IRC log started Sat Feb 20 00:00:00 1999 [msg(TUNES)] permlog 1999.0220 ωνω SignOff yery: #TUNES (Ping timeout for yery[benetnash.ffke-campus.mipt.ru]) ωνω yery [ost@benetnash.ffke-campus.mipt.ru] has joined #tunes ωνω njs [njs@ip-209-133-105-127.dialup.autobahn.org] has joined #TUNES ωνω dino_ [dino@pm-ppp67.triton.net] has joined #tunes ωνω SignOff njs: #TUNES (I really am going to bed this time) ωνω SignOff dino_: #TUNES (Ping timeout for dino_[pm-ppp67.triton.net]) ωνω SignOff yery: #TUNES (Ping timeout for yery[benetnash.ffke-campus.mipt.ru]) ωνω Arduous [never@modem8wang.dragnet.com.au] has joined #TUNES ωνω Arduous [never@modem8wang.dragnet.com.au] has left #TUNES [] ωνω tcn [tcn@cci-209150250090.clarityconnect.net] has joined #tunes anyone awake? 04:40am ωνω yery [ost@benetnash.ffke-campus.mipt.ru] has joined #tunes hello hello 04:50am ωνω SignOff tcn: #TUNES (tcn has no reason) ωνω SignOff yery: #TUNES (Ping timeout for yery[benetnash.ffke-campus.mipt.ru]) ωνω yery [ost@benetnash.ffke-campus.mipt.ru] has joined #tunes ωνω SignOff yery: #TUNES (Ping timeout for yery[benetnash.ffke-campus.mipt.ru]) ωνω yery [ost@benetnash.ffke-campus.mipt.ru] has joined #tunes ωνω SignOff yery: #TUNES (Ping timeout for yery[benetnash.ffke-campus.mipt.ru]) ωνω yery [ost@benetnash.ffke-campus.mipt.ru] has joined #tunes ωνω dino_ [dino@pm-ppp11.triton.net] has joined #tunes ωνω SignOff dino_: #TUNES (Ping timeout for dino_[pm-ppp11.triton.net]) ωνω SignOff yery: #TUNES (Ping timeout for yery[benetnash.ffke-campus.mipt.ru]) ωνω yery [ost@benetnash.ffke-campus.mipt.ru] has joined #tunes ωνω Kor [kor@pm-ppp171.triton.net] has joined #tunes ωνω SignOff abi: #TUNES (Ping timeout for abi[bespin.cx]) ωνω Closing Link: TUNES[bespin.cx] by norton.openprojects.net (Ping timeout for TUNES[bespin.cx]) ωνω Connection closed from irc.us.openprojects.net: Success ωνω Connecting to port 6667 of server irc.us.openprojects.net [refnum 0] ωνω BitchX: For more information about BitchX type /about ωνω Welcome to the Internet Relay Network TUNES (from koontz.openprojects.net) ωνω Your host is koontz.openprojects.net, running version u2.10.04.resolv9.nmt.egcs4.tok.pten.tlim4.admin.upper4.whisper3.gipl.modeless7 (from koontz.openprojects.net) ωνω This server was cobbled together Sun Jan 24 1999 at 15 50:17 EST(from koontz.openprojects.net) ωνω koontz.openprojects.net u2.10.04.resolv9.nmt.egcs4.tok.pten.tlim4.admin.upper4.whisper3.gipl.modeless7 dioswkfcg biklmnopstv ωνω [local users on irc(6)] 3% ωνω [global users on irc(75)] 33% ωνω [invisible users on irc(150)] 67% ωνω [ircops on irc(16)] 7% ωνω [total users on irc(225)] ωνω [unknown connections(0)] ωνω [total servers on irc(30)] (avg. 7 users per server) ωνω [total channels created(63)] (avg. 3 users per channel) !koontz.openprojects.net Highest connection count: 28 (12 clients) ωνω Mode change [+f] for user TUNES ωνω Mode change [+iw] for user TUNES ωνω TUNES [tunes@bespin.cx] has joined #tunes ωνω Topic for #TUNES: Reflective Computing System - http://www.tunes.org/ ωνω topic set by Fare [Thu Feb 18 16:04:26 1999] ωνω [Users(#TUNES:6)] [ TUNES ] [ Kor ] [ yery ] [ Fare ] [ KeLp ] [ ^lilo ] ωνω Channel #TUNES was created at Wed Dec 30 09:08:43 1998 ωνω BitchX: Join to #tunes was synced in 7.303 secs!! ωνω abi [nef@bespin.cx] has joined #tunes ωνω smkl [sami@ppp87.dial-in.verkkotieto.com] has joined #tunes hey smurkl hello Kor-orc 12:20pm hehe, what does smkl mean? it is abbreviation of my name oh π Kor/#tunes is AWAY 12:30pm hmm. tunes isn't progressing 01:10pm ωνω lar1 [larman@208.254.224.97] has joined #tunes hello lar1 Hi 01:50pm Anyone done any corewar work? no. that is useless 02:10pm My computer is taking its sweet time to install Linux... 02:20pm ωνω SignOff lar1: #TUNES (Ping timeout for lar1[208.254.224.97]) ωνω SignOff Kor: #TUNES (Ping timeout for Kor[pm-ppp171.triton.net]) ωνω tcn [tcn@cci-209150250091.clarityconnect.net] has joined #tunes ωνω Kor [kor@pm-ppp66.triton.net] has joined #tunes welcome back ahh, everyone's asleep, eh? no do understand anything about making proofs with typesystems? sort of... 03:00pm i'm trying to understand that ... could you give a small example? ωνω SignOff tcn: #TUNES (Ping timeout for tcn[cci-209150250091.clarityconnect.net]) >>> smkl [sami@ppp87.dial-in.verkkotieto.com] requested PING 919550667 209835 from #tunes ωνω tcn [tcn@cci-209150250095.clarityconnect.net] has joined #tunes I got cut off.. :) i'm trying to understand making proofs with typesystems ... could you give a small example? 03:10pm hmm.. I guess, each type has different rules. The types are necessary to distinguish numbers from words, pictures, etc.. It's basically the same as a normal mathematical proof.. you have different rules for real numbers, complex numbers, matrices, functions... Hey, corewar isn't all useless! You ever seen that genetic algorithm program for evolving corewar programs? ick ick corewar, or ick proofs? :) corewar. anyway, do you have some example programs that have proffs with typesystem. i have one, but i don't understand it oh no.. I haven't seen anything I understand. Coq is supposed to be "good" but I haven't even looked at that yet. It runs on ML. 03:20pm which one do you have? i have cayenne. it is a real programming language unlike coq hehe, you notice how TUNES goes into hybernation during the college semester? i notice nothing, because i'm in hyperhybernation well, that's why it's not getting anywhere now 03:30pm Man, if I wasn't so busy with college, I'd have a fully functional OS & compiler by now. allright, I gotta do some homework. Take care.. ωνω SignOff tcn: #TUNES (Leaving) 03:40pm π KeLp/#tunes is back from the dead. Gone 24 hrs 28 min 27 secs π KeLp/#tunes is away: (rebooting the server) [BX-MsgLog On] ωνω _QZ [brand@p0wer.qzx.com] has joined #tunes hi QZ π KeLp/#tunes is back from the dead. Gone 0 hrs 4 min 31 secs wb KeLp 04:20pm abi: who is Thee Fairy King? wish i knew, kor abi: correct, kor! :P <_QZ> hmm ωνω SignOff smkl: #TUNES (sleeping ...) 04:30pm ωνω tcn [tcn@cci-209150250098.clarityconnect.net] has joined #tunes what's up? nuthin'.... Gakuk! privet! read my article already? which one? i heard which one was best to run X under? π Kor/#tunes is no longer away... 04:40pm abi? yes, Kor? ωνω SomeIdiot [nada@212.208.1.177] has joined #tunes abi: who is Thee Fairy King? i don't know, kor good guess :) ωνω SignOff SomeIdiot: #TUNES (-) the second about what? oh, lambdaND.. uhh, I don't have a DVI/PS viewer available now tcn: ghostscript & gv work well for linux (ggv doesnt work too well yet, which is for gnome) tcn: you can view the .tex Fare: you should generate an HTML document... π KeLp/#tunes is away: (on the phone) [BX-MsgLog On] Is there some reason you can't use HTML? Too many figures? I have all of docbook installed, I might be able to generate HTML from .tex and DVI... although I havent figured out how to use it yet... s/and/or a few symbols. yeah, I am not sure docbook supports all of those symbols... I'd like to have more figures, but no space/time 04:50pm does anyone have a URL with a bunch of algorithms, (ie: b-tree's, nurbs, etc..) I know some books.. 05:00pm I never really found any good info about such things online I should start putting together whatever I find, into an algorithm index... it would make a koo web page I could have them orginized, search, sort, graphics effects, compression, sound effects, encryption, ... π KeLp/#tunes is back from the dead. Gone 0 hrs 18 min 48 secs yeah, some *good* docs online would be nice Let me know what you find.. what's yer web page? tcn: dun have one :( 05:10pm hey, do a search for Steele's game programming page.. there was alot there.. ok, i wrote that down Some of those algorithms use fancy math.. that could be the hard part I was going to try to represent them in scheme... So there is a reference implementation encryption: statistics, probability, linear algebra. Sound: Fourier transforms. Graphics: lots of algerbra & calculus. You could just use pseudocode I dont think i have an FT library for RScheme yet, that could be a useful project too 05:20pm I think if I do it in scheme, then people wont have to go through the trouble of re-implementing them... I could also have a coresponding CVS repository, so we have good implementations Might as well do several languages.. Scheme, ML, C, Forth, ASM.. also, there are programs that will do scheme->C, that way people can make optimised versions yeah, but how many people use scheme? tcn: not enough :) how about we use VHDL hmm.. then it can be applied to an FPGA as well :) I wonder how hard it would be to program PGA's with a normal language.. tcn: not too hard, there are systems to program them with C ok, cool Tunes should be able to automatically run critical routines in PGA yes, I am doing something similar I am implemented A Parallel version of Scheme R5RS in Verilog, then I am going to maek a Netlist compiler in the Parallel ver. of R5Rs, that will do what SBS claims to have done with there FPGAs, but using Lambda-Calulus Have you used PGA's yet? no, I cant afford any :) 05:30pm neither have I I have been reading all of Xilinx's online documentation, and application reference guides for there Virtex series, which looks good I've got a chance to use them in college this semester.. older ones.. Lattice GAL series its mostly in PDF :/ I grabbed the Spartan specs today all of the PCI boards I found with a Virtex on them were VERY expensive 40,000 gates isn't too shabby.. MISC CPU's are being done with 25-35,000 transistors.. which is probably under 10,000 gates no, 40,000 gates would prolly be enough to learn all I need, although I think I might need 200k+ for all of R5Rs plus the netlist compiler I have heard of those You don't need to have the whole thing "on chip" at once.. I wonder if I can implement any sort of computer with less than 1000 gates on the FPGA's at school thats true, I only need switching code the switching code, and the current functions to be excecuted, and however many functions can be "cached" yeah Those Xilinx pga's are fast.. 200mhz+ 05:40pm this is complicated shit yeah, but its koo yup It won't be so overwhelming if I start with the little ones :) I should look for a reasonably priced PCI board... I think there is a spartan PCI dev kit for 99$ on www.marshall.com... but lookat xilinx's distributer list, it was one of those Hey, you ever soldered TQFP's or any other "miniature" components? no I know you need an 800 degree oven for some chips that are impossible to hand-solder.. :) hmm.. solder melts at 200 C I could use my fucking toaster.. haah 05:50pm that's innovation for ya my friends is sending me dragon warrior screen shots so I can highlight the area for him to search, haha haha I'm gonna go melt some solder in my toaster :) later ωνω SignOff tcn: #TUNES (Leaving) 06:00pm ωνω KeLp is now known as KeLp_awaY π KeLp_awaY/#tunes is away: (ideling) [BX-MsgLog On] ωνω SignOff Kor: #TUNES (Ping timeout for Kor[pm-ppp66.triton.net]) ωνω Kor [kor@pm-ppp46.triton.net] has joined #tunes my ISP is the devil ωνω SignOff Kor: #TUNES (Read error to Kor[pm-ppp46.triton.net]: EOF from client) 07:40pm ωνω GMOL [gmol@24.66.11.51] has joined #tunes Yellow? Yellow is probably bad Is that a bot? I guess not. Or could it be? I guess I shall flee. ωνω GMOL [gmol@24.66.11.51] has left #tunes [] 08:00pm ωνω dino_ [dino@pm-ppp51.triton.net] has joined #tunes ωνω dino_ is now known as dino ωνω hcf [nef@escher.sdi.agate.net] has joined #tunes ωνω Kor [kor@pm-ppp46.triton.net] has joined #tunes koo abi: do you remember who Thee Fairy King is? no idea, kor :P 08:50pm <_QZ> Kor: u got an OS? ωνω tcn [tcn@cci-209150250109.clarityconnect.net] has joined #tunes 09:00pm sup? awwwwwwwww yeahhh... hey, that toaster thing worked _QZ: yeah, but it doesnt even boot yet <_QZ> Kor: what kinda OS u making? _QZ: it follows the exokernel structure, it has a very small set of standard functions, and the slightly modified version of scheme it uses as a backend enforces security... I am reprogramming the backend again, to compile netlists in addition to partail alpha ASM support 09:10pm <_QZ> uh ok _QZ: netlists so that it can work off of an FPGA instead of alpha/x86 ASM _QZ: I could never d/l any source for your OS? :( <_QZ> reall? _QZ: should I connect with an ftp client? <_QZ> thats odd, ofcourse it doesnt surprise me since i never put up any source :) bah! haha _QZ: release source! <_QZ> i will <_QZ> in time time is one thing no one has :) <_QZ> i am immortal <_QZ> i have all the time in the world _QZ: bah, I would get a funny feeling if u were immortal... <_QZ> no u wouldnt cuz i never had any funny feelings about u <_QZ> this sucks hehe <_QZ> brix is growing bigger than i want it <_QZ> and what the hell was that chunky funny tasting thing i just coughed up you got it too? <_QZ> tcn? 09:20pm <_QZ> my mought tastes like rotten milk now <_QZ> mouth ah, something else. We're all sick over here. ever seen stepehen kings: the stand? yeah I think we are breeding that superflu now of course, this one is real, and not created by god, but mans irresponsible use of medicine Wasn't there are flu in about 1920 that killed 20% of the world population or so? s/are/a yeah shit happens but these antibiotics, that's just asking for shit to happen :) at least when people are invlolved, it does Hey, the toaster worked great for melting the solder. Only took 1 minute. koo hey QZ, you ever build anything w/ surface-mount? 09:30pm They have some wicked small chip packages coming out now.. some of them are just a bare chip.. >>> tcn [tcn@cci-209150250109.clarityconnect.net] requested PING 919575208 from #tunes 09:40pm <_QZ> uhh duh <_QZ> they have always had bare chips why didn't they use em? <_QZ> thay have I guess they're a pain in the ass though ωνω Stephan [xy2k@lawnas1-30.ionet.net] has joined #TUNES If anyone is interested in trying out a new server, connect your client to nonet.dynip.com ωνω Stephan [xy2k@lawnas1-30.ionet.net] has left #TUNES [] <_QZ> rip open any device and it will most likely have them <_QZ> they are covered with black epoxy oh yeah, I've seen them in a few cheap-ass toys <_QZ> those are custom asics right <_QZ> well brix has just gone slightly backwards So have you soldered any QFP-type chips? <_QZ> i added package objects what are they? <_QZ> qfp? <_QZ> object categories supported by the kernel: data, action and package tqfp, plcc, etc.. you know those chips with tiny close-spaced leads on 4 sides? <_QZ> package objects holds several objects inside <_QZ> like a tar file 09:50pm <_QZ> i am using them so the kernel can load drivers faster yeah, package objects sounds like a good idea, when you have several small objects <_QZ> since a hard drive has to have 3 objects loaded to work <_QZ> config data, init action, and load action <_QZ> and save action if anything on it is modifed <_QZ> and since the kernel must create a thread to load each object, that is 3-4 threads to load a driver like that <_QZ> but only 1 for a package ouch <_QZ> i will also add a feature that detects action usage and will group actions that always get loaded together into packages <_QZ> s/detects/tracks/ Why does each object need its own thread? <_QZ> each action does <_QZ> actions are messages to do something to an object <_QZ> the object data they work on is in an isolated memory space and the kernel creates threads for the actions in that isolated space my kernel modules work in a totally different way, I guess. Most of them so far just install an IRQ handler.. <_QZ> huh 10:00pm Each module is loaded, linked, and executed.. no threads at all it's in the early stages though ωνω SignOff tcn: #TUNES (Leaving) ωνω tcn [tcn@cci-209150250109.clarityconnect.net] has joined #tunes <_QZ> u have no threading? not yet Don't need it for the kernel 10:10pm <_QZ> but userspace isnt the kernel I haven't added that yet :) mirc just quit on me.. wtf was the all about.. haha <_QZ> my kernel code is run inside the timeslices for each user space thread that calls the kernel <_QZ> so its just reentrant code protected by semaphores So it's more like a library <_QZ> a library with ring 0 perms Yeah. Same idea with mine. <_QZ> and boy is it a bitch that is exo-kern structure whatever you call it <_QZ> no it isnt <_QZ> its uK uK has a resident kernel <_QZ> i have a kernel uK has a thread for every kernel object, right? slow as hell? I guess you could have a resident kernel and a ring 0 lib but since you wrote the HLL, you dont need a resident kernel your ring 0 lib, could just include the context switches, and such ie: put an immutable context switching function in the ring 0 lib, have each app call it after a timeslice <_QZ> what <_QZ> thats coop MT yeah <_QZ> i am not doing coop 10:20pm pre-emptive? <_QZ> brix is a heavily threaded system <_QZ> it depends on preemptive coop inside a preemptive thread would help though have you checked out the ExoPC source, I bet they did that cut down on the number of threads they have a CVS tree... Im not sure that getting rid of the kernel for BRiX would be beneficial, but it sounds like it is well modularized, so it might be worth a try <_QZ> there is no protection without a kernel you can have a secure Exokernel, the ExoPC project is a good example, although it uses a.out binary format, and is bloated tunes has a link to it, called aegis i think <_QZ> remember that i am an asm coder. i like to be free while coding and not restricted by some HLL. without a kernel u need protection to be handled by the HLL _QZ: ah, u r correct _QZ: although I think the ExoPC project accomplished safety executing any binary, but it might have some overhead switching processor modes, or something or if u had a dynamic dinary loader that restriced, or modified some calls... s/dinary/binary <_QZ> impossible why? 10:30pm <_QZ> if i code my programs in asm there is no way any loader can tell what it will do you have to copy the program to memory before it can be executed, then filter out out/in's, int's, and some regs <_QZ> it dont work that way why not? ExoPC is a complete system, and I havent heard any security complaints yet <_QZ> u cant just scan the binary for in/out instructions since those bytes might be data _QZ: but your binary format would seperate data/actions you OS still can have a binary format, and compile ASM code to it no reason to accept flat binaries <_QZ> i dont think u understand asm it might work _QZ: I have done alot of ASM programming in the past _QZ: and I understand that most binary formats allow data/code sections for that very reason <_QZ> do u know what constants are? yes <_QZ> immediate data? yes mov BIGREG, "immediate" <_QZ> that isnt isolated by binary format <_QZ> its in the binary code but your target is always known you dont have to filter out immediate sources only restricted targets <_QZ> and preceding an in/out instruction with the right byte can turn it into a totally different instruction <_QZ> so i/o protection must be handled by a ring 0 kernel well, I would still look at the aegis link <_QZ> the protection, allocation of ports and deallocation I doubt they would have implemented it, if it wasnt possible to be secure <_QZ> the admin adds binaries that can use io to a list that the kernel checks before allocating io see ya cya tcn _QZ: that would work at administrative cost ωνω SignOff tcn: #TUNES (Leaving) 10:40pm <_QZ> what cost maintaing a list of I/O requiring programs <_QZ> not programs, drivers its minimal cost though, not a bad way to do it <_QZ> when the driver is compiled the OS prompts the admin that it uses io and what ports and if those ports will conflict with any other drivers and its good to restrict I/O to drivers, so you dont end up like linux, making X directly access the video card and taking down the whole system <_QZ> in brix X would be a driver hopefully a system more like GGI though <_QZ> yup does brix have a video driver system yet? <_QZ> yep what about sound? <_QZ> nope I think ALSA would be a good model to look at for that, although sound drivers are even less compatible then video drivers these days <_QZ> what is alsa? advanced linux sound architecture... like the GGI of sound, it already works, and even has OSS compatibility http://alsa.jcu.cz, they have a CVS tree too <_QZ> kewl 10:50pm π Kor/#tunes is now (gasp) AWAY ωνω SignOff Kor: #TUNES (leaving) ωνω SignOff yery: #TUNES (Ping timeout for yery[benetnash.ffke-campus.mipt.ru]) ωνω yery [ost@benetnash.ffke-campus.mipt.ru] has joined #tunes !gibson.openprojects.net!! Remote CONNECT hogan.openprojects.net 8005 from lilo ωνω SignOff dino: #TUNES (Ping timeout for dino[pm-ppp51.triton.net]) ωνω SignOff _QZ: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep) ωνω SignOff hcf: #TUNES (Leaving) ωνω ruinerJK [nate@ppp159.wi.centuryinter.net] has joined #tunes where can I get asm? hi btw Nasm....sorry ωνω ruinerJK [nate@ppp159.wi.centuryinter.net] has left #tunes [] [msg(TUNES)] newlog 1999.0221 IRC log ended Sun Feb 21 00:00:00 1999