IRC log started Sun May  7 00:00:01 2000
[msg(TUNES)] permlog 2000.0507
-:- SignOff eihrul: #TUNES ([x]chat)
-:- SignOff hcf: #TUNES (Leaving)
-:- smoke [smoke@15dyn249.delft.casema.net] has joined #tunes
<mog> smoke?
<smoke> yes?
<mog> i thought u were someone else :)
02:20am
-:- SignOff smkl: #TUNES (bbl)
-:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes
* Fare/#Tunes is back
-:- SignOff smoke: #TUNES (Ping timeout for smoke[15dyn249.delft.casema.net])
-:- smoke [smoke@15dyn249.delft.casema.net] has joined #tunes
-:- SignOff smoke: #TUNES (brb)
-:- smoke [smoke@15dyn249.delft.casema.net] has joined #tunes
-:- dalvarez [ircusr@212.68.72.98] has joined #tunes
-:- SignOff dalvarez: #TUNES (Ping timeout for dalvarez[212.68.72.98])
-:- SignOff smkl: #TUNES (Ping timeout for smkl[glubimox.yok.utu.fi])
-:- eihrul [lee@usr5-ppp115.lvdi.net] has joined #tunes
-:- AlonzoTG [Alonzo@216-164-130-209.s590.tnt2.lnhva.md.dialup.rcn.com] has joined #tunes
-:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes
-:- Connection closed from irc.us.openprojects.net: Success
-:- Use /Server to connect to a server
-:- Connecting to port 6667 of server irc.us.openprojects.net [refnum 0]
-:- BitchX+Deb1an: For more information about BitchX type /about
-:- Welcome to the Internet Relay Network TUNES (from king.openprojects.net)
-:- Your host is king.openprojects.net, running version u2.10.05.18.(ipcheck4-5) (from king.openprojects.net)
-:- This server was cobbled together Wed Apr 28 1999 at 12 02:19 EDT(from king.openprojects.net)
-:- king.openprojects.net u2.10.05.18.(ipcheck4-5) dioswkfcg biklmnopstv 
-:- [local users on irc(66)] 8%
-:- [global users on irc(414)] 47%
-:- [invisible users on irc(462)] 53%
-:- [ircops on irc(16)] 2%
-:- [total users on irc(876)]
-:- [unknown connections(0)]
-:- [total servers on irc(28)] (avg. 31 users per server)
-:- [total channels created(350)] (avg. 2 users per channel)
!king.openprojects.net Highest connection count: 146 (144 clients)
!king.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.dhs.org "
-:- TUNES [tunes@bespin.dhs.org] has joined #tunes
-:- Topic for #TUNES: TUNES, Free Reflective Computing System: http://www.tunes.org/ || Slate Language: http://www.tunes.org/~water/slate-home.html
-:- topic set by hcf [Fri May  5 20:43:58 2000]
-:- [Users(#tunes:10)] 
[ TUNES     ] [ smkl      ] [ AlonzoTG  ] [ eihrul    ] [ smoke     ] 
[ mog       ] [ thomas    ] [ corey     ] [ abi       ] [ Fare      ] 
-:- Channel #tunes was created at Sun Feb 28 08:48:06 1999
-:- BitchX+Deb1an: Join to #tunes was synced in 7.474 secs!!
-:- Mode change [-ws] for user TUNES
-:- dalvarez [ircusr@212.68.72.98] has joined #tunes
-:- dirt [mindfvck@niantic0304.mohawk.net] has joined #tunes
-:- SignOff dalvarez: #TUNES (chillout in the evening's sun)
-:- SignOff AlonzoTG: #TUNES (Ping timeout for AlonzoTG[216-164-130-209.s590.tnt2.lnhva.md.dialup.rcn.com])
-:- AlonzoTG [Alonzo@216-164-130-209.s590.tnt2.lnhva.md.dialup.rcn.com] has joined #tunes
-:- SignOff AlonzoTG: #TUNES (Ping timeout for AlonzoTG[216-164-130-209.s590.tnt2.lnhva.md.dialup.rcn.com])
-:- Downix [down@d-ma-boston-159.ici.net] has joined #tunes
<Downix> who runs the OS Review project?  I found an Os not on there
08:50am
-:- SignOff smoke: #TUNES (brb)
-:- smoke [smoke@15dyn249.delft.casema.net] has joined #tunes
<Downix> hey smoke, who maintains the OS Review subproject?
<smkl> perhaps core
* Downix/#tunes laughs.
09:50am
<smkl> just send a mail to review@tunes.org
<Downix> ok
10:00am
-:- mog [xdef@209-6-184-102.c3-0.wth-ubr1.sbo-wth.ma.cable.rcn.com] has left #tunes []
-:- lar1 [larman@adsl-63-204-135-211.dsl.snfc21.pacbell.net] has joined #tunes
-:- iStormy [stormy@rain.futuresouth.com] has joined #tunes
-:- water [water@tnt-10-225.tscnet.net] has joined #tunes
<water> hey all
<lar1> hey
12:00pm
<iStormy> hi
<water> hi storm, what's up?
12:10pm
<iStormy> i'm reading 6 months of comp.os.misc and it reminded me of you guys.
<iStormy> only 174 threads in 6 months, sad.
-:- vvater [water@tnt-10-225.tscnet.net] has joined #tunes
<iStormy> i guess that means you missed my answer to you.
<iStormy> i'm reading 6 months of comp.os.misc and it reminded me of you guys.
<iStormy> only 174 threads in 6 months, sad.
<iStormy> there you go
-:- SignOff water: #TUNES (Read error to water[tnt-10-225.tscnet.net]: Connection reset by peer)
-:- vvater is now known as water
12:20pm
<iStormy> why did newsgroup spammers all decide to add a 4 digit number to the end of their subject lines, allowing us so easily to killfile them?
-:- jack3lit [jack3lit@lar63.netscorp.net] has joined #tunes
<water> hi jack
<jack3lit> hey water
<jack3lit> what are you guys talking about?
<water> hm nothing at the moment
<water> you know what Tunes is?
<jack3lit> how fun
<jack3lit> no
<water> http://www.tunes.org/
<jack3lit> what is it?
<water> abi: Tunes?
<abi> i guess Tunes is a free reflective computing system at http://www.tunes.org or for programming languages what the internet is to networks or like flypaper for hallucinogenic programmers or discombobulated or obnubilated or AOTA
12:30pm
<water> yeah that about sums it up :)
<jack3lit> okay well thanks and bye!
<water> sure
<water> no questions besides that?
<water> oh great, now we have another lurker, most likely just as un-educated as the rest
<iStormy> i don't think he'll lurk long once he figures out we don't chat
<water> well if eihrul ever wakes up, maybe there'll be some decent discussion
-:- SignOff jack3lit: #TUNES ()
<iStormy> 
<iStormy> i wonder why bx defaults to repeating the last line in the channel in so many instances
<water> unless you have something you'd like to talk about?
<eihrul> water: awoke
<water> ask #linpeople
<water> what's up, eih?
<iStormy> my first forth word, converts hours minutes seconds to seconds: : hms rot 3600 * rot 60 * + + ; 
* iStormy/#tunes bows, no applause please.
<water> heh nice word
<eihrul> reading papers
<water> as am i 
<water> i'm still really bugged by the current syntax
<water> the code model is fanatastic, but making it coherent is a bit difficult
12:40pm
<eihrul> that tim sheard paper reminded me very much of lisp :)
<water> which one was that one?
<water> (sorry i've had a terrible week and a lot of my memory caches have been flushed)
<eihrul> interpreters to compilers through stating and monads (sorry, was getting food)
<water> oh yeah
<water> wow. very few of the usual irc nets i use are letting me on today
<water> everyone must be busy "loving" each other ;)
<eihrul> i've been looking through some of the compose papers
<eihrul> and they make it seem very bleak as to how to determine automatically what to partially evaluate
<water> hm
<water> yeah but that just means they didn't figure it out
<eihrul> i.e. estimating the speedup from partially evaluating 
<water> right
<eihrul> and most importantly, given a set of variables parameterizing a block of code, partially evaluate the wrt the variable that gives the best speedup :)
<water> probably very difficult with high levels of polymorphism
<eihrul> yes
<water> well, just default to a profiler and lazy optimizer
12:50pm
<eihrul> Self doesn't even handle it all that well, but it does handle it automatically
<water> well self is impressive almost only for the reason that its optimizations are totally transparent
<eihrul> yes
<eihrul> although it does illustrate a possibility:
<eihrul> partially evaluate everything possible that will give some speed benefit
<eihrul> and just use smart choices in which code to get rid of when you have to make room for more
<water> hm
<water> not sure about the second part there
<water> but ok
<eihrul> i.e. you could for instance eject the partially evaluated code that has decreased the least in size
<water> oic
<eihrul> as a very simple, but possibly good heuristic
<water> sort of like selecting partial-evaluations as in genetic algorithms? :)
<eihrul> yes, only the good ones remain
<eihrul> and also, if one partial evaluation results in way too much memory use
<eihrul> you could try partially evaluating less variables in the block
-:- water has changed the topic on channel #tunes to: TUNES - A Free Reflective Computing System - http://www.tunes.org/ || The Slate Language - http://www.tunes.org/~water/slate-home.html
<water> perhaps
<eihrul> however, then its back to the question of "which" :)
<water> well there's the possibility suggested by tunes docs which has no precedent
<eihrul> where?
<water> that there's some way to add meta-information to manually select optimizations while not breaking a uniform object model
* water/#tunes looks
<eihrul> annotations?
<water> yeah
<water> but tunes-style per se is definitely unprecedented afaik
<eihrul> but during idle cycles you could always breed new partially evaluated blocks and see which are more efficient :)
<water> too what end, though? it all could be for naught
<eihrul> so that when someone goes to the bathroom, they find their system running faster :)
<water> heh
<water> hm
01:00pm
<water> well there's nothing wrong with a thread (or thread-per-memory space) totally dedicated to optimization
<eihrul> yeah, that's the way i wanted to structure it
<water> and we can use language semantics to handle scheduling frameworks
<water> how well do you think the directory metaphor will work with programming efficiency?
<eihrul> well, it could be somewhat detrimental
<eihrul> well, not the directory itself perse... but any object is open to inspection
<eihrul> and this works against optimization
<eihrul> since you ideally should be able to observe the state of an object
<water> hm
<water> well the UI tools could be shaped to avoid overloading the system that way
<water> keep in mind of course that Morphic already has the same problem
<eihrul> a possible solution is that employed by jalapeno
<eihrul> for garbage collection
<eihrul> assuring that garbage collection only happens at specified safe points
<eihrul> which is basically just another kind of invasive inspection :)
<water> hm
<eihrul> but i don't know if there solution was particularly hardware-aided or software
<eihrul> s/there/their
01:10pm
<eihrul> but actually, it is quite dangerous to interact with objects that are already being worked upon
<eihrul> other than simple inspection
<water> hm
<eihrul> unless you completely halt the computation
<water> btw i've been considering grouping together parts of an object's protocol into sub-protocols for various reasons
<eihrul> and then insert your changes, and restart (much like a debugger)
* water/#tunes nods
<water> basically you have to fool with execution preferences in order to get a look at your computations
<water> it would be best to make this clear to the user
<water> as opposed to "here's the tool, go poking around and see where your framerate goes" :)
<water> anyway, i was thinking along the lines of the tunes idea of only providing the user with the relevant aspects of an object for a particular task
01:20pm
<water> but it's difficult to formulate a solution that's truly robust
<water> here's an example, though: we could group structural protocols, like addSlot etc, into one "category" implemented as a sub-object of root
<eihrul> though, addSlot has always seemed like the function of the meta-object to me
<eihrul> er rather a
<water> well have you thought of a good syntax for it?
<eihrul> nope
<water> btw there's a bit of a difficulty with namespace syntax
<water> it's intuitively obvious that you should be able to clone objects around by trying to set the ".." slot
<water> which is fine but it sort of violates the idea that "clone" per se is primitive semantically
<water> i guess i could base "clone" on ".. :"
<water> the problem is that a namespace is supposed to *contain* (intuitively) its objects
<water> so some concept has to give
01:30pm
<water> any thoughts/comments?
<eihrul> thinking
<eihrul> none right now
<water> i guess to elaborate, i should remind that the concept of namespace as container right now means the objects it contains are methods it can invoke in some sense
<water> :P
01:40pm
<eihrul> well, i'm just trying to relax and make a dent in my paper collection today :)
<water> oh ok
<water> yeah i guess i need to do those things as well :)
01:50pm
-:- SignOff smoke: #TUNES (z)
-:- dalvarez [ircusr@212.68.72.98] has joined #tunes
<water> hey dalvarez
<dalvarez> greetings...
<dalvarez> abi: seen Fufie?
<abi> Fufie was last seen on IRC 26 days, 2 hours, 46 minutes and 24 seconds ago, saying: hangs himself with a coax cable [Tue Apr 11 11:27:43 2000]
<water> hm sounds like a final message :)
02:20pm
<water> abi: tell dalvarez about beep
<water> gah
02:30pm
* water/#tunes is conversing on other channels right now, in case anyone cares where the activity is
-:- hcf [nef@me-portland-us427.javanet.com] has joined #tunes
<water> hey hcf
<hcf> hey
<water> syntax design issues suck, icuc
<water> i think Fare was smoking some seriously hallucinatory drugs when he wrote up some of the HLL design requirements
<Downix> yup
<water> every once in a while i say "screw it, why not just make arrow and forget all this language bullshit?"
<water> and then i remember very few people even grok arrow, and that slate's supposed to bootstrap the damned thing
* hcf/#tunes is away: (afk)
02:50pm
<water> maybe i should center Slate around the objects-as-function-expressions and derive everything from that and a few oo primitives
<water> comments/ideas?
<corey> water, what part of the design requirement dont you like?
<water> the problem with that per se is the fact that part of oo is necessarily about being able to vary the fine-detail of object identity, which fp doesn't allow
<water> hm
<water> well the "views of objects" idea for one
<water> where only the relevant aspects show themselves at a give time / interface
<water> that's a UI feature, but meta-programming it is intractable
-:- SignOff dirt: #TUNES (Leaving)
<water> hm?
<water> damn lurkers
<water> corey: does that help clarify?
<corey> not really
<corey> but ill add it to the collective
<water> well, read the HLL blurbs
<corey> i am
<corey> i read the views section
<water> unfortunately my http access is slooooooow today
<corey> you should mirror it locally.
<water> on my windows partition?!?
<water> my modem doesn't play with linux or beos
<corey> winmodem?
<water> no, pci pnp
<water> why do people always ask that like i'm some kind of idiot?
<corey> nm.
<corey> i had a nice courier v.everything external that i would have given to you, but i gave it to a girl.
<water> well i'm honestly sick of dealing with computers any more than project involvement absolutely requires
<water> anyway, Tunes HLL specs...
<corey> yeah
<corey> anything else?
<water> well i haven't even loaded up the page yet :P
-:- SignOff dalvarez: #TUNES (Ping timeout for dalvarez[212.68.72.98])
<corey> cant you share the connection on the windows box?
<corey> nm you dont want to talk about it.
<water> share it with what??
<corey> the mandrake one
<water> it's the same box
<corey> oh heh
<corey> sorry
03:00pm
<water> i have 3 oses on one box dude
<water> actually, i'll reconnect and see if that fixes it
<corey> kay.
<water> brb
-:- SignOff water: #TUNES (The Tao went that-a-way!)
-:- water [water@tnt-9-130.tscnet.net] has joined #tunes
<water> re
<corey> yup
<water> ok http is much faster now, got the page loaded
* corey/#tunes mumbles something about windows.
<water> ok Fare is totally clueless as to quotienting
<corey> hmm
* corey/#tunes reading.
<water> i talked to him about this entry just a few days ago
<corey> whats quotienting involve?
<water> heh
<water> treating a class of objects as the same thing
<water> or vice versa treating a single thing as a class of objects
<corey> how is fare totally clueless and how would you educate him?
<water> i wouldn't, it's not worth it
<corey> what do you mean?
<water> he's stuck in the "lisp is the greatest thing ever made" mindset
<corey> he doesnt need to know or you dont think hes worht talking to about this?
<water> the second one
<corey> the shining vision problem
<water> and i've already tried, as i stated before
<corey> blinds you to everything else
<water> including problems with lisp as a tunes candidate
<corey> why is slate better than lisp?
<water> well because it's not firmly bound and gagged to Cons cells, for one
<Downix> l(e(s)s) b(r(a(c(k)e)t)s)
<water> and also it supports an open lambda system which is as exploratively useful as Smalltalk / Self
<water> lol
<water> well, Slate's objects form directories of a sort
03:10pm
<water> which is like having all those brackets except it's simpler to inspect objects that way
<water> also Slate's objects are sets and not lists which means a higher level of abstraction (useful for a varied set of things)
<corey> a list is a ordered set?
<water> hm Fare's "Security" blurb on hll specs is good but totally inadequate for designing a HLL
<water> sort of
-:- SignOff hcf: #TUNES (Leaving)
<water> a list is an ordered multi-set
<corey> whose maintainer of the hll?
<water> hmmmm good question
<corey> water: why not you?
<water> because i'm not an orthogox tuneser it seems
<water> oh yeah, i forgot to mention i'm not a member :)
<corey> seems to me you would be the prime candidate
<corey> but you have your own projects.
<water> yeah well i'm also very busy
<water> exactly
<water> ok, on to my gripes about "HLL Semantics"
<corey> heh
<water> attirbutes is good... i have that covered so far
<water> context sucks really hard
<corey> why?
<water> well it's way too vague... people working on natural language have a better idea about context than Fare does
<corey> do you guys ever talk to each other?
<water> also, the way Fare uses it is very far beyond what most programming languages ever get into
<water> every few days or so
<water> he's usually not responsive enough
<corey> i see
<water> rewrite is all right
<water> i can handle rewrite, but it's via my meta-objects etc
<water> (as for context, my meta-objects handle most of those issues)
<water> reflective rewrite a la Maude can be handled by a pure language like Slate as long as it uses meta-behavior
03:20pm
<water> the rest of the stuff i can handle
<water> although lambda abstraction per se doesn't have a special construct in slate
<corey> access rights?
<water> well insofar as i can handle context, access rights are simple
<water> like right now, access rights can be handled by the fact that meta-objects intercept and can examine all messages
<water> they get to know who the sender is and can screw with the lookup process dynamically in order to implement all sorts of secureity features
<corey> i understand what your saying in regard to this
<corey> heh
03:30pm
<corey> hmm
<corey> types?
<water> hm
<water> well, extensionally, OO already provides types as objects' protocols
<corey> clarify
<water> the protocol of an object is the set of messages it responds to and what types it returns (per input values)
<corey> ok
<corey> i understand
<water> with meta-behavior, there's an added amount of flexibility
<water> as an example, i can implement number systems or algebras using slate objects
<water> of course it helps implementationally if i plug them into the right hardware operations ;)
03:40pm
<water> anyway, i can build lots of formally-verifiable type systems out of that
03:50pm
<water> does that answer a bunch of questions or no?
<corey> :)
<water> hm sounds like a "yes"
<water> keep in mind there are a LOT of issues to work out still
<water> mostly they are syntactical, though
<water> the semantics are relatively consistent
-:- SignOff eihrul: #TUNES (Read error to eihrul[usr5-ppp115.lvdi.net]: EOF from client)
-:- eihrul [lee@usr5-ppp115.lvdi.net] has joined #tunes
<water> re
<eihrul> re
<water> hm "evaluation strategy" sounds like a page from the Scheme spec
04:00pm
<water> doo dee doo
<eihrul> ?
<water> oh the conversation sort of died here
04:20pm
<water> eihrul: got any slate issues to talk about, even for later?
<water> lar1, corey, *anyone*: got stuff to discuss?
04:30pm
<water> PING
<eihrul> no
<water> okay
<water> i'll likely be on again late tonight
<water> how soon will you have a workable compiler, btw?
<eihrul> no idea
<water> not even a range?
<eihrul> maybe a month to implement
<eihrul> modulo specification time
<water> ok so it *does* depend on the stuff i'm working on? or do you mean compiler spec?
<eihrul> well, if the compiler is to be implemented in slate
<eihrul> then it would
<water> hm
<eihrul> otherwise, i was talking about compiler spec :)
<water> btw you know that vortex is written in cecil, right?
<eihrul> have not looked at vortex 
<water> ok
<eihrul> cecil's compiler?
<water> yep
<eihrul> will look then
<water> thx
<eihrul> but, independently of the slate spec
<eihrul> i can have a compiler to play with in lisp pretty easily
<water> ok
<eihrul> though, also, wrt to partial evaluation
<eihrul> there are definitely two parameters that should be partially evaluated
<eihrul> those being the 'map' and the meta-object
<water> er how are you using the PE term?
<water> you want more detail i suppose?
<eihrul> partially evaluating an object's methods wrt one of its other slots
<water> hm
<eihrul> or rather, 'self'
<water> ok that makes more sense
<water> well it depends on what primitive types you provide with the compiler. really
<eihrul> for meta-spaces, yes... structure, no
04:40pm
<water> hm
<water> well Self had primitive types
<eihrul> even then, how it optimized lookups of slots and methods
<eihrul> was via maps
<water> yeah
<water> don't maps work well with the slate object model already?
<eihrul> well, they're sort of orthogonal to bmo's
<water> isn't that good, though?
<eihrul> yes
<water> hm so is there a problem or is there an issue too vague to resolve?
<eihrul> nope
<water> ok
<eihrul> but whether to introduce optimizations like maps in the meta-objects themselves
<eihrul> or to put maps below meta-objects themselves 
<water> well even the bmo's can be composed of other objects, a la guarana
04:50pm
<water> toodles, bbl
-:- SignOff water: #TUNES (The Tao went that-a-way!)
-:- SignOff lar1: #TUNES (Hell yes!)
05:00pm
-:- Kyle_L [kyle@cr168790-a.nmkt1.on.wave.home.com] has joined #Tunes
-:- SignOff Downix: #TUNES (Read error to Downix[d-ma-boston-159.ici.net]: Connection reset by peer)
-:- hcf [nef@me-portland-us525.javanet.com] has joined #tunes
-:- mog [xdef@209-6-184-102.c3-0.wth-ubr1.sbo-wth.ma.cable.rcn.com] has joined #tunes
-:- SignOff mog: #TUNES (Read error to mog[209-6-184-102.c3-0.wth-ubr1.sbo-wth.ma.cable.rcn.com]: EOF from client)
-:- ult [ult@user-38lc63s.dialup.mindspring.com] has joined #Tunes
<ult> hmm hey
-:- SignOff ult: #TUNES (Leaving)
07:40pm
-:- lar1 [larman@adsl-63-204-135-211.dsl.snfc21.pacbell.net] has joined #tunes
-:- SignOff eihrul: #TUNES ([x]chat)
-:- water [water@tnt-10-37.tscnet.net] has joined #tunes
<water> hey all
<water> amazing: i leave the channel, and it goes dead silent :)
09:30pm
-:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us525.javanet.com])
-:- eihrul [lee@usr5-ppp115.lvdi.net] has joined #tunes
<water> re eih
<eihrul> re
09:50pm
-:- SignOff Fare: #TUNES (Ping timeout for Fare[lantier.enst.fr])
-:- hcf [nef@me-portland-us143.javanet.com] has joined #tunes
-:- smoke [smoke@16dyn66.delft.casema.net] has joined #tunes
-:- SignOff eihrul: #TUNES ([x]chat)
-:- SignOff smoke: #TUNES (Ping timeout for smoke[16dyn66.delft.casema.net])
-:- smoke [smoke@16dyn66.delft.casema.net] has joined #tunes
-:- SignOff lar1: #TUNES (Hell no!)
-:- SignOff hcf: #TUNES (Leaving)
-:- SignOff iStormy: #TUNES (Bye)
[msg(TUNES)] newlog 2000.0508
IRC log ended Mon May  8 00:00:01 2000