you employ the services of Tortise and Achilles ? Tril, didn't know there was a strict definition of layer. What I meant was that there will be dependancies on computations that you don't understand but can use...TUNES will have these.. constraining the metamodel's scope? you mean refining the meta-metamodel? jdl: well, i think that the role of formal construction in Arrow is more powerful than you realize. Tril: No, part of the metamodel describes exactly which models may be interpreted with it. GMOL: Yes you can depend on abstractions. But the user (and authorized programs) ALWAYS have the potential to go inside those abstractions, so they CAN understand/change them. gmol: unlike current systems where reification (un abstraction) requires you to get the source code, and requires a compiler to make use of -:- _water [water@ppp-tnt-112.tscnet.net] has joined #tunes GMOL: Forgive me, but there's a number of research articles confirming the utility of little languages. I suggest that your inability to see Prism's usefulness is due to lack of experience on your part, not a flaw in Prism's purpose. GMOL: I didn't mean strict definitions of layers, I meant strict ENFORCEMENT of layers at the system implementation level. s/Prism's usefulness/Prism's usefulness if successfully implemented/ that's fine. But it doesn't change the fact that there will be dependancies, so TUNES will have layers, my original statement that you disagreed with. I.e. a kernel boundary, a library boundary, a method-invocation boundary -:- SignOff water: #TUNES (Ping timeout for water[ppp-tnt-58.tscnet.net]) 06:00pm water: What do you mean by the role of formal construction in Arrow? -:- _water is now known as water gmol, but we don't have to cluster unrelated abstractions together. A kernel is filesystems + device drivers + IPC + memory management + tasking and so on, each of these are orthogonal in tunes. jdl, name one gmol: A "layer" is a group of abstractions that are "at the same level" but we don't force the user to choose a certain group of abstractions. so you can build your own "layers" to program at. jdl: forcing all objects to be arrows and their structure to be built from arrows allows semantics to be built within the system and kept there. it allows many things to be expressed in Arrow which are "intangible" in current systems. GMOL: Don't have any on hand. Search ACM digital libary for "little language" in title, you'll find at least one water: I think the fact that Arrow ontologies are defined within Arrow is VERY powerful, but also very difficult to do. I'm not sure if its even possible. -:- Kaufmann [Kaufmann@200.224.105.231] has joined #Tunes -:- SignOff Kaufmann: #TUNES ("Nationalism is an infantile disease. It is the measles of mankind." - Albert Einstein) water: (I.e., what is the ontology of an ontology? Where does the recursion stop?) jdl: You mentioned possibility of reflection in prism. How much thought have you given to this? -:- _water [water@ppp-tnt-181.tscnet.net] has joined #tunes how about you name a title or an authour...not even one of the many you mentioned? Tril: It hasn't been a priority. Basically, you'd have to solve the same problem Brian does -- what is the metamodel of a metamodel? GMOL: Hold on, I might have a few left in my 'research' subdirectory... -:- Kaufmann [Kaufmann@200.224.105.231] has joined #Tunes -:- SignOff water: #TUNES (Ping timeout for water[ppp-tnt-112.tscnet.net]) <_water> jdl: i should express something more to you... Hey. Is this the Tiny Tunes channel? :) -:- _water is now known as water Kaufmann: No... no it's the musical OS channel :) -:- water has changed the topic on channel #Tunes to: Tunes: A Free Reflective Computing System Seriously... I found Tunes through an article on osOpinion, and it seemed interesting. The site pointed to here. water: you should express something more to me...? oh yeah, beholder's water: Just use tortuse and Achillies to translate IRC to paper... Now what is the project's focus, anyway? jdl: 'meta-modelling' as you call my ontologies may form a loop. in fact, in a complete system, they should form many loops. Kaufmann: Depends who you ask :) Not only that jdl I really don't see how you're meta meta models tie into this. Do you intened on designing a tonne of languages in that use prism as an internal format for their compilers? 06:10pm Kaufmann: we're trying to decide right now :) jdl: i.e. the meta-model of the meta-model of ... may coincide with the original domain. Kaufmann: a free software system where everything is an object, at all levels. * Kaufmann/#Tunes is getting confused... is this one of those Lisp-y thingies? :) I'm focused on writing actual code not really.. GMOL: I couldn't find any of those papers, but I'll look it up in the ACM later on. It's a little too hectic here to do it now... :) It didn't seem so, but someone started getting recursive, and I thought about it. * water/#tunes plays "Also Aprach Zarathustra" on the stereo system. water: one moment... back to the big picture. We need a system that will give formal semantics so that the computer has the ability to "work" with it's program, verify functionality and ease program design, yes? Kaufmann a lot of tunes membes like lisp, but lisp is not what we're making (i mean, Also Sprach Zarathustra...) * tcn/#tunes likes Forth So, are there any more details? GMOL: a syntax tree? GMOL: There's only one meta-metamodel, and it's only used by Prism compilers. There's only one kind of compiler, and that's Prism compilers. Prism is a framework which can encompass a number of mini-languages and mini-compilers (which are (MORE) there's a lot of utopian ideals. GMOL : And so all programs can interoperate.. <_QZ> tcn: yer nick completion is broken we're trying to put together a workable theory and make software for it. Kaufmann: Yes . It's not an OS, or a language. But if we succeed, OSes and languages will both be obsolete. run inside the Prism compiler). (DONE) Tri,l agreed Tril: obsolete is a strong word :) _QZ: huh? Tril: how so? <_QZ> tcn: do u have ansi in it? jdl if you make prism reflective "a number" becomes infinite :) jdl: "obsolete" is not mis-applied, however. _QZ: no kaufmann, it will be one unified system where you can do everything using one interface, whether you are programming or using it. Ok our method of formal semantics will likely be in the form of some sort of language that is expressable on the computer and the action of endowing the computer with the abilty to understand this language, yes? <_QZ> tcn: u have something in it thats making bitchx goto terminal color instead of the color i want the text to be jdl, waiting for those papers...must be a big subdirectory water: I'm looking back to the thing I said 'hold on' about... I'll answer in a sec * Kaufmann/#Tunes will just sit down and observe for now (that is, if you don't mind, of course). gmol: I'll do it later, too much activity here now _QZ: go figure :) GMOL: yes. Specifically, a persistent hyperlinking system, with ability to refer to uncompleted computations as objects. Kaufmann oh I thought you were going to contribute lively and informative input! Kaufmann: TUNES is open to everyone. water: you said 'the meta-model of the meta-model of ... may coincide with the original domain.' (MORE) gmol: (and communicate with all drivers as objects, and all protocols, etc) water: Perhaps, but I'm not sure that enables understanding. BAZ is a FOO BAR. FOO is a BAR. BAR is a FOO BAZ. What are these things? Tril: but I'm still kinda floating here. jdl: 42? kaufmann: IRC logs are available :) http://tunes.org/files/irc Kaufmann: Shhh! You'll give it away! ;) abi: Prism is also at http://www.teleport.com/~sphere okay, jdl. abi: Arrow? well, Arrow is a homo-iconic information representation system intended to support programming and human languages in the way that Tunes should. or at http://www.tunes.org/papers/Arrow/ Prism? it has been said that Prism is a way to program with multiple specialized, interoperable languages. or at http://www.teleport.com/~sphere 06:20pm Kaufmann: actually, here are a collection of different projects which are similar. All share some goals with tunes jdl: what i'm saying is that making information translatable between any two theories (perhaps by several steps of translation) and making the translation reversible results eventually in circular definitions at all levels. _QZ: does this look right? _QZ: or this? Now...why do you need a LLL in here or initially bootstrapping it? Argh... big words... getting dizzy... feels like reading GEB all over again... Kaufmann, you're ahead of me, i haven't read it yet.. jdl, sure. Why can't you just go the subdirectory and DCC me an arbitrary one? Kaufmann: sorry. water: okay... I guess what I'm saying is that for a CPU/OS to 'run' Arrow (rather, a program modeled in arrow), you have to define a transformation to the CPU's syntax ..and I don't know if that's possible the way you're talking. But I don't know it isn't. It's hard to wrap my brain around the concept. jdl: guess what? an ontology does that! (multiple ontologies may do that, in fact!) the cpu's ontology is included as a target for the arrow interpreter _QZ: ok now? now is probably time to start practicing programming wow! abi's right for once! water: I agree... I got confused while I was typing :) ok -:- Revned [none@mid-tgn-ngn-vty4.as.wcom.net] has joined #Tunes hi rev water: Bottom line -- good luck bootstrapping the system. Sounds hard. :) <_QZ> wow, full house OK Kaufmann you get to explain TUNES to the next person that asks :) -:- s_r [s_rr@phila-dialup598.nni.com] has left #tunes [] Hello, why do we need a LLL? Can anyone answer this? Heh <_QZ> GMOL: cuz its better jdl: thanks. rev can :) GMOL: Fare can answer it. GMOL: not all papers in directory are related to little languages. But I think I have a free moment now, I'll get it for you. * Kaufmann/#Tunes feels kinda stupid, actually i agree that we need a LLL, only that it would be sufficient to pick an existing open-source project and adapt it to our various needs. you need an LLL to get direct access to the machine LLL? i think LLL is really an executive to be used as compiler target or a HLL that encompasses an LLL LLL = low level language sorry why do we need a new LLL ? Forth ain't new gmol: we don't. Now stop asking Kaufmann: don't worry, you kind of walked in the middle of the biggest discussion we've had in ages. :) it's over 30 years old for christ's sake :) it's older than you are!! Fine, why do we need a new OS? jdl: I gather from this that this isn't a very large group. (... or something to that effect, but better-sounding.) tcn, you intended on deisnging some sort of compiler/interpeter with formal semantics in Froth do you? Kaufmann: I think everybody's here, actually :) Except Fare maybe GMOL: No existing OS is integrated, none has everything as an object, none allows all languages to interoperate, and the most important reason, because there are currently multiple OSes. TUNES would unify different environments. water: What ontology will the Arrow interpreter natively interpret? (Does that question apply, or am I misunderstanding the purpose of the arrow interpreter? IS there an arrow interpreter?) GMOL: no but Forth gives me a user interface and an extensible language, at the same time, with little effort. jdl: the ontology that everything is castable into the arrow type. MAN, my connection sucks. Let me restate: Tunes is this magical system that will represent and allow programs to talk to one another and relieve all the problems of programming...why do we need to design this thing under a new OS? * tcn/#tunes nods to kaufmann water: please elaborate * Kaufmann/#Tunes nods back to tcn (... not really sure why, but anyways.) Kaufmann: are you near Leiden, Netherlands? Tril: I think you're exaggerating TUNES (potential!) capabilities a bit. 06:30pm GMOL: I remembered the paper's title! "Little Languages, Little Maintenance." Tril: yeah, only a few thousand miles away :) I'm in Rio de Janeiro, Brazil. (God knows why, but I am.) jdl: the translation of information into arrows is probably the closest thing to an 'absolute' ontology that Arrow could have. the arrow ontology includes the basic logical relationships over arrows devoid of other meaning. GMOL: The concepts of TUNES are most useful when they are applied to the entire system. Why? that's great tcn, but I why should Forth be used as an LLL for a propsed system like Tunes? -:- Revned [none@mid-tgn-ngn-vty4.as.wcom.net] has left #Tunes [] Tril please be specific what concepts applied to what system? The OS? Kaufmann: You're connected through the netherlands, try another server (see www.openprojects.nu/services/irc.html) GMOL: Forth is expedient.. (I'm assumbing the N in TUNES is "Nevertheless") brb... one minute Tril: there are few better choices... water: how will you translate information into arrows? Is that ANY, arbitrary information? Or just specific kinds of information? Anyway, my problem is really with my connection. i'm back water: see recent question I'll BRB, people. -:- SignOff Kaufmann: #TUNES ("Nationalism is an infantile disease. It is the measles of mankind." - Albert Einstein) GMOL: TUNES is a system that allows you to integrate all kinds of information and programs GMOL: When you put the concepts of an OS into tunes it lets tunes control the hardware (that's what OSes do) jdl: if i interpret functions, propositions, and stuff in terms of relations, then arrows are the true statements within that relation, and the collection of those arrows is the relation itself. Tril...still does not answer my question, why do you need to design such a proposed system under a new OS? Tril: Right.. if TUNES didn't control the hardware, it would be slower. -:- Kaufmann [Kaufmann@200.224.105.235] has joined #Tunes * Kaufmann/#Tunes thinks it's better now. Fare sent me a postcard how nice GMOL: under an existing OS, you'd end up simulating most of a new OS anyway. why bother? GMOL we never said we did. It can be initially implemented for any platform, and become its own os later. GMOL: Linux, for example, controls the machine.. therefore it's a virtual machine. GMOL: Notice how tcn and I have contradictory opinions. and yet we get along just fine :) water: To be more concrete: how would I encode the concept of a program which printed "Hello World" into Arrow? Presumably, once it's in Arrow, it can generate machine code to print that (we'll assume), but how do I get it in there in the (MORE) first place? (DONE) * Fare/#Tunes is back GMOL: because we need CONTROL Fare: did you send this postcard? I don't see your name.. jdl: are you asking me how to give information to software? you can't be THAT naive, can you? GMOL: page tables. multitasking. protection. jdl: the system contains a model of the machine that responds to actual changes to that machine. water: No, I'm asking how you'll present that specific concept to that specific piece of softare jdl: I think the answer to your question is you draw arrows representing the program. 06:40pm Well why not develop (as in code and produce and test) under a normal OS and then once you have apeice of software that realizes all those wonderful goals, implement an OS using the software design strategies that you have made? jdl: hardware state-changes map to arrows. GMOL: that's a tactic GMOL: and we intend to do that, TOO GMOL: I started to do exactly that.. starting from scratch began to look very attractive :) that's our OTOP subproject Tril, lots of people have differing opinions on things. I don't really care if someone does. If you seriously want to get a project done, you have to afree on something. however, why *forbid* parallel development of a low-level framework? GMOL: Why? Oh so you're going to do both at the same time are you? afree? agree water: fair enough Contraidcory goals can screw up design of a single thing... water: when Arrow is bootstrapped for the first time, what ontologies will it initially contain? (If you haven't decided this yet, that's okay) jdl: state-machines are modeled in mathematics by arrow diagrams of transitions between states as nodes. GMOL give an example well, I'm not able to manage, so I let people develop in parallel jdl: the ontologies for the arrow construct, mostly. if you're a good manager, feel free to step in... water: I'm very familiar with state machines and David Harel's statecharts. I created a design pattern for coding statecharts. :) I see no contradiction If one designer wanted to build the Eiffel Tower out of metal and the other out of matchsticks...do you think we would have anEiffel Tower? I see complementarity jdl: ok. cool. :) GMOL: a little diversity is a good thing GMOL: difference: it's SOFTWARE, not HARDWARE GMOL: no, we'd have two. :) GMOL, if each Eiffel tower can build the other, why does it matter? the information is not done at the expense of other information people invest their TIME in things that they WANT to do Towers don't build things. water: Can you describe that ontology in natural language? GMOL: We want to build BOTH, eventually, anyway. It doesn't matter which one first. to change that, you must either 1) CONVINCE them, or 2) PAY them GMOL: metaprograms do jdl: sure. most of it is in the paper. (hint, hint.) * water/#tunes nudges jdl * jdl/#tunes looks innocent water: Okay, I'll just look in the paper. I should probably get going anyway. What do you have left to do, then? Fare: if you pay them, they may just pocket it ;) Well, I'll be off. When do you people come here? tcn: that's Fare's goal anyway a system of trust for distant programmers jdl: well, for my squeak implementation, i _need_ a lazy evaluator. Fare: paying someone is just another form of convincing :) water: Theory-wise, do you have anything left? Kaufmann: afternoon and evening, mostly Okay then. Bye. You guys are delusional. You have no goals. The very basic design of things like Retro are not even wrong, ask someone who has made an OS and he'll tell you that. Fare as a PhD student in CS, I don't know how you can't point that out. There is nothing concrete, you are kidding yourselves if anything useful will come of this. THe people here are full of claims not based in reality. jdl: the basic theory has been outlined, but more will come, i am sure. You are not even listening to simple reasoning..enjoy. -:- SignOff GMOL: #TUNES (Read error to GMOL[elvis.scripps.edu]: EOF from client) water: I don't know squeak, and I'm buried with my Prism compiler implementation, or I'd help Fare: I think we made him mad -:- SignOff Kaufmann: #TUNES ("Nationalism is an infantile disease. It is the measles of mankind." - Albert Einstein) GMOL smashes his modem in disgust Tril: Some people can't handle organized anarchy :) was GMOL accusing me of having no basis in solid reasoning? jdl: well, he also doesn't understand the concept of what we're trying to achieve. water: he just likes to argue with people I'm beginning to believe that's all he does water: no. He wouldn't accuse you, because he knows he doesn't understand the Arrow paper. He's accusing the tunes project as a whole ok. GMOL's basic problem is that he sees TUNES as a unified project, when it's really just a bunch of people all working on their own thing. he's evaluating Retro according to TUNES's goals GMOL: you're not completely wrong, but you exagerate jdl: I could make TUNES a unified project -- by firing everyone at sitting alone :) s/at/and/ 06:50pm If maneesh said something important, most people would miss it.. because they ignore his complaining.. I hope that other members would then split another project from the same free basis * jdl/#tunes pronounces Fare absolute dictator of all things TUNES uh-oh he already was, sort of * jdl/#tunes pronounces himself absolute dictator of everything else :) he never was! Okay, org chart: tcn: in charge of Retro hahaha Fare: in charge of TUNES Brian: in charge of Arrow Tril: in charge of type system dotted line connecting all these things ;) me: in charge of Prism Tada! We're organized! Yay! * water/#tunes is going to go vomit now, having been exposed to the concept of authority. water: didn't you go to boot camp? water: AYE AYE, SIR! >>> tcn [tcn@cci-209150250119.clarityconnect.net] requested PING 928634099 795280 from #tunes i hate authority. no, "hate" isn't strong enough a word. water: Yeesh... what are you doin gin the Navy? same here studying the center of cybernetic evil. no reason for hate, the simple logical fact that anyone who claims to be authoritative is thereby proven to be the opposite by claiming. hey Tril, I saw another review of the Matrix :) <_QZ> water: why would u join the navy? everyone knows the army is better :) * water/#tunes slaps _QZ around a bit with a large trout tcn did you see the Matrix, though/ qz: you don't get to travel in the army ? <_QZ> tcn: yes u do yeah, to bosnia and fort bragg review of the matrix uhh see the actual movie yes, see it for yourselves. * water/#tunes admits he watched it four times. Tril: no it was funny.. it was written by the guy who normally writes the gay column for this magazine.. <_QZ> tcn: u read a magazine that has gay articles? I don't read the gay articles <_QZ> sure :) it only has 1 ...and why would it matter if he did? Sheesh. <_QZ> sure :) it's a political magazine for christ's sake anyways the guy also reviewed this movie Existenz ... eheh.. sounds like something AlonzoTG would do.. abi: tunser? i haven't a clue, water abi: tuneser? rumour has it tuneser is on average a young white male hacker or programmer interested in music and rpg's. or one who strives to create the perfect CD database software haha <_QZ> water: is it scary being trapped on a boat full of horny men? :) ok, time to refine that definition Gotta go... see y'all later. Good conversations. later -:- SignOff jdl: #TUNES (Leaving) 07:00pm -:- iepos [iepos@d3.t1-6.tecinfo.com] has joined #TUNES tunes isn't all white people qz: i'm hardly intimidated by a thousand people who are extremely sensitive to hypnosis and mind-control. hey iepos you're not white are you? ;) tcn: ok, i'll remove that. abi: tuneser? i don't know, tcn ummmm......... rumour has it tuneser is on average a young white male hacker or programmer interested in music and rpg's. or one who strives to create the perfect CD database software <_QZ> well it has the word average meaning that not all are white QZ! <_QZ> Fare! QZ: I need your help for updating the asm HOWTO <_QZ> Fare: and yer asking me? <_QZ> do u maintain it? well, could you help make the "pointer" section in the end anew? Fare, I finally got tired of cellular automata; i'm playing with combinators again now :-) ... still haven't been able to get Curry's "Combinatory Logic" yet though... <_QZ> Fare: Assembly-HOWTO? <_QZ> hmm u are its maintainer <_QZ> u do know that i have never written an asm program for linux? abi: tuneser? bugger all, i dunno, water damn! Fare: I need to borrow your brain for a minute Fare: each object in the Persistent Store has a header that specifies its address, size, and flags.. and a pointer to an "external address" field 07:10pm * Fare/#Tunes sent a patch to the cvsweb author and debian maintainer <_QZ> Fare: what is it that u want me to do to it? iepos: did you see John Tromp's page? the extaddr specifies where to find the object on disk tcn: I don't understand your need to embed the "external address" in the objects (or on another machine) tcn: I'd do quite the reverse: a mapping *from external addresses* to objects!!!!! fare: yes i saw it. actually quite awhile ago when i first started playing with combinators... i don't complete understand the significance of his "U" combinator though. <_QZ> tcn: why does the object need a pointer to itself? <_QZ> tcn: it shouldnt know or care where it resides iepos: feel free to contribute to the VMs.html review page... (or the Glossary, or whatelse) qz: this is the object header, used for GC/PS <_QZ> tcn: so this header is not part of the object, internal to it that is? if the object is in RAM, 'addr' is the RAM location. If it's on disk, 'extaddr' says where. It could be in both places at the same time. qz: object headers are used for storage management <_QZ> tcn: are these headers persistent? 'extaddr' is free-format. It can be any length necessary to specify disk blocks, network addresses, etc.. qz: yeah tcn: the mapping should really be reverse most objects are not externally visible (not directly) only a handful are <_QZ> tcn: ok so an object gets loaded to ram and its extaddr is changed and the header gets saved to disk. the power then goes out and on the next boot the object is GONE Fare: wrong "external" and the mapping goes reverse Fare: 'extaddr' refers to external storage! oh now I see what else would it be? had me worried :) I thought you were going to have this pagewise? per-object disk flushing is not such a good idea, anyway page-wise disk flushing that's no problem for big objects, and small objects could have a copying GC to improve locality... yeah qz: when an object is loaded into ram, extaddr stays the same.. the object stays on disk but perhaps you're right and an extaddr field is more useful than I thought but its name should sure not be extaddr <_QZ> tcn: oops i thought u said extaddr gets changed but u did say that addr is changed it should be persistaddr or such Fare: that's too long while the in-mem address would be cacheaddr or such <_QZ> tcn: what if the object is loaded by 2 machines? why too long? 07:20pm use emacs's abbrev-mode with emacs's abbrev mode, nothing is too long it ain't my style another tactic: use a nickname, when you type, and do a search and replace globally when you're done tcn: but it's a way of commenting your files for a group project and it's a sound design. I like to keep identifiers short but not *too* cryptic at the very least, document it prominently somewhere! sure <_QZ> tcn: how do u handle that? maybe sometime I'll add a foreign-language feature where you can rename 'extaddr' to something more meaningful to you, in French or whatever. (and include a link to the documentation near the definition of the field name) qz: the object has the same extaddr on both machines qz: and it's marked as a distributed object, and handled accordingly but I haven't gotten there yet :) <_QZ> what if machine A has write access and it changes the object, how do u handle machine B notification? here's where I define extaddr currently: extaddr Address in external storage ... followed by a whole paragraph QZ: only one machine has the patch pumpkin at a time <_QZ> Fare: what? mutual exclusion mutex you lock access to the object <_QZ> do u not support write-by-one read-by-all? sure, you can do that and through a proxy you can write-by-all too <_QZ> uhh then shouldnt u be keeping the headers in the store? I already made a multiple-in multiple-out queue <_QZ> bbiaf, eating good idea * Fare/#Tunes goes sleeping instead Bye! see ya bye bye -:- SignOff tcn: #TUNES (dinner) 07:30pm tril: i'm going to write up some quick reviews of the reflective languages on the review page, ok? water suure sounds good! Maude is first however if yo just use cvs co on bespin, you can update all these reviews yourself.. well, if i could use it, then i would! someone has to do it, and I just hate having to keep writing "review by water" in the comment water: can you login to bespin? i told you your password earlier today.. right. what about ftp? can i just upload the text files that way? ok. You failed to get CVS to work at your home computer, right? Well, you can run CVS on the shell account. water: there's a very rough review page on reflection -- make it yours! * Fare/#Tunes is away You check out a copy of the tunes pages into somewhere on your account. fare: ok Then modify them however you like, and check them back in. They are in html so ftping text files to your account, then editing the html is fine tril: i know how cvs is supposed to work, but how can i make it happen? telnet? :) Water: Right. Telnet or ssh ok. let me see. ok i'm in. i'll use the instructions on the collaboration page. CVSROOT should be already set for you on the local system it is just CVSROOT=/cvs do you think i can get the files strictly via telnet? 07:40pm what, session logging or copy from the scrollback buffer? i guess.. but they are large, so it would be awkward right. i'll have to ftp them. you know, i've got about 20 windows open right now. :) if you like convenience, get a telnet with zmodem in it. i'm about to add another. say, what's the ftp address? telnet and ftp both use A records so anything that works for one will work for the other ok that worked. err... it's just one machine :) ok.. there's my source code. i should make a quick readme -:- MikeM [Administra@c72705-a.rchdsn1.tx.home.com] has joined #Tunes -:- MikeM [Administra@c72705-a.rchdsn1.tx.home.com] has left #Tunes [] done. are you putting the source in cvs? better discuss that first ok. how would i use ftp to activate cvs. You don't no, the source is still in the html directory. oh you use telnet to activate cvs 07:50pm welcome to unix. and ftp handles all the file transfers? ftp is only for sending and receiving files. telnet is only for typing interactive commands at a unix shell -:- billyboof [hatefull@nrwc-sh2-port163.snet.net] has joined #tunes right. i understand that much. hi billy so what is the question? you want to download the entire tree of cvs stuff ? or just part of it? hi water... tril: not if i don't have to. tril: i'd rather just do it file by file. you don't have to download any of it, if you know how to use a unix editor and aren't too lagged billy: i'm learning to use cvs tril: no thanks, i've had enough of ed and vi. water: that's cool... i haven't touched anything to do with coding in a couple days, been too busy.. just online to check email/say whats up... ok. -:- SignOff billyboof: #TUNES (sleep time :<) going to eat dinner, back in a bit ok, i'll muck about for a bit 08:00pm -:- SignOff water: #TUNES (The Tao has left the building!) -:- water [water@ppp-tnt-8.tscnet.net] has joined #tunes hello? hey, water, what's up? oh. i was looking for Tril, but he's still eating dinner. anything on your mind, iepos? not sure. i'm been studying combinators the last while. they seem interesting... oh, yes. i looked at them, but not very closely. i should go back and read over your arrow paper again... arrows seem interesting... i have the papers on them too. really? cool. i don't completely understand them, though.... arrows are symbols, meant to convey information, right? 09:00pm sure. they represent binary associations between objects (that happen to be other arrows) okay then... is the meaning of an arrow restricted only to statements, or can an arrow represent something else, like for instance a book, picture, or idea well, these associations can be used to bind statements to variables, and these statements can refine the definition of the object that the arrow represents. hmmm, so can an arrow represent something other than a statement, then? yes, but the idea is that arrows can always be interpreted as statements to some kind of declarational system (even one with a dynamic semantics) okay... that makes sense... so... the arrow system gives you a way of expressing information... were you hoping that automated reasoning would be easy in the system... or what other purpose will the arrow system fill as far as computers are concerned... ? well, there is a style of logic called 'arrow logic' which deals with objects that act like arrows. it is decidable, and often easily computable. but that would be only one part of a large system of semantics. hmmm... that sounds interesting... what kind of reasoning could be done with that logic? just discovering the basic relationships among arrows. it would give them no semantics either. like whether certain arrows compose like vectors do. or whether inverses of certain arrows exist. well, in a finished system, aren't you going to want to be able to tell the system "a -> b" and "a" and have the system deduce "b"... could arrow logic do that? i mean implication for "->" ... sure, you would just have to connect the arrow system to the user-interface using mappings made of arrows. if you generalize the mapping idea, you get my idea called an ontology. 09:10pm hmm... i'm not sure i completely get what you're saying... ok. umm... i should go back and read your draft again the user changes the state of the machine (hardware or software), which allows the software to respond. changes in state can be made into arrows. you say you expect a cybernetics background though and i don't have one... do you know of any introductions anywhere (preferably on the web)? principia cybernetica: abi: pcp? rumour has it pcp is Principia Cybernetica, a web-based public theory of cybernetics and systems theory oh great, no url. wait a minute... really, okay i'll have to take a look at that. i think it's www.vub.ac.be i found it somewhere with a search engine... ok i'm definitely going to read over this... it looks interesting... i would definitely convince them of modifying their theories if i had the time. or at least try... Whoa... Who am I? Where do I come from? Where am i going? :-) What is truth? What is good and evil? ... :-) hehe :) 09:20pm -:- HickServ [thrustit@209-68-229-42.dialup.cust.tfb.com] has joined #tunes buenas noches amigos como estais? what's up not much estoy bien, mi amigo there was an intense debate earlier. que bueno oh que lastima i'm sorry that you missed it. yes no, no. there were no harsh words, just great discussion. i would have been out of place like always however really? my knowledge on the subject of os design and reflective computing is rather limited well, ok. anything on your mind? not really just the fact that i have very little actual experience coding in assembler :( i have to admit that i'm getting tired, so that i won't be very good company. i've just read a lot oh well i had to pick my mom up from th airport im tired as well i found the best way to learn assembly was on a very simple computer. for me, it was the apple ][ i mean i just need more practice i still program on it: i have an emulator for windows 09:30pm but i know so little about linux syscalls well, it's easier to practice on a simple system i can't even do file io oh i did asm in dos :) oh. syscalls are nothing but memorization i liked that * Tril/#TUNES back hey tril well do you know where i can get docs on them ? tril: where can HS get docs on Linux syscalls for asm? abi pcp pcp is Principia Cybernetica, a web-based public theory of cybernetics and systems theory uh like principia mathematica? sort of. The URL is on the review page!!! abi should know it, though For PCP, not asm linux.. i think it's www.vub.ac.be what about it? any idea on those sys calls? Hick didn't krz already tell you that.. some include file to read?? and Fare gave you lots of pointers.. some Forth implementation that uses them? please don't remind me * Tril/#TUNES reminds Hickserv unistd.h isn't my idea of documentation it's poorly commented um... HickServ, you must be missing something. These are the kernel system calls. yes They are fully documented in man pages AND everywhere on the net. They ARE C Functions. blarg mehates c all you need to know is the calling convention, i.e format of args on the stack and what interrupt it is (someone said 0x80?) I don't see what documentation you would need. It sounds very straightforward. if you know asm well there are 4-5 functions with many arguments i don't know which one to use tril: i'll just keep writing up review entries and upload them later when i figure everything out. plus i don't know asm that well i don't have much experience that is 09:40pm HS: Dive in and then write a document explaining how you did it for the next person that's what im aimen for :) what do you mean dont know what function to use? You wanted to call the kernel, you got to know what one you wanted, or how would you know it's in the kernel? there needs to be more docs for the beginner tril: where are the DB files for review entries? water: none yet. It's all HTML still it be easier just to tell me the basic format for read write open close water: I'm planning to work on the DB soon. really? i remember browsing earlier and finding a bunch of text files forming review entries. HickServ: No one will tell you that. If you don't know how to read man pages, you need someone to tutor you in C. or was that Wiki? water: Well, some of them are in Wiki but probably not all.. Paul Dufresne was copying review into wiki, but stopped due to lack of interest. oh im not saying you can learn everything from a file nothing works that way great. just great. I mean, lack of interest from OTHER people. Not lack of his interest. -:- SignOff iepos: #TUNES (Leaving) no one was helping him. oh. well, why is it too hard for one person? * Tril/#TUNES thinks HS doesnt know C i know c not very well i know it better than any other language but um i've been self tought all the way i can't really take classes im not old enough :\ URL: http://tunes.org/list/tunes/9905/msg00058.html URL: http://tunes.org/list/tunes/9905/msg00058.html water: just read that, it's his post about it ok * HickServ/#tunes thinks tril is synicle cynical? no I am frustrated at a bug in the program I'm writing. hmm.. i just read it. and i remember when it was posted. 09:50pm excuse my spelling tril: is SWiki compatible with Wiki? water: i don't know. I seriously doubt that Squeak Wiki uses an internal format of 1-file-per-page, rather it would use squeak objects, no? yes, that must be true. notice we're not using the original wiki, either. it's been modified with CVS support what is the interface to the Wiki? WWW vi? (god, i hope not.) ok. i'll try to get at it. cgi, using HTML forms and POST what is the question well, the question is where to start. this effort may require re-defining old terms, i don't understand -:- Kaufmann [Kaufmann@200.224.105.249] has joined #Tunes Wow. You still here? anyone feel free to corrupt your mind and ask me a question :) i'm here Does such a thing as a TeX/LaTeX display tool for MacOS exist? i think it's called MacTeX. i have emTeX (for DOS). www.tex.org a display tool? Not the full package? hey in nasm does .something designate a label or a proc? I just need to be able to read Fare's god-forsaken paper, and the latex version was the smallest one Good grief. I'm so drowsy I can barely punctuate. Oh God, it's getting worse. Just let me know when I stop making sense. Fare's paper? Lambda and Non-Determinism Yeah, that one hick I dunno! dvi2ps ! or dview, then both are part of MacTeX 10:00pm tril doesn't know you guys :( you can also read the TeX with a text-editor (fun fun fun!) Kaufmann why don't you just read Fare's paper in HTML and save the trouble? water, yes, I really do feel like finding my old TeX reference manual. Waitaminnit, something just occurred to me. Damn. I could have sworn I had GhostScript. ok, tril. this Wiki is semi-workable. water: I hate the wiki. I never encouraged anyone to use it. That's why it died. I hate the capitalizations, and i hate using its forms in lynx. I hate how it is separate from our main page. well, what DB package can you get that generates html? and it doesnt restrict access water: I already basically decided all that have you been around or read the logs? abi mot-2 mot-2 is Programming-free DB/Web framework for Apache/PHP/MySQL @ http://tomato.nvgc.vt.edu/~hroberts/mot/ woohoo she got it right this time I'm planning to try out MOT-2 and I hope it works for the whole site. What? Kaufmann abi is an info-bot ok, so if i go ahead and create a bunch of small text files, they can be loaded in when the package is read? So I gathered i dont know what you mean "the package is read" ready. ok. yes. sorry. typo. ... and thus does my bit-per-second rate seem to deteriorate even further. Jesus christ. I don't understand. It's going at a nominal 46777 bps, and the speed is decent for everything else... except WWW. Damn Tim Berners-Lee. It /randomly stops loading files and declares that they're done/, for God's sake!!! this is your web browser? what web browser? tril: btw, you'll notice that if the topic doesn't immediately interest me, i often quickly forget. water: referring to the wiki post? brb (a minute) yes 10:10pm Netscape, but it's doing the exact same thing in iCab. It only loads Slashdot up to the header. This is very bloody frustrating. Shit, this is too much for me. I have to sleep. Good night, people. i'm back -:- SignOff Kaufmann: #TUNES ("Nationalism is an infantile disease. It is the measles of mankind." - Albert Einstein) fair enough. i'll write up as much as i can and place it in my html directory. i also volunteer to help move entries to the database when the time comes. yeah, once I have a framework, everybody should be able to help with that. ok. it'd be nice to see a web page with nicely-formatted listings of languages by different categories. or even sorting along arbitrary categories 10:20pm *yawn* hehe om real programmers don't get tired. they just get thirsty for coffee! :) (i should talk) my sister is out partying all nyte when her kerfue is midnite yeah well im sorta in a rut alonzo: girls go figure she's a GREMLIN!!! she drives a gremlin mobile THe tail pipe is about to fall off... when i get this command line crap and file io stuff done im home free now that isn't much!!! -:- SignOff HickServ: #TUNES (Leaving) 10:30pm ok. maude's done. it looks very interesting. all it uses is re-write logic and equations, apparently. "logical reflection" - higher-order logic at least. oh, wait. it must be logic about self - something possibly different. well, someone obviously figured out how to maximize Forth's power. i'm downloading papers now. 10:40pm on to Pliant. 10:50pm ok I'm satisfied with my tic tac toe program now I'm going to bed gnite. i'm thoroughly impressed with these languages. have fun with pliant ..does it run in windows? i don't know, but its gnu, so i can compile it i like maude more though never mind * Tril/#TUNES is away: (afk) [BX-MsgLog Off] 11:00pm guess my uptime =) 18 hours! hey y0 I gots to sleap... but if you have anything quick to say... sorry, i was looking at another screen i'm updating review <_QZ> 61 days great... ok. not really. I need to write my prototype qz: ? shuttin down for the nyte... -:- SignOff AlonzoTG: #TUNES (Have Nice Day :)) <_QZ> my uptime on borg <_QZ> s/on/for/ good night <_QZ> cya qz: oh you too? ok. <_QZ> me too? oh. i thought you were leaving <_QZ> oh i thought u were leaving <_QZ> u said good night nope. i'm still doing web research. i said good night to alonzo <_QZ> oh he was long gone on my screen when u said that oh 11:20pm -:- SignOff water: #TUNES (The Tao has left the building!) [msg(TUNES)] newlog 1999.0606 IRC log ended Sun Jun 6 00:00:00 1999