IRC log started Sat Feb 5 00:00:02 2000 [msg(TUNES)] permlog 2000.0205 -:- smoke [smoke@15dyn126.delft.casema.net] has joined #tunes -:- SignOff smoke: #TUNES (One day sheep will rule the world) -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us1019.javanet.com]) -:- ult [noone@user-38lcmpl.dialup.mindspring.com] has joined #tunes -:- SignOff washort: #TUNES (Ping timeout for washort[d140.narrowgate.net]) -:- Fufie [stig@tunnel-44-10.vpn.uib.no] has joined #tunes -:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes -:- SignOff ult: #TUNES (Read error to ult[user-38lcmpl.dialup.mindspring.com]: Connection reset by peer) -:- bineng [Anders@j141.ryd.student.liu.se] has joined #tunes -:- Downix [down@d-gnaps-71.ici.net] has joined #tunes morning all morning what's up bineng? well.. the usual 06:40am useless coding and plotting to take over the world? apart from the coding part, yes are you up to something? right now not much, downloading M13 ic. I had a glance at it, didn't see much improvements over M12. well, I submitted a few bugs I found did they fix 'em? looks like it it waas nothing major ok small thing while viewing files on the desktop I'd fixed the code on my own copy, then submitted the change you did that? pretty much, I just dcc'd the fixed snippits to the people on mozilla.org irc server, told them where the hole was ic.. wasn't it hard to locate the right pieces of code? only 2 lines dif. nope 06:50am helps that I'd logged all errors during compile what platform are we speaking about? Linux -:- mattis [martin@pool1-249-83.cm.starport.se] has joined #tunes 07:00am * Downix/#tunes is away: (Auto-Away after 10 mins) [BX-MsgLog On] leave #tunes * Downix/#tunes is back from the dead. Gone 0 hrs 1 min 4 secs hmm -:- mattis [martin@pool1-249-83.cm.starport.se] has left #tunes [] 07:10am -:- hcf [nef@me-portland-us644.javanet.com] has joined #tunes * Downix/#tunes is away: (Auto-Away after 10 mins) [BX-MsgLog On] * Downix/#tunes is back from the dead. Gone 0 hrs 21 min 47 secs -:- SignOff Downix: #TUNES (Downix has no reason) -:- ult [noone@user-38lcn5t.dialup.mindspring.com] has joined #Tunes -:- washort [washort@d108.narrowgate.net] has joined #tunes -:- lar1 [larman@dialup-209.244.106.79.SanJose1.Level3.net] has joined #tunes Hey 11:00am -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us644.javanet.com]) -:- SignOff lar1: #TUNES (Ping timeout for lar1[dialup-209.244.106.79.SanJose1.Level3.net]) -:- ult_ [noone@user-37kbaqe.dialup.mindspring.com] has joined #Tunes -:- SignOff ult: #TUNES (Read error to ult[user-38lcn5t.dialup.mindspring.com]: Connection reset by peer) -:- ult_ is now known as ult -:- SignOff bineng: #TUNES ( ) -:- SignOff ult: #TUNES (Ping timeout for ult[user-37kbaqe.dialup.mindspring.com]) -:- water [water@tnt-9-162.tscnet.net] has joined #tunes on for a minute hey eih, sorry i was pretty incoherent the other night anyone here? no oh. heh just me k well bbl then -:- SignOff water: #TUNES (Leaving) -:- ult [noone@user-37kbap5.dialup.mindspring.com] has joined #Tunes 01:40pm -:- SignOff Fufie: #TUNES (Ping timeout for Fufie[tunnel-44-10.vpn.uib.no]) -:- mibin [mibin@62.11.104.137] has joined #tunes -:- pyro [tcn@cci-209150250077.clarityconnect.net] has joined #tunes -:- lar1 [larman@1Cust128.tnt22.sfo3.da.uu.net] has joined #tunes -:- SignOff pyro: #TUNES (pyro has no reason) Hey What could one do with a Multiproccessor SPARCstation? -:- pyro [tcn@cci-209150250158.clarityconnect.net] has joined #tunes compute pi to really large digits? Heh I was offered one for free but its big and I need to justify it to mom why? its free!?!? Its big and if you had *how* big? tell mom to screw off unless of course its 20' by 20' it's small enough dammit size of small refigorator then tell her to help you carry it so? :) You should see my garage It full of crap get rid of some other crap lar1 you sure thats not one of the 240volt machines? at the worst, you lay it on its side, and tell her its a bed hah haha Heh eihrul! then when she asks you why her power bill is for 4000 dollars ult: Haven't seen the thing yet what'll you say? Hmm how old is this thing? Project for science class :) Dunno what speed? i guess speed is an important issue and if I can get CL to run at Bison speeds, etc... then I might reconsider it for various projects I do. He said the guy that has it got a better machine and cant keep this one, and he dones't want to toss it pyro: No idea... he said somthing about it not being made by sun though well.. 02:30pm lar1 well if its free and your picking it up via car (shipping won't be a problem) you could always toss it Yeah SPARCs look like good processors pyro: Do they run 68k? no they're sparcs! Any idea of what one could do with such a box? Any thing that SPARC does real special that Intel doesnt? pyro: Ah. I didn't know a sparc was a kind of chip you can put Linux or netBSD on it, if Slowaris don't cut it :) Hmm, ok... it has 3-level page tables, compared to the usual 2 256 bytes, 64k and 16M, I think. Sounds interesting I'll just claim that I can learn a whole bunch from it... so you can put a small thing on a 256 byte page.. where on the x86 you'd have to waste a whole 4k page and as far as the size goes.. what is it, about 3' high, 1' wide, 2' deep? He said about 2x2x3 feet I think Its not _that_ bit big deal. I had some speakers bigger than that :) Hmm, good point :) What kind of moniter do these puppies take? it doesn't have one? I am not sure I have a 19" sun monochorome on my sun 3/50 haha what other crap have you got ? :) 02:40pm You don't want to know! ;) in the garage are ya? Nah I just use that for storage :) My room is too full... -:- SignOff washort: #TUNES (Ping timeout for washort[d108.narrowgate.net]) you should be able to use any monitor with that any multiscan monitor, I mean 02:50pm ugh Hmm Thats most any monitor, right? that's enough OS work today Heh at least I got the display/keyboard drivers done Coo ERG My pointing device is making this damned clicking noise I think I got somthing stuck under it -:- washort [washort@d108.narrowgate.net] has joined #tunes -:- SignOff lar1: #TUNES (BBL) -:- SignOff pyro: #TUNES (later.) 03:00pm -:- SignOff ult: #TUNES (Ping timeout for ult[user-37kbap5.dialup.mindspring.com]) -:- rares [rares@wtrb-sh9-port88.snet.net] has joined #tunes -:- Fufie [stig@tunnel-44-22.vpn.uib.no] has joined #tunes morning 'nite Fufie: hey -:- SignOff washort: #TUNES (Ping timeout for washort[d108.narrowgate.net]) 03:40pm -:- water [water@tnt-10-67.tscnet.net] has joined #tunes * eihrul/#tunes puts on his raincoat. hey eih sorry i was incoherent the other day... had been working too much around some nasty fumes :) np, i've become used to it :) lol -:- bineng [Anders@j141.ryd.student.liu.se] has joined #tunes hey bin hello how's it going? could be worse 03:50pm work continues to suck, or course :P I don't have much of that now actually.. not much in school lots of time to tinker with freebsd btw, can anyone recommend a good non-corporate-oriented news service or ticker? i've been devoting my free time lately to getting slate going that's good. Any progress? well, eihrul and i resolved how methods would work but did we resolved how they'd be invoked? simple enough to explain? s/resolved/resolve and we seem to have a good working object-function union idea heh i'm letting eihrul explain stuff :) * bineng/#tunes listens.. we're basically using the meta-object to set the read-write access on the various slots and using a special "result" slot to refer to the code of the object-as-function some of the slots are input variables, and others are local variables does this 'result' slot represent the return value? yes and the function is invoked when this slot is accessed, then= when you ask for the result, it gets computed based on the object's current state so it's lazy 04:00pm note that the actual code is in the separate object referred to by 'result' ok. What if some variable the function depends on changes? so that the main object acts as activation record and the 'result' is a sequence of message-sends then the result changes, i suppose but if the result was already calculated, there must be a way to detect such changes well, you could use data-flow dependencies i mean, track them what and how? and how are vars kept persistant? There has to be an action for this... abi forget and how water: I forgot and how hehe won't you answer her? ;) heh i'm not sure how we'll have the actual code arranged just yet, there are a lot of issues that interact but the immutability policy would assist in keeping the semantics close to data-flow I don't understand hm what is this data-flow? there are programming languages based on the flow of values, not variables' state such as? each 'variable' is just a name for an immutable value * water/#tunes looks for a data-flow lang url well, they're not common. sisal and val are examples it's an idea related to functional programming basically also, a lot of compilers create data-flow graphs to study and optimize algorithms can't say I'm familiar with it well, consider a function invocation as the node of a graph with i/o between the function-applications being the arrows in the graph 04:10pm so, you don't deal with variables, just persistent function-applications if you have this graph information stored, then whenever one of the input nodes changes value, updates to dependent data happens automatically how does it differ from functional programming? it's very similar, but there was some syntactic/semantic difference iirc i think it's basically just concurrent fp, but i haven't looked at it in a while ok sounds like fp well, there was an actual data-flow processor architecture being played with i think that's why dfp is lumped into a separate category I heard of that processor I think the only reason i know much about it is apld abi apld? apld is, like, advanced programming language & design at http://cseng.aw.com/bookdetail.qry?ISBN=0-8053-1191-2&ptype=179 it's an online book as pdf damn the link's bad hold on putting 'advanced' in book titles or project names is kindof pretentious :) yeah, it's just an overview of different paradigms but it does get into the nitty gritty 04:20pm n/m the addison-wessley cs site is just down aw isn't too bad any of you know docbook? a little in a chapter I use sect1 for the sections, right? like \section{} in LaTeX? yeah i think so no manual? :) have the online html reference but references only tell so much :) generating documentation here bin: any more questions? i don't think i really explained much 04:30pm hmm.. I think I pass for tonight k 04:40pm water: hrmm, perhaps a news page or stamp pages with last-modification dates wrt slate? i'm not sure news about a not-yet-implemented language is really that interesting well, i keep looking to see if you added any more information to the page :) and it gets tedious hm i'll let you know, ok ko 05:10pm besides, you usually find out about stuff first i mean before i post it yes, but its usually more coherent once its all been processed and written out on the page >:) or atleast more specific yes, that processing takes some doing are you working anything on the page? yes, semantics is the current focus when i get that very well-defined, i'll rewrite the rest to support it right now, there are a few inconsistencies since i am experimenting with a few ideas bin: btw, we figured out that block-scoping amounts to inheritance hierarchies of the activation record objects since each one contains the functionality and state of all its parents Nope, that sentence means nothing to me hm we have some communication problems :( you grok block-scoping, right? I think that's the part I don't know huh? so much easier to explain, eh it's like the nesting of contexts in pascal and derivatives like in lisp? each variable is only visible in the contexts it's defined in and all sub-contexts ok what about 'activation record objects'? those objects that act like functions are really activation records for their 'result' code 05:20pm and within an object, variables are inherited from parent objects (as well as method code) do you see a hierarcy of objects as one object? you mean in slate? yes well, it's a good oo tool to be able to view an object's entire protocol and state by "collapsing the inheritance hierarchy" slate should provide both, i guess, but it's an environment issue I guess the question could be restated as if objects can be seen as self-contained if desired. oh yes k btw, that's a refactoring idea I didn't know that yeah, a lot of refactoring involves moving code up and down the hierarchy -:- air [brand@p0wer.qzx.com] has joined #tunes I'm not sure about the nature of the hierarchy you're speaking about inheritance, sorry :) so B would inherit from A instead of the contrary? no no A would contain method c instead of A''s parent ahh, that makes sense (more sense at least) it applies to variables too 05:30pm but in usual oo, refactoring gets really complicated and in Slate? heh i'm working on it :) -:- SignOff rares: #TUNES (Ping timeout for rares[wtrb-sh9-port88.snet.net]) i think data-flow graphs (func prog style) should make it a lot simpler eih: btw, refactoring applies to meta-object hierarchies as well as inheritance hierarchies er... mo inheritance hierarchies well, since they're just object hierarchies i don't see why it wouldn't :) since mo functionality has the basic effect of a bank of switches, it would be good to have varying 'preferences' hm i really need to refine the mo spec mo? meta-object ofc.. 05:40pm maybe i should rename them I think so. IMO "meta-object" isn't very informative name yes, but will the name be any more intuitive than meta-object? e.g. meta-space or behavior controller ... the mo idea is behavioral reflection For me it seems natural to just say the object is in a 'space'. No meta. well, we also want namespaces "functional space"? separate from meta-spaces no, that doesn't sound very suggestive just brainstorming i know you could just say 'behavior' hm and consider the meta-object just a representation of that (which it already is) sounds good yes, but the object has a behavior too maybe meta-behavior? yes, but that's the behavior of the object representing the behavior :) no it ain't gah :) well what do you suggest, then? the behavior of the meta-object is not inherently related to the behavior of the object as far as i see it, or atleast not concretely... since the meta-object can be changed yeah, and the mo behavior can be changed independently of object behavior in fact, i'm forcing them to be orthogonal * eihrul/#tunes ponders why the word behavior suddenly seems to be spelled 'oddly.' i.e. i don't want one to be able to duplicate the functionality of one by the other 05:50pm which means you could always detect changes in meta-behavior by looking for messages to the mo hm... *certain kinds* of messages to the mo ones that update its state water: what do you aim for Slate to become? Anything useful in practical appications? s/appications/applications/ sure... any application needing lots and lots of flexibility or concurrency or where you want to have flexible simulations so it's not just a project for experimentation? no i'd like to actually use this system good to know yeah, keeps me from adding crappy features ;) it didn't help with unix, afaik >:) heh so, is it meta-behavior or what? meta-behavior is the behavior of the meta-object though... nothing seems to sound quite right no it isn't! then what is it? meta-behavior is not behavior meta behavior to me is what the meta-object describes not quite the object contains the description of its own behavior by referring to methods, variables, etc the mo controls how those slots are used 06:00pm doh heh i was thinking of behavior at the wrong level from the level you thought of it :) oh my behavior was the behavior of the behavior :) as you pointed out yes :) hmm... what to call it then.... because meta-behavior is just as bad as meta-object yep but that would be the proper formal term for its function * eihrul/#tunes nods. maybe meta-space after all? or perhaps we should keep looking yes yes what? well, meta-space is just another way of stating meta-object oh except inverting the relationship between the object and the meta a bit by saying the object lives in the meta-space rather than the object has the meta-object well i like the intuitive idea of dragging object icons between visual spaces yes but in slate but perhaps that would be better used for other semantics any object can be considered to contain or to contain or be contained by s/to contain or to contain/to contain heh well, honestly nothing comes to mind * eihrul/#tunes can't think of a *single* word that describes a meta-behavior/meta-space/meta-object. because it all deals with the representation of the object you know about "mirrors" from self? unfortunately, no hm different concept, n/m it's a kind of meta-object mostly structural does it involve the evil word 'reify'? :) 06:10pm kinda it allows you to edit an object by making the object's structure into a complex of self objects but it doesn't really add anything to the object, just "reifies" :) 'edit'... i take it visually? no, the visual stuff is the morphic overlay but morphic uses mirrors to get that structure, yes rmm, well my definition of editing lends itself to saying 'interactive modification' so probably just in my definition of the term :) though, a mirror doesn't sound too different from the object itself... which is probably 'hence the name' right though how is a mirror more useful than a reference to 'self'? 'self' doesn't respond to messages other than primitive kinds and, of course, to it's own methods :) mirrors act as wrappers how so? (what things do mirrors allow one to do or what operations do they provide) read about them in the progref k though i'm already reading a paper :) while talking to you do you think i am ever not reading something? what about when you sleep? heh i read volumes about the void :) but you've already read them... or else you wouldn't be able to read them nonsense (unless your subconscious is very creative) you grok the void or not? the void is void umm.. does someone know how I unmount thise FS of mine under freebsd? 06:20pm