IRC log started Thu Apr 29 00:00:01 1999 [msg(TUNES)] permlog 1999.0429 -:- SignOff _QZ: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep) -:- An [atu6@mwimba.line-pro.es] has joined #Tunes -:- An [atu6@mwimba.line-pro.es] has left #Tunes [] -:- An [atu6@mwimba.line-pro.es] has joined #Tunes -:- An [atu6@mwimba.line-pro.es] has left #Tunes [] -:- Connection closed from king.openprojects.net: Success -:- Use /Server to connect to a server -:- Connecting to port 6667 of server king.openprojects.net [refnum 0] -:- BitchX+Deb1an: For more information about BitchX type /about -:- Welcome to the Internet Relay Network TUNES -:- Your host is king.openprojects.net, running version u2.10.04.resolv9.nmt.egcs4.tok.pten.tlim4.admin.upper4.whisper3.gipl.modeless7 -:- This server was cobbled together Sat Jan 23 1999 at 21 33:38 EST -:- king.openprojects.net u2.10.04.resolv9.nmt.egcs4.tok.pten.tlim4.admin.upper4.whisper3.gipl.modeless7 dioswkfcg biklmnopstv -:- [local users on irc(7)] 3% -:- [global users on irc(89)] 35% -:- [invisible users on irc(166)] 65% -:- [ircops on irc(12)] 5% -:- [total users on irc(255)] -:- [unknown connections(0)] -:- [total servers on irc(34)] (avg. 7 users per server) -:- [total channels created(76)] (avg. 3 users per channel) !king.openprojects.net Highest connection count: 16 (15 clients) -:- Mode change [+f] for user TUNES -:- Mode change [+iws] for user TUNES -:- JOIN activated by "TUNES #tunes tunes@bespin.cx " -:- TUNES [tunes@bespin.cx] has joined #tunes -:- Topic for #TUNES: reflective OS at it's best: http://www.tunes.org -:- topic set by Downix [Wed Apr 28 04:46:23 1999] -:- [Users(#Tunes:6)] [ TUNES ] [ Tril_ ] [ abi ] [ Fare ] [ Tril ] [ ^lilo ] -:- Channel #Tunes was created at Sun Feb 28 08:48:06 1999 -:- BitchX+Deb1an: Join to #tunes was synced in 7.804 secs!! -:- Mode change [-s] for user TUNES -:- AlonzoTG [Alonzo@client-151-200-125-173.bellatlantic.net] has joined #tunes -:- NetSplit: lucas.openprojects.net split from koontz.openprojects.net [08:26am] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [lucas.openprojects.net] -:- Netjoined: lucas.openprojects.net koontz.openprojects.net -:- AlonzoTG [Alonzo@client-151-200-125-173.bellatlantic.net] has joined #Tunes !carter.openprojects.net!! Remote CONNECT zheng.openprojects.net 8005 from lilo om topic? topic is http://tunes.org (Reflective Computing) http://tunes.org/papers/Arrow/ (Arrow paper available!) http://tunes.org/~dem/tunes/Specifications.html (Tril's draft specs) 11:50am -:- SignOff AlonzoTG: #TUNES (Have Nice Day :)) -:- s_rr [s_rr@phila-dialup056.nni.com] has joined #tunes -:- s_rr [s_rr@phila-dialup056.nni.com] has left #tunes [] -:- GMOL [gmol@24.66.11.51] has joined #tunes -:- GMOL [gmol@24.66.11.51] has left #tunes [] -:- tcn [tcn@cci-209150250096.clarityconnect.net] has joined #tunes -:- SignOff tcn: #TUNES (tcn has no reason) -:- tcn [tcn@cci-209150250096.clarityconnect.net] has joined #tunes buenas dias! -:- hcf [nef@me-portland-us103.javanet.com] has joined #tunes howdy hcf 01:00pm hoy tcn sup? oh, just chillin how's the lighting stuff going? no change, havnt practiced hehe.. I hear half the high schoolers around here having been playing hookie all week :) "there's a bomb threat!" 01:10pm * Tril_/#TUNES is back from the dead. Gone 22 hrs 1 min 27 secs anyone else want to quit? hehe I'm sticking it out :) sure you are, you caused all th is this 01:30pm hmm yeah, it would have been fine if nobody ever wrote any code :) you did some work, causing everything to be messed up well, GMOL did say the tunes project should be disbanded if no progress happened soon Maybe if I'd pretended to be a CS professor I woulda got more respect :) maybe if you keep working you wont need respect since you will have a working system. "I have a PhD, I know everything" did you get the insane babblings I just sent out? hmm nto yet I get plenty of respect, really. I only have 2 outspoken critics :) just more symptoms of not having any hll code. i'll change that soon. I have to read this arrow paper first, though. Hey, maybe we can decipher it if we work as a team what don't you understand? 99% of it jk I haven't read 90% of it 01:40pm -:- Tril_ has changed the topic on channel #Tunes to: must read arrow paper: http://www.tunes.org/files/papers/Arrow_Introduction_v8.htm NOW!!!!! I skimmed through it yesterday, read a few parts that looked interesting.. there's a lot of jargon though so tell me what `'s it for huh? whats the arrow system for i'm quizzing you just quoting from the abstract would be sufficient oh, it can represent *anything* really :) is it really that confusing? it's transcends the state-of-the-art? come on that's the kinda thing that bugs me have you thought about why current systems are so braindead? isn't that why you're in TUNES? yup so why do you think they are so "dumb" and what can be done to fix it? Well, I did start reading about cybernetics.. Wiener's books are really concise and readable.. changing the subject? i said what do you think, not what wiener thinks hmm.. the math & cs jargon will take a lot longer to learn 01:50pm well forget about the arrow paper right now. what do YOU think is wrong with current systems? ok hang on, I'll read from my notebook from last year, before I ever heard of tunes :) ok. Allright, I said they're too slow, bloated, inefficient.. needlessly complicated, counterintuitive, ugly, forbidding... built up over the past 50 years without a fresh start.. ok.. but why? there have been fresh starts. but they had the same problem. as a system evolves it ends up as you describe There were some good developments, but they didn't get too far.. bad management, etc.. * Fare/#Tunes is back take Apple, for instance. They made everything proprietary, which ruined their market share. Plus, they got bloated. take lisp machines -- greed and arrogance i'm not thinking of marketing decisions, just system design ones (cybernetics) nobody ever designed a cybernetic computer system?? tcn: when you have a PhD, you know that you know next to nothing but you still pretend otherwise? uh? hehe not YOU , personally, fare 2 more years and I could have a BS degree.. we all know what that stands for.. Whatever you do will be insignificant, but it is very important that you do it. -- Gandhi BS -- BullShit? 02:00pm * Fare/#Tunes ordered "Categories for the Working Mathematician" by MacLane from Amazon.co.uk so that I can give names to all these supposedly common situations that arise when studying implementations I forgot what PhD stands for physical doctor philosophy Phucking Dweeb I mean the students' definition :) PUsh Here Dummy anyway oh yeah where were we? you both must read the arrow paper now! * Fare/#Tunes recommends that all diplomas be automatically granted to everyone who reaches the age six, so that real work and study may take place afterwards * tcn/#tunes nods agreement why can't anyone at age six do real work and study? maybe age three, or even birth, is a better idea, tho -- the first years are particularly important Tril: I'm not fluent in that language yet you don't need to be. just read it over and over until you get it. that's what i did for the tunes pages at least you will have it memorized * Fare/#Tunes fetches some bodily food one of my professors explained to me why my paper was rejected at Reflection'99: aaaaa! I get it! "when a french researcher has three ideas, he writes a paper. when an american researcher has one idea, he writes three paper. when a french researcher submits a paper to an american conference, he must not only use the english language; he must also use the anglosaxon style. or else, his american readers won't be able to understand his paper. you have three ideas? according to what you just said, you have to write 9 papers. 3 for each idea Tril: exactly isn't that a book? that is, if I write in american style :( 02:10pm it's an American conference being held in France? haha Tril: indeed it doesn't matter _where_ the conference is held, only _who_ is in the committee board. I like the WWW better. no committee board. or rather, everyone his own committee board -:- SignOff hcf: #TUNES (Leaving) I think I undetstand the arrow paper. But I already undertstood it before I read it, so that doesn't necessarily mean the paper explains it well. Brian's assuming the readers basically know everything he does you mean he chose an ontological frame that you don't understand? we need model-level reflection to translate his ontology into yours :) ahah I used to think we had the same idea and we just couldn't communicate it to each other, but I guess we had different ideas. I still don't understand "ontological" oh what's your idea? Fare: s/ontology/theory/ In formal terms, an ontology is the statement of a logical theory. Tril: in your sepcs, I don't like the idea of "moving an object from a type to its completementary" that mostly defeats the purpose of typing! my idea was just a simplified programming concept... unless you're doing something like LISP -- trivial typing Fare: should objects be allowed to have more than 1 type? tcn: there are lots of possible type frameworks tcn: most "typed" programming language have a notion of "principal type", tho 02:20pm Fare: but what do *you* like? tcn: i.e. a one privileged type from which all other valid types can be deduced Fare: I only mentioned that in the Reference section, which I said to ignore. tcn: I like the ability to dynamically move between multiple type systems hey, we coulda downloaded nuclear bomb specs over the net :) you will see what I meant after I make the system i'm not sure what you mean to defeat the purpose of typing i mean the type and its complement together (unioned) equal the supertype typing = giving an abstract invariant on an object tcn: you still can't? or in your library? I don't understand "homo-iconic" Fare: me either :) is it what we've been calling "uniformity" throughout years? I'm hetero :) homo= the same icon=symbol . It means there's one primitive that the whole system uses. (An Arrow in this case, in my case membership in a type). he coulda just said that you have to know latin (or at least enough latin) 'icon' has so many meanings now.. too vague Fare: invariant is only relative, since I repreesnt all dynamism in terms of "change of type" :-~ a type is invariant *at that moment*, and types are used to define state. But when state changes, type also changes (because they are the same) what's a :-~ :-& pretzel >>> Tril_ [dem@bespin.cx] requested PING 925421387 930325 from #TUNES 02:30pm good, you're reading the arrow paper. s/arrow paper/email/ how the heck is the system supposed to foresee the consequences of using its own ontology and adjust it for the relativity of ontologies tril: who are you asking? Brian :) i'll ask in email. Fare: whats :-~ the notion "homoiconic" is flawed. Even in lambda-calculus, you have a three-entry grammar uhh even with SKI (or just SK), you need one binary construction, and two or three nullary constructors but everything is a lambda expression the text layout for arrow* is annoying Tril: every lambda-expression is a lambda-expression, indeed but then, in any system FOO, all FOO-expressions are FOO-expressions, too! are you trying to imply that this revelation means nothing?? it's like saying "everything is an object". It doesn't have a definition, it is the one and only premise. rather, that whatever valid intuition behind would require much narrower formalization before it become any useful I wonder if arrow* is something Brian just discovered, but others have know for a long time. I've done that sort of thing. well, whoever knows it, they didn't make it widespread for instance, some deisgn choices could be related to a particular reference for K-complexity of systems I'll have to re-read arrow* after some looong sleep hehe oh shit, I got a calculus test tomorrow 02:40pm quick integrate 1/(x^6+1) dx what do *you* call "calculus" ? -:- GMOL [gmol@24.66.11.51] has joined #tunes maneesh? hey GMOL finish that cancelled letter yet? Fare: like what tril just said hey maneesh no...hey guys perhaps dicussing here would be better? (just finished my cell bio exam) I'm half asleep, but go ahead oh yeah it's 1am there :) no, only midnight Tom, I downloaded your latest version of retro, comments regarding GDT and IDT obviously don't apply...but I see no multitasking suppourt or memory allocation, am I correct ? gmol: correct, afaik giveup? my calc says (sqrt(3)*ln((abs(x^2+sqrt(3)*x+1)/abs(x^2-sqrt(3)*x+1))/12 + arctan(2x-sqrt(3))/6 + arctan(2x+sqrt(3))/6 + arctan(x)/3 gmol: all multiprogramming will have to be done in the bootstrapped HLL :) Tril: which is your calc? Tril: maple? actually, I'm borrowing my friend's TI-92 GMOL: Multitasking is my next project oh and just where are you going to take care of your mutual exclusion routines? in the HLL? Tril: Try int 1/(x^6) as an approximation :) tcn: we had to do that in homework my next paper proposes a formal account of the notion of implementation; next paper illustrate its use, and thrid paper shows how to deduce a "metaobject protocol" from it GMOL we'll have to have atomic things in asm; the rest in HLL (ultimately, the HLL compiler will manage the asm alone) (but not at first) GMOL: There is memory allocation. Everytime an unallocated page is accessed, the pagefault handler allocates it :) I'll improve it later, when I'm working on the persistent store. how atomic? do you understand the tradeoffs associated with at what point you abstract (semaphores messaging, pipes etc.)...and you really can't make a decision on your multitasking until you've decided on your memory model, how is that going to look? GMOL: the LLL is really an executive to be used as compiler target hmm indeed, the choice of memory model is foremost in it GMOL: Use an algebra of implementations to determine the most efficient at the time it's also an interactive debugging environment for really low-level subsystems 02:50pm I mean in terms of paging, segmented, what are you're boundaries, do you know what appropriate size limits are, how will this affect mutlitasking? i.e. DON'T make tradeoffs. Specify what you want and let the system make tradeoffs Tril: that's possible when bootstrapped Tril, a simple ukernel written in assembler can swtich entire memory models dynamically. the initial bootstrapper will have wired-in tradeoffs tcn? tcn is gone. or doing his project for his own education. he's reading email GMOL: the "ukernel" can, but the question is: can the rest of the system? GMOL: ever read my glossary entry on ukernel? You're telling me that you are going to implement an algorithm that will dyanimcally change between approaches like shared memory and say segments? GMOL: eventually, we should all this written in assembler no no, the initial asm bootstrapper will have wired-in policies I'm using Forth now too At least that's portable are you guys (tcn, Fare) still designing a scheme interpreter ? yup, or something like scheme, anyway anyone know much about DKlisp? can't do it on the 386 my friend...how you setup your mem model will greatly affect multitasking routines...if you're using shared mem, you'll probably have state variables (usualy in epxected places)..you just can't dynamically switch to a segmented model.. and if the bootstrappers job is only to load the HLL, why are you bothering with this? actually.. (mem management, multitasking etc.) the HLL needs it you can if your compiler lets you recompile your programs while they're in memory -:- SignOff Tril: #TUNES (Ping timeout for Tril[140.160.164.200]) * Tril_/#TUNES is still here You can use segmentation on top of paging, on the x86. I don't see the problem. segmentation seems pretty useless, though :) GMOL: something to do, while he waits for the HLL design (me)? I'm talking like say having system calls accessed through execute only pages, and then through selectors...you expect to dyanimcally switch between the two? not anytime soon? So tcn what does your memory layout look like? well, I'm not too interested in segmentation, because most architectures only have paging GMOL: stop asking me, it's pretty vague.. it'll be whatever it needs to be. what my question? 03:00pm GMOL: mem model welll you'll have to know before you implement xtasking.. not only that have you considered how your choices will effect the implemntation of parallel/distributed computing? -:- Tril [dem@140.160.164.200] has joined #TUNES I'm keeping my options open. you have't even considred these things yet...the reason I'm pounding you with questions is becuase you don't realize what's invovled here anyone can make a smple uniprocesser kernel when you get to more advanced issues, it takes time to understand the issues... tradeoff's etc. GMOL: we *can* dynamically switch model. Only this will obsviously not be a "realtime" kind of operation And I garuntee you that no man alone can conquer(Fare just a sec lemme finish with tcn) making such an OS on his own... GMOL: it's more of a save&restore kind of thing and without seriously considering these implications before hand you are doomed I've considered them look at dos, windows...we'll never have clean environments under them due to the root of how they are designed.. I'm still considering.. it's new territory GMOL: you still don't get it that once a LLL is written in the HLL, we throw out the LLL we wrote in asm. Yeah, we *rewrite* the bootstrap & system stuff at high level, and throw out the old prototype/bootstrap stuff WAell starting off about a bootstrapped kernel is not the hard part...do you know the amount of research that goes into issues in these types of things? lot's and there are smarter people than you and me behind it... SO you're going to make a half baked prototype, design a new HL under that prototype, then throw out the prototype? HLL HOWEVER, I don't have the burden of writing 3 papers about every idea I come up with! GMOL: basically. academia is (mostly) a joke as is industry, and government 03:10pm three cheers for the nonprofit sector Now explain to me why you're going to do that instead of starting with a perfectly fine exisiting environment? GMOL: Like I said. Because he's bored. GMOL: name a 'fine' existing one just like you, he thought TUNES looked like it was going nowhere. So he decided to do something, even if it wasn't very useful. it's for programming experience more than to benefit the project (tcn let me know if I am off track) tcn: That is an ignorant statement. There are stupid poltics in acacdemia jsut like everywhere else...but there are some frightengly intelligent people out there that come up with truly beautiful ideas the system is flawed, the people are good a few of 'em everywhere else, like the tunes list? Computers wouldn't be around without it...neither would all this neat stuff in OS's...a good system, just about any system out there has lots of tools to help you compile analyse and do stuff without having to write it yourself (windows, linux, windows NT...) We could have made much better progress with totally different social/academic institutions. It's pretty amazing that the system we have accomplished what it did. :) Tril: I'm doing Retro to learn, but I hope it'll become useful, after lots more work. just sec I'm talking to my dad on phone got your mail, tril 03:20pm Fare: any comment on the mail I sent to the tunes list? ril: which? which is what oh, I like it tunes is a non-organization with no single direction sums it up good organizations are inherently flawed subject: Re: This project is going [multiple directions, ... Tril: looks fine Tril: I have nothing to add or remove it although something is essential for the convergence: some kind of spirit and also, I wish we were further so stop slacking hmm i wonder if jiml read the arrow draft tcn: still around? is anyone willing to read about my implementation stuff "live"? on irc? sure 03:30pm somone wanna answer my question as to why use a half baked OS to develop on? Fare: yeah Fare: go ahead One person's learning experience shouldn't hinder a whole group gmol, i'm designing a hll entirely separate from retro. It probably won't be built on top of retro. GMOL: you can ignore it :) just beacuse retro is actively being developed in the open, don't mistake it for where the project is going So now you're saying that HLL won't be impliemnted under retro? over that is Retro is just for trying out ideas GMOL , fare and tcn want to make a HLL on retro. thats fine. but I am working from the top down. That is, I'm designing a HLL first (almost done) , then AFTER it is designed, someone will decide what to implement it on. you have a compiler or a specification ready? i have what fare calls "Design notes" at http://tunes.org/~dem/tunes/Specifications.html you can tell me what you think GMOL: I'll use your 0S/2 if it allows reflective modification tcn...you maybe taking this personally, and for that I am sorry, but you haven't even begun to adress real issues in OS design, and it will be a matter of time before you realize that you can't do something of the magnitude you are proposing, nor do it well enough to be better than what is available...I am being objective, and reasonable, there isn't much point in designing a half baked ukerenel... Referring to the half baked ukernel I wrote? so, we start from operational semantics, aka transition systems, aka categories: a space of "objects", with a "reduction relation", a system of "states" with "transitions", or a category of "objects" with "morphisms" (three vocabularies, one idea) you can also use ASM or Rewrite Logic -- that's equivalent GMOL: You gotta start somewhere. This is my 2nd generation of Retro.. the first one was just a 16-bit bootstrapped forth Fare: OK so far GMOL: 3rd generation should be distributed so, to implement an abstract category Abs with a concrete category Conc, Fare: how about types and functions? :) 03:40pm chose a subset Obs of observable states in Conc Tril: yeah, forgot that one -- categories are also isomorphic to typesystems :) Obs is given by a canonical inclusion (aka a monofunctor) j:Obs->Conc what's that mean? now, the implementation is defined by an interpretation function phi:Obs->Abs it means that we can identify objects in Obs with their image by j in Conc so the set of implementations is given by quotienting the interpretation function spiff and that the arrows in Obs between o1 and o2 are "the same" as the arrows of Conc between j(o1) and j(o2) what's an observable state? Tril: no, every implementation has its own phi and j an "implementation" of an abstract system Abs in Conc is the data of Obs, j, phi. You won't get there without...people spend entire careers working in something like distributed computing...I don't think you've done this stuff long enough (you yourself stated that you've only worked with OS's for a year). What you are proposing is building a kingdom from a sandcastle...believe me. It's like saying well the next sandcastle will have feature xxx, I start and redbuild evrytime, eventually it will be a castle. Meanwhile other people are reallly bui without learning a lot more that is the most important condition we want is called "safety". In category vocabulary, it says that phi is a functor one doesn't learn anything about really building castles from working with sand. it means that you can complete a diagram phi(o1)=c1,phi(o2)=c2,o1->o2 with an arrow c1->c2 i.e. if we can observe some concrete behavior, it is a correct abstract behavior. what's an observable state? Tril: an observable state of the concrete system is a state where you can say "look! we have implemented such state of the abstract system" it's some kind of "safe point", to which you can give a valid abstract meaning whereas the concrete systems may go in lots of intermediate, non-meaningful states. is the abstract category like the source (specification) of a program? and the Conc is like the language it gets compiled to? the abstract category is the operational semantics of a language/program that we want to implement a non-meaningful state is one that isn't expressible in the given HLL? take, for instance, the operational semantics of the join-calculus 03:50pm don't know join calculus (operational semantics are more interesting for non-pure languages) ok, of LISP, then take an abstract LISP interpreter well.. i know lisp, maybe not its operational semantics it's a step-by-step description of what state the system takes i.e. we start with such state of memory/registers, and evolve according to such rules. it has environments, and stores yup, so a "state" of LISP will be the data of an expression, a environment, a store, a continuation. -:- SignOff tcn: #TUNES (dinner) (and auxiliary registers) (beep me several times by /msg when ready) so our concrete system will also have memory, registers, whatever and we want to express the fact that somehow, it implements the abstract system so an observable state is one of the states in the concrete system which is reachable from the abstract system? within the concrete system, we find a subset that will model the abstract system as faithfully as posssible this will be the subset of observable states. no; there may be more reachable states than are observable thre may be lots of "transient" states that will be reached, to which no precise abstract meaning can be given so in-between operations of the abstract system? the abstract system has NO in-between operations if I interrupt a cons operation, that is unobservable? but if I look at the concrete state right after a cons operation, it is observed? but when you implement it with Conc, you generally cannot avoid them. YET, you want to preserve atomicity of operations, and all the properties of the abstract system as long as it is only observed at kosher observable states (beep me) ok I guess I understand that part. what is the "faithfully as possible" part? 04:00pm Tril: it is that depending on requirements, we may have several kinds of implementations often, you don't want all the conditions to be fulfilled, but only "enough for the sub-problem at hand" you mean resources like free memory and optimization? there are many possible implementations which one is better? i.e. you don't want to completely and faithfully implement the infinite set integers, when your computations will never overflow from n-bit registers with n konwn in advance dont confuse infinite integers with unbounded integers. I do want unbounded integers implemented correctly. and, anyway, when we implement infinite abstract systems with finitely resourced computers, we can't be complete and faithful Tril: there will always be a limit on the integers you can implement on your system -- because the memory is finite if I tell teh system to keep growing the integers, it should tell me I need to add more ram yes, yes, let me come to that point so the most important criterion is safety: any observable abstract behavior of the implementation must be a correct behavior in the abstract system. it's a condition on phi now comes soundness: so you are saying that if I give you an observable state, you will tell me what abstract state it corresponds to? the implementation can always reach an observable point, from wherever it is Tril: exactly is that one of your functions? Tril: that's the definition of an observable state phi? phi: Obs->Abs associates the abstract state to the observable concrete state what is the compiler function ? is phi the inverse of that? 04:10pm yup, the compiler is the "inverse" of phi surely the compiler was written first because phi is not necessarily injective, phi would be much more complicated the compiler function is nondeterministic the compiler gets to choose a valid concrete state to implement an abstract state actually, you always write the compiler and phi at the same time at least in your head ok and the point of all this? well, we can give other conditions: soundness, weak or strong liveliness, realtime w-or-s liveness completeness, faithfulness, etc so that we can specify what we expect from an implementation 04:20pm -:- Use /Server to connect to a server -:- Use /Server to connect to a server -:- Connecting to port 6667 of server irc.us.openprojects.net [refnum 1] -:- BitchX+Deb1an: For more information about BitchX type /about -:- Welcome to the Internet Relay Network TUNES (from carter.openprojects.net) -:- Your host is carter.openprojects.net, running version u2.10.04.resolv9.nmt.egcs4.tok.pten.tlim4.admin.upper4.whisper3.gipl.modeless7 (from carter.openprojects.net) -:- This server was cobbled together Sun Jan 24 1999 at 15 50:17 EST(from carter.openprojects.net) -:- carter.openprojects.net u2.10.04.resolv9.nmt.egcs4.tok.pten.tlim4.admin.upper4.whisper3.gipl.modeless7 dioswkfcg biklmnopstv -:- [local users on irc(14)] 5% -:- [global users on irc(111)] 36% -:- [invisible users on irc(197)] 64% -:- [ircops on irc(11)] 4% -:- [total users on irc(308)] -:- [unknown connections(0)] -:- [total servers on irc(34)] (avg. 9 users per server) -:- [total channels created(81)] (avg. 3 users per channel) !carter.openprojects.net Highest connection count: 22 (21 clients) -:- Mode change [+f] for user TUNES -:- Mode change [+iws] for user TUNES -:- JOIN activated by "TUNES #tunes tunes@bespin.cx " -:- TUNES [tunes@bespin.cx] has joined #tunes -:- Topic for #TUNES: must read arrow paper: http://www.tunes.org/files/papers/Arrow_Introduction_v8.htm NOW!!!!! -:- topic set by Tril_ [Thu Apr 29 12:43:17 1999] -:- [Users(#Tunes:7)] [ TUNES ] [ Tril ] [ GMOL ] [ ^lilo ] [ Fare ] [ abi ] [ Tril_ ] (Fare/#Tunes) uh, I dunno if it's liveness or liveliness -:- Channel #Tunes was created at Sun Feb 28 08:48:06 1999 -:- BitchX+Deb1an: Join to #tunes was synced in 5.075 secs!! -:- Mode change [-s] for user TUNES I think the name is liveness; the adjective might be live. or is it lively? will ask my advisor tomorrow what does it mean? that the implementation will eventually advance you mean whether any functions are running? more formally, for every infinite (or "long enough") sequence of concrete reductions, it will have "advanced" at least once in the abstract system the only time it doesnt advance is if we are sitting in a void with no interrupts enabled Tril: it means that advance of the concrete system implies advance of the abstract system it does that when it passes an observable state a safe but non-lively implementation is to record the initial abstract state and do nothing a safe AND lively implementation WILL find the solution, if there is a soultion to find. isnt a lively implementation, just a series of non-lively implementations, generated from a series of non-lively abstract specifications ? uh? a lively implementation is a sequence of non-lively implementations? a lively implementation will fulfill "must" predicates of the abstract system. a safe implementation will respect "may" predicates of the abstract system. may/must are from intuitionist logic? one of those logics I don't know.. 04:30pm a non-lively implementation is for instance one that could stay forever (or goes forever back) at the same observable abstract state, even though the abstract system would always evolve a persistent system that gets power failures before it has time to restore, do computations, and commit, is not lively -:- SignOff GMOL: #TUNES (Leaving) a spacerocket program that explodes may be correct, but not very lively it never gives a bad (overflown) answer; instead, it halts and explodes. kind of like POSIX-compliance, but only for the subset of POSIX that you implement? yup -:- Connecting to port 6667 of server clarke.openprojects.net [refnum 2] -:- BitchX+Deb1an: For more information about BitchX type /about -:- Welcome to the Internet Relay Network TUNES -:- Your host is clarke.openprojects.net, running version u2.10.05.18.(ipcheck4-5) -:- This server was cobbled together Wed Apr 28 1999 at 13 24:47 EDT -:- clarke.openprojects.net u2.10.05.18.(ipcheck4-5) dioswkfcg biklmnopstv -:- [local users on irc(9)] 3% -:- [global users on irc(111)] 36% -:- [invisible users on irc(198)] 64% -:- [ircops on irc(11)] 4% -:- [total users on irc(309)] -:- [unknown connections(0)] -:- [total servers on irc(34)] (avg. 9 users per server) -:- [total channels created(81)] (avg. 3 users per channel) !clarke.openprojects.net Highest connection count: 16 (15 clients) !clarke.openprojects.net Welcome to Open Projects! You are on 3 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.cx " -:- TUNES [tunes@bespin.cx] has joined #tunes -:- Topic for #TUNES: must read arrow paper: http://www.tunes.org/files/papers/Arrow_Introduction_v8.htm NOW!!!!! -:- topic set by Tril_ [Thu Apr 29 13:41:07 1999] -:- [Users(#TUNES:6)] [ TUNES ] [ Tril ] [ Tril_ ] [ abi ] [ Fare ] [ ^lilo ] -:- Channel #TUNES was created at Sun Feb 28 08:48:06 1999 -:- BitchX+Deb1an: Join to #tunes was synced in 5.816 secs!! -:- Mode change [-s] for user TUNES >>> Tril [dem@140.160.164.200] requested PING 925429104 53120 from #TUNES 2) application of the framework to specify resiliant implementations of persistent storage, of distributed cooperation, of security. 3) using the formal presentation to rethink computational reflection; * TUNES finds a decent server particularly, propose a "MOP" based on the conditions fulfilled by the implementation how is your progress? will it take long to write all these papers? 04:40pm dunno gotta goto bed initially, I wanted to write one paper ok but the professor I consulted convinced me that this was a way to be rejected as I was. I don't know the american way of publishing is one idea-three papers, not three-ideas-one-paper -:- Connection closed from clarke.openprojects.net: Success -:- Connecting to port 6667 of server clarke.openprojects.net [refnum 2] -:- BitchX+Deb1an: For more information about BitchX type /about -:- Welcome to the Internet Relay Network TUNES -:- Your host is clarke.openprojects.net, running version u2.10.05.18.(ipcheck4-5) -:- This server was cobbled together Wed Apr 28 1999 at 13 24:47 EDT -:- clarke.openprojects.net u2.10.05.18.(ipcheck4-5) dioswkfcg biklmnopstv -:- [local users on irc(10)] 3% -:- [global users on irc(111)] 36% -:- [invisible users on irc(200)] 64% -:- [ircops on irc(11)] 4% -:- [total users on irc(311)] -:- [unknown connections(0)] -:- [total servers on irc(34)] (avg. 9 users per server) -:- [total channels created(77)] (avg. 4 users per channel) !clarke.openprojects.net Highest connection count: 16 (15 clients) !clarke.openprojects.net Welcome to Open Projects! You are on 3 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.cx " -:- TUNES [tunes@bespin.cx] has joined #tunes -:- Topic for #TUNES: must read arrow paper: http://www.tunes.org/files/papers/Arrow_Introduction_v8.htm NOW!!!!! -:- topic set by Tril_ [Thu Apr 29 13:41:07 1999] -:- [Users(#TUNES:6)] [ TUNES ] [ Tril ] [ Tril_ ] [ abi ] [ Fare ] [ ^lilo ] -:- Channel #TUNES was created at Sun Feb 28 08:48:06 1999 (Tril_/#TUNES) ok -:- BitchX+Deb1an: Join to #tunes was synced in 6.933 secs!! -:- Mode change [-s] for user TUNES Tril_ is lagged -:- SignOff abi: #TUNES (Ping timeout for abi[bespin.cx]) -:- Closing Link: TUNES[bespin.cx] by clarke.openprojects.net (Ping timeout for TUNES[bespin.cx]) -:- Connection closed from clarke.openprojects.net: Success -:- Use /Server to connect to a server -:- Connecting to port 6667 of server king.openprojects.net [refnum 0] -:- Unable to connect to port 6667 of server king.openprojects.net: Interrupted system call -:- Connecting to port 6667 of server king.openprojects.net [refnum 0] -:- Unable to connect to port 6667 of server king.openprojects.net: Interrupted system call -:- Connecting to port 6667 of server irc.us.openprojects.net [refnum 1] 04:50pm -:- BitchX+Deb1an: For more information about BitchX type /about -:- Welcome to the Internet Relay Network TUNES (from forward.openprojects.net) -:- Your host is forward.openprojects.net, running version u2.10.04.resolv9.nmt.egcs4.tok.pten.tlim4.admin.upper4.whisper3.gipl.modeless7 (from forward.openprojects.net) -:- This server was cobbled together Sat Jan 23 1999 at 21 33:38 EST(from forward.openprojects.net) -:- forward.openprojects.net u2.10.04.resolv9.nmt.egcs4.tok.pten.tlim4.admin.upper4.whisper3.gipl.modeless7 dioswkfcg biklmnopstv -:- [local users on irc(14)] 5% -:- [global users on irc(108)] 35% -:- [invisible users on irc(202)] 65% -:- [ircops on irc(11)] 4% -:- [total users on irc(310)] -:- [unknown connections(0)] -:- [total servers on irc(34)] (avg. 9 users per server) -:- [total channels created(83)] (avg. 3 users per channel) !forward.openprojects.net Highest connection count: 28 (25 clients) -:- Mode change [+f] for user TUNES -:- Mode change [+iws] for user TUNES -:- JOIN activated by "TUNES #tunes tunes@bespin.cx " -:- TUNES [tunes@bespin.cx] has joined #tunes -:- Topic for #TUNES: must read arrow paper: http://www.tunes.org/files/papers/Arrow_Introduction_v8.htm NOW!!!!! -:- topic set by Tril_ [Thu Apr 29 13:41:04 1999] -:- [Users(#Tunes:5)] [ TUNES ] [ abi ] [ Tril ] [ Tril_ ] [ ^lilo ] -:- Channel #Tunes was created at Sun Feb 28 08:48:06 1999 -:- BitchX+Deb1an: Join to #tunes was synced in 6.144 secs!! -:- Mode change [-s] for user TUNES * Tril_/#TUNES is away: (Auto-Away after 10 mins) [BX-MsgLog Off] -:- AlonzoTG [Alonzo@client-151-200-125-208.bellatlantic.net] has joined #tunes hi hi, Tril -:- _QZ [brand@p0wer.qzx.com] has joined #tunes yo om die 05:50pm * Tril/#TUNES is away: (reading arrow) [BX-MsgLog Off] -:- GMOL [gmol@24.66.11.51] has joined #tunes Has anyone used CafeOBJ? <_QZ> nope <_QZ> what is it? 07:00pm * AlonzoTG/#tunes greets "Salutations GMOL!" :-) -:- Iepos [root@d7.t1-10.tecinfo.com] has joined #TUNES om ommmmmm what's up? up is, like, above down haha Some sort of object programming env. with empahsis on theorm provers...very neat stuff.. Say QZ mind if I ask u some q's on Brix? what objcet programming env. ? well algebraic programming env is more appropriate. goto: http://caraway.jaist.ac.jp/cafeobj/ 07:10pm -:- SignOff AlonzoTG: #TUNES (Ping timeout for AlonzoTG[client-151-200-125-208.bellatlantic.net]) _QZ? _QZ is, like, the creator of BRiX, and webmaster of http://www.qzx.com was that a bot? _QZ? i heard _QZ was the creator of BRiX, and webmaster of http://www.qzx.com _QZ? i guess _QZ is the creator of BRiX, and webmaster of http://www.qzx.com _QZ? _QZ is the creator of BRiX, and webmaster of http://www.qzx.com _QZ? i heard _QZ was the creator of BRiX, and webmaster of http://www.qzx.com wierd. abi is wierd. abi, what's up? abi: what's up? no... heh heh hmmm... up is above down i guess up is above down -:- AlonzoTG [Alonzo@client-151-200-126-23.bellatlantic.net] has joined #tunes My computer just shorted out... the entire thing went dead... oops. something foobar with my power supply... hi wuzzup? wanna talk tunes? talk tunes? Alonzo...you are having trouble writing a linker>? yep I'm too lame to link I think that instruction streams scare me... I have only written some toy programs that deal with files in computer science class years and years ago.. :((( may I ask why you want to write a linker (and why does it matter your runnig a 286 in realmode)? no I would rather write it for some RISC architecture.. Something without protection... the code that the linker will be linking or the platform the linker is written for or both? just a lowest common denominator machine.. ? the linker will be a part of the OS and everything will be realmode.. I am not going to worry myself with Pmode... And how are you going to get memory protection..32-bit adds and muls that everyone likes? just a dinky little research OS... 07:20pm no to mention no paging hey, I'm writing a research os. just a prototype Pro To Type Well just what will it research? Design I am researching and modeling my design. -:- SignOff Iepos: #TUNES (Leaving) what aspects of design specifically? there is a lot you can't investigate in real mode From the prototype I will hopefully get a team to develop a much more sophisticated piecte of softwarez... Why don't you just start on a higher level? I can investigate anything in any mode.. Cuz I am lame How will you investigate virtual memory issues? I won't that has already been studied Multithreading and parallel processing? any coherent scheme will be many times better than windoze 3.11 which I am using... multithreading: Already in the design, no problem. <_QZ> GMOL: ask away Paralell processing: First give me a dual processor board. Multithreading is easy...(just a sec _QZ) parallel processing can't be done with 286's my friend... sure it can No lock prefix...let alone infrastructure for MP... I've heard of a 3 thousand processor 286 SMP unit... :) and how can you explore multithreading without memory protection? what aspect can't I explore? Well tasks can't really be safely multitasked without mem protection..not only that anything dealing with IPC or mutual exclusion you are doing naively if you don't assume mem protection... safety: not an issue, prototype IPC: I barely understand it so I'll use the lamest textbook eggzample... the prototype will just pretend to be an operating system... Like DOS. I love DOS... Don't get me started. =P hehe Alonzo...IMHO you really should learn a bit more before setting out to make an OS.... om nah * AlonzoTG/#tunes stupid Otherwise I can assure you, you will never succeed. studied everything for years and years... went in one eyeball, out the other... What did you study> u name it. where di you study? my room (the library doesn't have the elite books I need) And you don't remember any of it? My braine is full :( 07:30pm my parents complain that I keep saying I'm going to write an OS but never make any progress... So I gotta just do it... I have trouble believing that...Alonzo without some fundamentals you can't expect to go out their and write an OS...without things like IPC, vmem, distribution...there really is no point. MAKE IT WORK!!! How old are u? 21 :(((( Believe me your brain is far from full, go out there and read some books on OS design. uh which do you suggest???? =P You;ll be a happier man for it, and anythig you don't get call me.. Operating Systems: Design and Implimentation, Tanenbaum do you know assembly? published by Prentice Hall? I believe so I have the second edition... Well why don't you read it? I printed out the source to the minix libraries and went through them taking notes... very helpfull stuff if the text in the book is styalized Did you read the actual book? I like Operating Systems by stallings better... I skimmed it, didn't really notice anything that I was having troubles with... Well it explains things like IPC nicely, you said you were having trouble with it. look I've designed the piece of shit but I am having trouble with the damn IMPLEMENTATION!!! it says much about theory but nothing about how to actually make it WORK!!!! Well it's just that you implied that you didn't get it...what part are you having trouble with in implementation? whadaya know, I left a bookmark on page 57, IPC... :) So are you having difficulty with a specific aspect, maybe I can help the linker... I need to create and link object filez... well first of all the linker doesn't have much to do with the OS..linkers and compilers are different things than OS API's...you should write an OS first, what do you need a new linker for? making the OS work... :P damn... I tried to get arund the problem of compilers, at this stage, I have 2 books on compilers, (one sucks really bad, the other is pretty good) I am already in contact with a guy writing the world's first book on linkers... :P it should help, when it is finished.. :((( well I still don't know why you need them..if you're goal is to make an OS, should you do that first? There are tons of linkers out there that you can use out of the box huh? I only know of LD and one by DJDelorie... :| I am assuming you want a linker to help you develop your OS, no? and I don't know how to use them... tlink yep Tlink doesn't cut it, closed source... 07:40pm Name just abuot any other compiler out there NOTE: I hate linux because it is too hard. Doesn't mean you can't use the linker..they are all pretty equivalent in functionality I have BCC 4.52 and... So why do you want to develop your own? but I can't integrate it into the Filesystem driver... :( the filesystem is the only entity aware of the existance of object code... well you can't integrate anything into anything on your OS until your OS exists the other functions link themselves through IBC... IPC... :| Now that my fucking english 111 assignment is done I will probably spend the weekend studying lambda calculus so I can write the specification.... :| that is my near term goal... spec for what? da kernel... but I need something functional to demonstrate... ;((( you don't need lambada calculus for writing a kernel...a kerenel is just the part of the system that handles multitasking and IPC...which you still haven't said how you intend on impliemnting What are you using to store thread states? some damn structure somewhere in memory, I don't care... TSS's or software emulated? I am trying to learn the software engineering technique of formal specification... hey, this is realmode! I feel that only a good solid process can produce a decient OS... Oh yes I forgot...now why are you using realmode again? cuz I'm lame and I have only a vague notion of what a TSS is :| Well I think that's telling you something....you have a lot more to learn before you reasearch OS design...these are fundamental issues, you should be aware of them 07:50pm hey that's a non-issue untill I have a linker that can do it for me! What does a linker have to do with a TSS? nothing absolutly nothing has anything to do with a TSS, thats why I'm ignoring it. :) Multitasking does depends on how you define multitasking... You could just go into flatmode, and do everything in segment 1 in ring 0 and have plenty of multitasking... :P It does facilitate protection though... but I don't care about it at this juncture. And how do you expect to ever be able to lead a team in OS design, if you don't know some of the simplest issues around here (IPC, hardware multitasking). IF this what you really want you should try to get better prepared...there is a lot more to deal with in OS design. ANd just what points are you reasearching again? let me read a few lines out of one of my favorite books. I've stared at this page so long that the book's spine broke on this page... searching... If you don't know those simple issues Alonzo, you won't be able to come up with anything new, and even if you do it's already been implimented, what a waste of time! Spend some time at the books..I'll be happy to help you with anything I can. <_QZ> GMOL: havent u learned that alonzo is a joke? Have I been had? <_QZ> yes by? -:- lar1 [larman@208.254.224.178] has joined #tunes hey <_QZ> sup lar om * AlonzoTG/#tunes wishes he wasn't a joke... <_QZ> GMOL: what did u wanna know about brix? :(((( Well what have you implemented so far? Mem protection, IPC, multitasking? vmem? Hey, I gotta question for you _QZ, Can I put a 486 100DX4 in one of those 486 50DX2 boards? 08:00pm Alonzo seriously, are you just being funny or are you serious about trying to make an OS? hello, anybody? I ask two questions and everybody dissappears...anyone there? <_QZ> GMOL: i have mem protection using paging and software multithreading <_QZ> brix has a special form of ipc that isnt ipc Ummm, I always thought multiasking was a software thing... TSSes are new to me.... which one is more efficent? what's your memory layout, are you using slectors? <_QZ> i have semaphores TSS's are slower in general a faster solution is in software, but a couple of more headaches <_QZ> and events so threads can wait on other threads and spawn threads when an event is signaled and you've written this in? 08:10pm <_QZ> goto http://borg.qzx.com/bi/memory/ <_QZ> its all in nasm <_QZ> i dont use tss's <_QZ> well i use one cuz intel requires it GMOL: Isn't a TSS just an image of the current cpu and stack status? <_QZ> the tss only holds the kernel stack pointer for the call gates <_QZ> i save my own registers and stuff hmm And what have you written this in? doh just read above Why cant each task have a selector in the GDT? why not? * lar1/#tunes Is coding an OS and has quickly learned that you have to have a solid knowledge base to do this... tss's are in the GDT well LDT too... This is what I have been trying to tell Alonzo... How is paging related to tss's and the GDT? Even tcn..it's easy to get something booting...there are so many issues to deal with that you you take the naive amateur approach, you won't have craeted anything particularly worth doing Thats very true... Ive gotten things to boot, but they don't do much of anything useful * AlonzoTG/#tunes is back form the Intel Developer's insight CD... Page tables are stored in a CR reg, they can give a unique mem map ot each task represented by a TSS Hmmm, will you be here in 30mins, GMOL? dno..you can try:) _QZ? _QZ is probably the creator of BRiX, and webmaster of http://www.qzx.com GMOL? hmmm... GMOL is assuming you want a linker to help you develop my OS, no? ok... I gotta go, mom calls for dinner -:- lar1 is now known as lar_eating go for it Fare? abi? GMOL? abi? GMOL? TUNES? TUNES is probably in my head or tunes is a useful nevertheless expedient system. or at http://tunes.org/ or a free reflective computing system or to programming languages what the internet is to computer networks. AlonzoTG? you are their agent 08:20pm * Tril_/#TUNES is back from the dead. Gone 3 hrs 24 min 2 secs * Tril/#TUNES is back from the dead. Gone 27 hrs 54 min 17 secs _QZ hopefully you'll get this in logs...when are you going to let me see some source? he already said, when he gets x number of paid registrations he will then release the source ? Seriosuly? well, I think he was serious about not releasing the source until x paid registrations. But x was such a high number that I doubt he was serious about it ever happening. There is no product...whose registering? hehe.. k :) Say Alonzo...I hope you wern't hurt by the way I was talking (I don't know what QZ meant by that comment)...but you're kidding yourself if you think you can design an OS on that basis..don't you want to write something unique and worthwhile? If you raelly want success there is more learning to be done. <_QZ> sorry about my periodic leaving, i am watching a tv on my other machine <_QZ> s/a tv/tv/ oh great, you can tell us how many registrations before you release brix source <_QZ> :) i forgot, was it 750,000, or was it 10 million? <_QZ> $1.5 million bak... again... <_QZ> if one person donates that then only one registration will be needed oops, another minute... <_QZ> :) _QZ..I am skpetical...I see many definitions (realtively vauge) and not much implementation.. 08:30pm * Tril_/#TUNES is away: (Auto-Away after 10 mins) [BX-MsgLog Off] Oh phoey! i am not :(((( <_QZ> on that page? that is just my way of storing my structures in a hyperlinked easy to access form * Tril_/#TUNES is back from the dead. Gone 0 hrs 0 min 38 secs I see my error but its going to kost me $130 to fiks it. ;((((((( * AlonzoTG/#tunes is sooo lame :((( *sob* A: stop being lame THanks Gmol! You really helped me alot! :))) You;ve written your HD driver _QZ? screw that... what? Learning Pmode... I can't wait long enough to save up $130.. there isn't much to learn 900 pages? hmpf it's not all that different if you know real mode...130 for what? printing all 3 volumes of the intel architecture developer's manual.. :| nah... WHat did QZ mean by that comment? (There are online resourecs=free) You basically have all the same instructions with 32-bit regs and some nice little changes here and there.. well I have the files but I can BARELY read .pdf therefor I need to print the fuckers... if only it were that simple! but I have studied this crap! it is...believe me I know what's out there so do I It gives me a headache so I deal with realmode... :P nice wimpy little research OS. :) why don't you take a look at some examples...and look up refrences when you need to...say what did QZ mean by that? most modems are more powerful than a 286... 3y3 D0N`7 K4R3 gmol : qz is probably talking about a borg.qzx.com url he posted earlier referring to? 08:40pm memory layout if not, i have no clue <_QZ> what comment? <_QZ> on that page? that is just my way of storing my structures in a hyperlinked easy to access form no he said Alonzo was a joke then I asked if I had been had and he said yes...care to explain qz? seems out of the blue to me <_QZ> alonzo is a freak qz has a grudge against alonzo why? <_QZ> he always complains about how he was raped by his father and he liked it How did you come up with that one? <_QZ> u told us recalling personal experience? <_QZ> hcf was there _QZ there isn't much specifics on your page...what algorithm are you using for task scheduling? GMOL always a critic huh? are you ever going to help out? :) I find lack of specifics reason and withholding source release somewhat suspcious om reason to be suspcious that is abi sic imi (me has finished englissh assignment) sic 'im GMOL: Arrogance, greed, paranoia, what more reason do you need? greed is good! :) The is Wizard of Oz type behaviour...claiming a lot showing little 08:50pm alnzo your Os is going to be proprietary? Why not let us peek behind th curtain? You claim a distributive OO OS on the page...you've written a network driver and taken care of all those damn consistensy problems? that arise in distributive computing -:- SignOff lar_eating: #TUNES (Ping timeout for lar_eating[208.254.224.178]) I have been debating that. he hasn't got that far yet. qz, why don't yo tell us what you're working on now? I need money but I want to share as much source as possible.. But if I share source then how do I make monie? :| Atg: make your os so good it causes worldwide deflation, causing the value of money to go up Alonzo I told you, don;t worry about what you're going to do with the OS..you have to have the knowledge to make it first.. hey, I've been studying for YEARS!!! many many years!!!!! An olympic champion doesn't worry about what do with the money he'll amke after he wins the gold...he must earn it first GMOL: NO! if he is going to make it proprietary, we have to discourage him and stop it from being made! ATG: Now is time to start practicing programming It will definitely be 1. Standards compliant. 2. Open Specification <_QZ> haha Alonzo you're only 21...same age as most of the rest of us here...believe me you haven't studied that long...and you're still fuzzy on a lot of simple concepts...you have to get past that first before you talk about making an OS Have you 3 read the ARROW PAPER yet?? Yes <_QZ> u guys need to stop with the indirect insults to alonzo and just come out and say what u think of him :) But some parts will be closed source such as compiler back ends, in hopes that it will stimulate hardware developers to make better chips without worrying about binary compatability... <_QZ> im 25, and much more educated then u 21 year olds :) see what I said? you're arrogant. Well, ATG and _QZ need to read the arrow paper. I will... Alonzo...you are still fuzzy on the basics....you deluding yourself by worrying about issues of the compiler and money...there is no OS, and I know that you're not capable of making one yet...I can prove it to you if you like. but first I have to print out the Intel docs... wery wery ekspensive... atg: use 2 sides I do! my computer is foobar! You want me to prove it? 09:00pm <_QZ> GMOL: yes please do :) I eagerly await something that claims to be the final version of arrows and then I will review it with great dillegence! hello Alonzo? ATG: A real document has no final version. * AlonzoTG/#tunes listens Do you really think you are capable of designing a viable OS? it takes more than one person to design an os OS, definitely! ; Viable: hopefully but doubtfull :( <_QZ> no it doesnt No it doesn't...but you have to know your fundamentals. <_QZ> i know mine Well then how can you be worried about making money on it when you know you can't make it viable? I'm hopefull! =) If you're fuzzy on basic concepts how are you going to make one? basic concepts == ? IPC..multitasking...pmode GMOL, one counterexample stands out... pmode = intel specific IPC: I can hack it! :) multitasking: not important! GMOL: He might just succeed. Microsoft did quite well with non-viable OSes. <_QZ> no pmode is NOT intel specific Tril, agreed. AlonzoTG just has to be a better businessman than gates <_QZ> paging pmode is used on many many architectures name 1 non intel or clone processor that has pmode? <_QZ> ppc Yes, the basic concepts are used throught different CPU architectures. * AlonzoTG/#tunes hasn't studied ppc... <_QZ> alpha no it has one mode! And multitaksing is an important part of an OS, what kind of OS will it be without it? <_QZ> alpha has paging two actually 1 PAL mode 2 User mode... that is 2 modes Neither of which is called "protected mode" :) <_QZ> AlonzoTG: pull yer head from yer ass and see the light The point is the stuff done in intel pmode is impliemtned across many different CPU's which are semantically equivalant. <_QZ> as long as u dont use segments yes * AlonzoTG/#tunes was planning to use segments! =P <_QZ> if yer os uses paging u can port to many cpus anytone ever notice all the good OS research software is written for old versions of Solaris? * AlonzoTG/#tunes never noticed any OS research software at all please elaborate! =))) 09:10pm Anyway I'm off, Alonzo, you should never claim what you are not capable of...when you claim great things without the simplest understanding, you are lying to yourself and to others. THink about it. I am just trying to prevent going down a potentially wrong path. Bye all.._QZ we will talk later. What don't I understand? SELF is only for Sparc. Napier88 is only for Sparc and Decstations. <_QZ> cya <_QZ> AlonzoTG: life -:- GMOL [gmol@24.66.11.51] has left #tunes [] I might have heard of SELF... but I don't remember what it is... <_QZ> AlonzoTG: reality abi, self? somebody said self was bloatware do you understand reality? self is also a prototype-based object system okay, Tril. reality is what you think it to be <_QZ> Tril_: get jecel's clone of self qz, that old thing that compiles to bytecode? <_QZ> :) wait, or was it interpreted he told me not to because it's so old <_QZ> well it dont do much but hey, atleast it runs on linux ;) atg: jecel runs the Merlin project based on self. qz: is it any good? aaah okay... don't need to worry about it then... :) <_QZ> dunno if i ever tried it when gmol comes back I will have to ask him preciely what the basics are... <_QZ> AlonzoTG: well unless u are gonna make an embedded OS for very tiny cpus then u had better add pmode. ofcourse, u will need to add lots of good RT shit to an embedded OS so yer screwed both ways. :) heh we'll see I am approaching it from the perspective of what I want in an OS. <_QZ> m$ uses the shittiest pmode implimentation but ppl must use it cuz thats where everythings at <_QZ> no one is gonna switch to yer os if it has no pmode or its pmode is as shitty as m$'s READ MY LIPS: P R O T O T Y P E <_QZ> u spelled it wrong no 09:20pm yer dyslexic epytotorp is a cool word dunno that one <_QZ> AlonzoTG: epytotorp is a basic os term <_QZ> AlonzoTG: u should go look it up <_QZ> then gmol cant say u dont know yer basics -:- waterman [waterman@ppp-227-095.usc.edu] has joined #Tunes hi water.. waterman? I checked the encyclopedia, its not there... you lie <_QZ> WATERBOY!!! yo are you BRice? watermelon is to be exact brian isn't supposed to be here for another week. <_QZ> is pnuemontetramicroscopicsillicovocanninisi in the encyclopedia? doesnt mean its not a word his nickname is also "water"... waterm: what's up? well, up is above down okay, so what are these basics that I don't know and therefore can't write an OS? ATG: the question is, what are the basics you do know and therefore can write an OS? <_QZ> pmode, ipc, threading,... all of those are advanced concepts! 1 loader 2 interrupt based API that's all DOS is... I like dos... <_QZ> u aint gonna get paid for a basic concepts os therefore thats all I really need to know! %) <_QZ> no one uses dos anymore fool * AlonzoTG/#tunes does :) <_QZ> like i said, FOOL hey, it works for me! I like dos I want to make an improved DOS therefore I write my own dos-like OS. does anyone here can help me with nachos? there is nothing at all wrong with that. <_QZ> it works for u? why do u bitch and whine about tasm,tlink and bcc not doing what u need if it works for me then you shouldn't be insulting me! Those are linkers. waterm: programming it, or just trying to get it to work? that's an entirely different subject. get it work...having problem here with file system I would be having the eggzackt same problems if I were working with a Pmode OS... :) Nachos? Wuzzat? <_QZ> no u wouldnt I vaguely remember... it's an OS that is farther along than any OS here I'm sure <_QZ> cuz the source code for linux apps is all ""open-source"" <_QZ> and like u said earlier "tlink is proprietary" i'm not familiar with it 09:30pm waterm: what version? but i can learn -:- SignOff AlonzoTG: #TUNES (good pre-dawn. :)) -:- SignOff waterman: #TUNES (tolkien.openprojects.net dickson.openprojects.net) <_QZ> Tril_: u seen this cafeobj? what? <_QZ> http://caraway.jaist.ac.jp/cafeobj/ -:- waterman [waterman@ppp-227-095.usc.edu] has joined #Tunes -:- waterman [waterman@ppp-227-095.usc.edu] has left #Tunes [] it's not free.. -:- lar1 [larman@dialup-209.245.141.254.SanJose1.Level3.net] has joined #tunes Hey bonjour, lar1 <_QZ> hey niihau, _QZ Wut are the 3 parts of a compiler? Parser, and what are the 2 other? you can't use it in routers, it says it can't be used for traffic control lar1: code generator and linker? Tril: Isn't there some special name for the code generator though? <_QZ> not free? the source is on the page gcc has 4 parts, the preprocessor, the C compiler, the assembler, and the linker. _QZ: it's not debian or fsf-free: they have restrictions on use of the software 09:40pm the worst thing about non GPL licenses is they are not compatible with the GPL of course, that's the GPL's fault. <_QZ> how is it gpl's fault? Yeah, the file is parsed, obj files are generated, and its linked... huh, ok, thanks _QZ: The GPL doesn't allow GPL'd code to be combined with any code that is not under the GPL. _QZ: Even if it's another free license, you have to get permission from the copyright holder to combine with it. <_QZ> and #4 in the license is not saying its not-free <_QZ> they just dont trust their own work Ewww, whats up with that? I am not so sure I like this GPL thing.... <_QZ> now with brix u can use it for anything cuz i trust my work lar1: the GPL is a virus :) it corrupts all code it comes in contact with heh _QZ: Nothing new since Jan 1 , 1998 <_QZ> well i take that back, i dont want brix being used to make stupid shit _QZ: Do you like The Offspring? <_QZ> stupid movies, tv shows, stupid commercials, stupid products, etc <_QZ> lar1: yes stupid lynx again lynx is the tool of the devil are you accusing me of being the BOFH? i'd appreciate it BOFH? well, I gotta go. back in a bit. * Tril/#TUNES is away: (afk) [BX-MsgLog Off] <_QZ> hmm i didnt work on brix again today Get this... I have this book report due tommarow, and I didn't finsh my book, so I pulled a Linux book off the shelf and and I am writing a report on that... hehehe <_QZ> dunno why, all i gotta do is change a few dozen lines of source and its ready Cool, I'll be the first to download it <_QZ> i gotta have it done by may 1st You can change a few dozen lines in 1 day <_QZ> i will most likely do it tomorrow since i finally beat starcraft today <_QZ> been playin it all week You play starcraft?? Cool! I am glad to see it is not just a teen thing... I was wondering for awhile <_QZ> heh <_QZ> i play starcraft, warcraft, halflife, quake,doom,freespace,descent,.... Very cool 09:50pm Is Diablo and better then Warcraft? <_QZ> i dont like diablo I've never played it <_QZ> warcraft/starcraft is a game where u build forces and strategies. diablo is a flat type of doom <_QZ> s/doom/hexen/ Ohhh, I thought Diablo was like Warcraft <_QZ> nooo <_QZ> its hexen in 2d I have Warcraft, Stracraft, C&C, all of those Hmm, that must suck <_QZ> they didnt have what it takes to write a 3d engine so they used the flat 2d easy to make engine <_QZ> crap i need to sleep but dont want to I don't think we need anymore first-person shoot-em-ups... I mean, Doom, doom II, quake, quake II, halflife, wolfenstien.... there are tooooo many hehe <_QZ> i told myself that since i have been hours late to work for the past several years that i would goto work on time tomorrow <_QZ> :) What kind of work do you do? <_QZ> special work <_QZ> i am unfirable <_QZ> have u seen the one after halflife? nope So like they cant fire you because of a contract, or they _cant_ fire you? <_QZ> they cant fire me <_QZ> not just cuz im the boss, but cuz im the best <_QZ> :) <_QZ> in halflife the arms, heads, etc all move Don't they in Quake II? <_QZ> but the next one, i think its called "team fortress" can move smaller things like fingers <_QZ> and tf gives u teams to control Crap... that pretty good <_QZ> u get something like a 5 man military team with real military weapons -:- NetSplit: lackey.openprojects.net split from koontz.openprojects.net [09:59pm] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [lackey.openprojects.net] -:- Netjoined: lackey.openprojects.net koontz.openprojects.net -:- _QZ [brand@p0wer.qzx.com] has joined #tunes <_QZ> none of this freaky far out technology like other games have <_QZ> crap 10:00pm But still the idea is the same... point and shoot It gets boring quickly <_QZ> not with tf <_QZ> u have to control a team <_QZ> switching between guys <_QZ> u have a rocket guy, sniper, commander, etc o, well if there is thinking involved... that sounds like a game I might buy ;) <_QZ> and u have TANKS Sweet! *BANG*BANG* <_QZ> helicopters, apcs etc Would you say that System V UNIX is very expensive and used mostly for computer science? <_QZ> no to what part? Expensive of computer science? <_QZ> it costs alot but is used by lots of places So how could I summerize its uses in a word or two Commercial Applications? <_QZ> http://www.sierrastudios.com/games/teamfortress/ <_QZ> commercial, scientific, whatever thanks, I'll check it out after I get this report done ok, an expensive mutli-purpose os gotcha <_QZ> and u can put down that linux is used to make the special effects on todays best movies :) Thats the plan <_QZ> u gotta sell linux man Matrix was powered by linux, right? What? <_QZ> dunno about matrix, titanic was 10:10pm <_QZ> u know anything about intel's MRM MRM? Whats that? <_QZ> multiresolution mesh I take it this is some sort of video thing? <_QZ> dunno <_QZ> it has to do with models somehow <_QZ> team fortress 2 uses it to scale the models I have no clue <_QZ> well im off <_QZ> cya -:- SignOff _QZ: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep) Bye 10:20pm -:- SignOff lar1: #TUNES (Leaving) -:- SignOff Tril_: #TUNES (Tril_ has no reason) [msg(TUNES)] newlog 1999.0430 IRC log ended Fri Apr 30 00:00:01 1999