IRC log started Thu Feb 17 00:00:00 2000 [msg(TUNES)] permlog 2000.0217 -:- SignOff smkl: #TUNES (Ping timeout for smkl[glubimox.yok.utu.fi]) -:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us926.javanet.com]) -:- hcf [nef@me-portland-us932.javanet.com] has joined #tunes -:- smoke [smoke@15dyn89.delft.casema.net] has joined #tunes -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us932.javanet.com]) -:- hcf [nef@me-portland-us932.javanet.com] has joined #tunes -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us932.javanet.com]) -:- hcf [nef@me-portland-us932.javanet.com] has joined #tunes -:- smoke_ [smoke@vengeance.et.tudelft.nl] has joined #tunes morning hi hoy 02:50am how were the requirements for the HLL derived? (www.tunes.org/HLL/requirements.html) it seems to be based on the philosophy described in 'WhyNewOS.html', which I do not fully agree with.. it all seems to be based on money :( 03:00am money? 03:10am yup -:- abi [nef@bespin.dhs.org] has joined #tunes the paper assumes that it's `terrible' that programming takes up so much precious time and money. and almost the entire piece on security does not apply to recreative programming it's not that i don't agree on the important points, but it's all brought forward as if to stimulate the corporate world to dig the ideas of Tunes. (which is, in my humble opinion, ridiculous) hmm.. we can't have that.. what? (since i said so many things at once) 03:50am searching for a goal for tunes is not necessary imho sucking up to suits is not a Good Thing grin well it's not bad, but it's unnecessary perhaps i interpreted WhyNewOS incorrectly.. i get the feeling that, in a way, it's more of an excuse than a foundation ask Fare fare? i think fare is sometimes looking at another screen and interruptible with a beep (/msg Fare ^G) or at http://www.tunes.org/~fare/ or pronounced Fahree brb.. need shower 04:00am -:- SignOff smoke_: #TUNES (One day sheep will rule the world) -:- SignOff Fufie: #TUNES (bbl) -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us932.javanet.com]) -:- SignOff smkl: #TUNES (changing servers) -:- Phigmeta [nospam@207-193-6-155.noc.sbis.net] has joined #tunes -:- Phigmeta [nospam@207-193-6-155.noc.sbis.net] has left #tunes [] -:- SignOff thomas: #TUNES ([x]chat) -:- smokie [tw026024@zaalf15.twi.tudelft.nl] has joined #tunes -:- SignOff smokie: #TUNES (BitchX-75+Deb1an -- just do it.) -:- bineng [Anders@j141.ryd.student.liu.se] has joined #tunes -:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes Context! 09:10am -:- ult [noone@user-37kbaig.dialup.mindspring.com] has joined #Tunes switch 09:20am eh I got TeXmacs working. Anyone that knows how to type a logical-not symbol? 09:40am ^ erm, ! even. that "-" with a small bend on the end '!' doesn't do anything 10:00am -:- SignOff bineng: #TUNES (Ping timeout for bineng[j141.ryd.student.liu.se]) -:- bineng [Anders@j141.ryd.student.liu.se] has joined #tunes -^ thou meanest? hum? nah.. that doesn't do it either 11:10am I mean the negation sign in logic.. can't find a way to write it ~ will do as a replacement, you mean? yes i haven't used texmacs ok but ~ is just so common I'd prefer the "real" sign... wonder if it's present. 11:20am sure probably just in a different font -:- SignOff ult: #TUNES (Leaving) 11:30am -:- Fufie [stig@tunnel-44-11.vpn.uib.no] has joined #tunes -:- rares [rares@wtrb-sh2-port212.snet.net] has joined #tunes -:- hcf [nef@me-portland-us232.javanet.com] has joined #tunes -:- rares [rares@wtrb-sh2-port212.snet.net] has left #tunes [Segmentation] what is meant with `hygienic macros' (context: scheme) ? is it a technical term or just a metaphor for how clean it is? metaphor it's about unintended variable capture ok ah! i actually read about that :) 12:10pm -:- SignOff bineng: #TUNES ( ) -:- thomas [thomas@193.217.63.152] has joined #tunes hei 01:10pm hi thomas 01:20pm howdy, smoke 01:40pm -:- NetSplit: sterling.openprojects.net split from devlin.openprojects.net [02:00pm] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [sterling.openprojects.net] -:- Netjoined: sterling.openprojects.net devlin.openprojects.net -:- Fare [rideaufr@esmeralda.enst.fr] has joined #tunes -:- Plundis [plundis@130.238.23.252] has joined #tunes -:- smoke [smoke@15dyn89.delft.casema.net] has joined #tunes -:- abi [nef@bespin.dhs.org] has joined #tunes -:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes -:- hcf [nef@me-portland-us232.javanet.com] has joined #tunes -:- thomas [thomas@193.217.63.152] has joined #tunes -:- air [brand@p0wer.qzx.com] has joined #tunes anyone know about mailservers? hey what is a continuation? a continuation is probably the reification of the computation, as embodied by the current contents of the call stack. what does that mean? heh 02:50pm -:- eihrul [lee@usr5-ppp195.lvdi.net] has joined #tunes -:- SignOff smoke: #TUNES (z) -:- pyro [tcn@cci-209150250079.clarityconnect.net] has joined #tunes pyro: hey, any progress with retro? -:- SignOff pyro: #TUNES (pyro has no reason) -:- pyro [tcn@cci-209150250052.clarityconnect.net] has joined #tunes air: yeah, I've got console I/O working 03:00pm dammit, my connections sucks tonight 03:10pm -:- NetSplit: lackey.openprojects.net split from sterling.openprojects.net [03:48pm] -:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [lackey.openprojects.net] -:- Netjoined: lackey.openprojects.net sterling.openprojects.net -:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us232.javanet.com]) -:- water [water@tnt-9-86.tscnet.net] has joined #tunes hello all wb * water/#tunes notes 200 emails in the last 30 hours ok, MI doth suck when two behaviors happen to intersect... yeah if there were a way to guarantee no conflicts, it'd be ok like making mi less expressive well, i've been trying to find some way to express combination... i had considered inheritance by part-objects -:- hcf [nef@me-portland-us735.javanet.com] has joined #tunes is something wrong with /.? hi hcf 04:00pm water: the only intuitive notions i can come up with are composition or simply dispatching to all possible alternatives (forking) how is forking different from mi? no... it's not different i mean how to handles intersection of behavior oh if not any single one, then execute both! :) but alas, that isn't really beneficial ok i agree those are the two intuitive possibilities but i only think composition will work or the intersections of the two different methods themselves :) most likely mi will only work for co-inductive objects but then this lead me to interfaces... and how you must specify in what way you expect the given objects to act... so that it can unambiguously decide the correct behavior from a set of behavior hi water so? abi has ~3000 less factoids now hcf: cool water: rmm, i'm not saying 'pro' or 'con,' just that interfaces allow the problem to be solved in said way... k there are also aspects to consider for composing objects but weaving as it has been implemented is difficult to meta-progran wow. a lot of posts to the squeak mlist are concerning syntax ideas * eihrul/#tunes ponders. though, i've been thinking that expressing the idea of being the receiver for a method is not necessarily that useful when expressed as the child of both the receiver and a prototypical method while it does handle the scope semantics really well hm? it makes manipulating the method and the receiver individually more difficult why? unless either parent can be referred to by name :) in which case, then it's fine 'receiver parent' eh? parent is just an accessor method well... with MI, you have multiple parents :) 04:10pm oh you 're discussing mi still so having an implicit 'parent' slot is not as useful when there's multiple parents of course not i only did that because i wanted mi kept out of the core language in that case, mi still allows an intuitive definition of a receiving a message :) in what case? it binds the particular object and behavior together the receiver and method, respectively what is 'it'? it, the use of mi to express sending a message to a receiver what are you talking about? the idea of binding a receiver to a method as in message sending wtf does mi have to do with that? well, i'm thinking in terms of scopes... you want the closure to *inherit* the argument and receiver?!? no... the closure inherits the receiver and its prototypical closure uh that seems to break the conceptual model where at? i mean that a closure should not have the behavior of the receiver it should *invoke* it depending on which conceptual model... the one that makes sense, of course :) 04:20pm in terms of scopes it does... i hardly find it intuitive that a function should start acting like one of its arguments that's crap in that the scope within the method is both the scope of receiver and the method well... if you consider a behavior a temporary extension of an object :) sorry dude i can find a big flaw one would not want one's closure able to respond to messages that have nothing to do with the computation especially if another object sends the messages (hm Slate's syntax resembles NeoLogo's somewhat) * eihrul/#tunes ponders... from within the method itself it's somewhat desireable, oh well... just a thought nothing consequential we can get the same functionality without mi yeah, with relative addressing! :P dude this debate has been repeated identically within oo lang design for the last 20 years and every time mi breaks concepts in exchange for no additional expressivity you may as well use macros as mi i never said mi was perfect... and in it's current state i wouldn't use it either mi doesn't have a 'state' but it works nicely for specifying unions so long as their are no intersections :) it's a concept that doesn't work for inductively-defined objects at all the word has a state... in the form of the ideas behind it screw the word mi works only for co-inductive ideas which most languages don't use hm a few smalltalkers seem to dislike message-passing syntax for flow-control well, what others options are there? unification, special purpose control constructs... 04:30pm singleton objects supported by the vm (:P) eh? how does that tie in with flow control like true and false oh... but they don't signifigantly effect message-passing syntax :) as applied to flow-control... well i don't even like the singleton idea * eihrul/#tunes is more a fan of unification. i mean, it's okay if you compile it into the vm because you like it but in the core language it's a terrible thing within slate, we can get singletons very easily though how do you put unification into an OO framework? :) however, we still haven't demonstrated anything near the idea that we can compile bmo stuff dynamically into the vm eh? nothing, just a branch well, if you have a reification of the vm... i don't see what's too impossible about the matter just a matter of making a good compiler for it hm. are we ready yet for describing the run-time model? might as well try if we're not... we'll find out k hm maybe not now :) i guess i'm busy enough digesting incoming mail 04:40pm -:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us735.javanet.com]) -:- hcf [nef@me-portland-us735.javanet.com] has joined #tunes well, how are we on namespaces' issues? what was the issue? not *the* issue all the issues with namespaces for instance implementation, syntax, etc oh... as well as consistency and the conceptual model the relative addressing i consider an unresolved issue :) i didn't know it existed explain your ../ oh namespace shifters hm i need better names for those they specify relative locations sort of how sort of? well the problem is of course that references don't have to form a tree so it's not about "location" otoh... hm then what does ../ actually mean... ok i think i have a way to explain it the problem is that ../ only works if you have an unambiguous way to say the "a resides in b" 04:50pm you do... somewhat but only in the inverse er, converse i mean to me, a possible solution is to have an implicit container for the evaluator at all times so that all objects created and destroyed by default are part of an object representing that container that then says the global namespace is the namespace of the evaluator... hm maybe and maybe not >:) the namespace of the evaluator could cut through all the namespaces well, it would have to :) but it would be like having /usr/bin/ in your path variable roughly hm who needs namespaces anyway? :) intentional programming... that's it heh um no besides, how else do we have two different methods of the same name ? eh? well if all methods are just objects, then objects are just namespaces for those objects s/those objects/those methods/ yeah, and? :) so i think that idea nicely generalizes to a global (world-wide) namespace system what, that the namespaces of objects that are a consequence of the evaluator are within the namespace of the evalutor as well? huh? ok, you lost me somewhere then i'm saying containership as a metaphor is useful, not just expedient :) i mean it's a nice intuitive model of a system well, i thought we already had it... or almost did 05:00pm using what? well, that would hinge on pseudo-multiple-inheritance anyway you look at it :) er... perhaps not yuck then how else do you propose to introduce containership? look you're going to have to have some phenomenally-persuasive arguments to make me even consider mi "a references b" => "a contains b" er no doh! "a references b as a slot" => "a contains b" well, isn't it somewhat binary in that an object needs to know where its contained? :) no the object doesn't need to know the *evaluator* (aka system) does oh yes i remembered an idea i got recently so objects outside of the evaluator don't quite have a container... which addresses mi sigh eihrul: no! though no objects can exist outside of it not true well, if the evaluating knows where objects are contained but they don't... where are you getting this stuff? what happens if an object outlives the evaluator? oh god evaluator /= program it's a persistent object like everything else yes... besides all the info is in the system image but umm, reexplain your idea more clearly please :P "more clearly" being akin to "go find a black hole" no... it means, clarify... -:- lar1 [larman@1Cust198.tnt20.sfo3.da.uu.net] has joined #tunes i'm saying that the containership relationship will effectively form a file system eihrul: i know what the fsck it means -:- fire [no@209-68-229-223.dialup.cust.tfb.com] has joined #tunes hey all an object doesn't *know* where it belongs the system merely uses it where the *system* knows it is that's fine.... amen water :) i already dig that... to me, a possible solution is to have an implicit container for the evaluator at all times so that all objects created and destroyed by default are part of an object representing that container oh i see the problem in your parsing ;) that implcit container is not fixed, dude 05:10pm in that the containment relationship for the evaluator is determined as it is for all objects? ugh well, define 'implicit container' that's quite vague... the evaulator indeed is a system object no it ain't yes it is... water: can you compile with squeak? it's the same as having a "current path" in a shell if it doesn't mean the same containment relationship as for other objects then it means something else quite vague... fire: compile to what? machine code so if it is not such, explain executable fire: not within the squeak image i see fire: not yet ;) bummer water: eh? eih: current path yes, i'm damned well aware of that... but current path is equally vague... do i even have to explain that? anyone here used lex and can help me? grrrr look.... fire: not here in that the containment relationship for the evaluator is determined as it is for all objects? that's what you're saying to me... :\ eih: you're not helping me help you :) i don't get what you're asking in that for C contains an object created by the evaluator, and D contains an the evaluator, C = D... this is what's getting grokked dude *every object is created by the evaluator yes.... but *every* object in every namespace this rules out the notion of nested evaluators :) no it doesn't explain how you think it does as i said... don't reiterate, explain for C contains E, and O is created by E, D contains O, and C = D initially... speak english!!!!!!!!!!! grrrrrrrr the container of the evaluator is the container of the objects it creates no!! to me, a possible solution is to have an implicit container for the evaluator at all times so that all objects created and destroyed by default are part of an object representing that container the evaluator creates objects in the currently *selected* namespace the selection is implicit to object creation this selection exists independently of the evaluator then? yes just like in a shell and hence is not the same kind of 'implicit container' as you inferred it was... it's implicit to evaluation, isn't it? well, when you say container... i take it to mean the normal containment relationship container of *what* 'implicit container' don't assume ask you asked 17 wrong questions 'an implicit container for the evaluator' FOR NOT OF they can be used very similarly, afaik :) a container implicitly used by the evaluator an implicit evaluator for the evaluator to create objects within 05:20pm this is so FSCKing frustrating * eihrul/#tunes hates English. how's your spanish? it's just a shell metaphor, idiot * eihrul/#tunes sighs. rusty but good IDIOT... it's not that that is anyway what's confusing i alluding 10 times to shells your ambiguous use of terms that you yourself define is what is confusing eihrul: Use <>! dude you need to get a new parser or something i bet lar or fire would easily not have been confused fine hell even atg would probably not have had problems then talk to them for help... no i like you most of the time but you assume without questioning your assumptions which really irks me to no end well, i didn't know you so hated assumptions... i'm a mathematician at heart but please refrain from name-calling i don't feel it's appropriate to the discussion nor useful, in any way well i got pissed off a long time ago i had to express it and i don't think the landlord would like it if i broke all my walls well, you express it at the expense of others... oh so i should hit myself? :) or something that won't hit back... i don't have any place for things like that but anyway... namespaces my only comfort is taoism/zen, and they can't be invoked while talking to coders now you grok? yes... it was just an incorrect assumption indeed