IRC log started Sat Mar 25 00:00:01 2000
[msg(TUNES)] permlog 2000.0325
-:- Jinli6819 [Jinli6819@tns01419.singnet.com.sg] has joined #tunes
<Jinli6819> Hello! Everyone.
-:- SignOff Jinli6819: #TUNES (Bye Bye)
-:- SignOff ult: #TUNES (Leaving)
12:40am
-:- Lothos_ [lothos@cdslppp26.ptld.uswest.net] has joined #tunes
<Lothos_> hello :)
01:30am
-:- SignOff eihrul: #TUNES ([x]chat)
<Fufie> hello
01:50am
<Lothos_> is this channel usually very active?
<Fufie> no
<Fufie> people are usually busy in other windows
<Lothos_> heh, ohh well :)
<Lothos_> i'm writing an OS, I thought it might be a good idea to lurk here and absorb any insights you all might have
02:50am
<Fufie> that sounds like a decent idea
03:00am
-:- SignOff water: #TUNES (Ping timeout for water[tnt-9-232.tscnet.net])
-:- SignOff Fufie: #TUNES (brb)
-:- dalvarez [ircusr@212.68.72.98] has joined #tunes
-:- SignOff smoke: #TUNES (Ping timeout for smoke[16dyn86.delft.casema.net])
-:- mathias is now known as mathias-basket
-:- Fufie [stig@tunnel-44-26.vpn.uib.no] has joined #tunes
-:- smoke [smoke@16dyn164.delft.casema.net] has joined #tunes
-:- eihrul [lee@usr5-ppp96.lvdi.net] has joined #tunes
<eihrul> los
07:20am
-:- SignOff smkl: #TUNES (blah)
-:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes
-:- SignOff Fufie: #TUNES (Ping timeout for Fufie[tunnel-44-26.vpn.uib.no])
-:- Fufie [stig@tunnel-44-26.vpn.uib.no] has joined #tunes
-:- SignOff Fufie: #TUNES (Ping timeout for Fufie[tunnel-44-26.vpn.uib.no])
-:- water [water@tnt-9-186.tscnet.net] has joined #tunes
<water> 'lo
<eihrul> hey
08:00am
<water> who's lothos?
<eihrul> no idea... a newcomer i guess
<water> ok, hi lothos
<water> hm i have a lot of new papers to scan for ideas
<eihrul> on?
<water> subjects are quite varied
<water> all from RI, though
<water> mostly they seem to relate closely to one or another aspect of slate
<water> btw i always throw away papers that mention C++ :)
08:10am
<eihrul> heh
<eihrul> what about java papers?
<water> those get scutinized before downloading
<water> otoh i do the same with ML papers
-:- tcn [r@cci-209150250174.clarityconnect.net] has joined #tunes
08:20am
<water> the same issue gets all 'typeful' when you're discussing ML
<water> hi tcn
<water> tcn: make sure you give nate the low-down on forth
<water> tcn: he'll be around again on sunday icuc
<water> so the only thing ML'ers ever add to research is to get long-winded about type systems, it seems :)
<tcn> hey water
<eihrul> luckily i haven't read any real ML papers, yet...
<water> well if you're looking on dissertations about types, look no further :)
<water> hm i might wind up throwing out as much as half of these new papers (who knows?)
08:30am
<tcn> heh
08:40am
<water> does anyone know anything about 'comprehensions'?
<smkl> like list or set comprehensions?
<water> yes
<water> apparently there are some for monads, and there's a few papers here about collections in general
<eihrul> having hcf withdrawals? :)
<water> heh :)
-:- Fufie [stig@tunnel-44-26.vpn.uib.no] has joined #tunes
<water> hey fufie
<water> what can you tell me about comprehensions?
08:50am
<Fufie> it's a long word
<Fufie> it might be difficult to comprehend
<water> bah :)
<water> i thought you might know about list comprehensions for some reason
<Fufie> list comprehensions?
<Fufie> you use many long words water, where is this taken from?
<water> well foldoc has an entry
<water> wadler uses them a lot in his papers
<water> and some recent papers have things like set and collection comprehensions
<Fufie> Sorry, the term comprehensions is not in the dictionary
<water> no, list comprehensions
<Fufie> sounds pretty invented to me
<water> why?
<tcn> hey.. I found a Forth GC
<eihrul> wow...
<eihrul> the seventh sign, indeed
<water> if it can be generalized to an abstraction over sets or general collection types, then it seems to have more substance than just "sugaring"
<Fufie> *shrug* maybe it gives meaning to the people who started using it.. 
<water> bah you're just being a lisp evangelist now
<eihrul> water: constructor of sorts?
<Fufie> nah
<water> well how do you justify your stance?
<Fufie> four noted uses of the word.. and they give an example from Haskell which doesn't use the term
<smkl> there are set comprehensions in set theory, like {x | 0 < x < 1} 
<Fufie> very nice
<smkl> list comprehensions are copied from there
<water> hm
<Fufie> smkl: yes, doesn't make it a valid term except if you treat lists as sets though
<water> i've read a lot on set theory but i've never seen that notation called a "comprehension"
09:00am
<Fufie> seems as if theorists run out of words ;)
<water> what else can they do?
<water> invent nonsense words?
<Fufie> invent brand new words
<water> such as?
<Fufie> like list-zummies
<eihrul> well, it is better than overloading terms already occupied
<water> yeah that'll get them respect ;)
* eihrul/#tunes shrugs.
<eihrul> i don't think fufie is respecting them much now for their nomenclature :)
<smkl> list comprehensions are a lot like set comprehensions though
<Fufie> I thought the point was making good terms and establishing terms.. not getting respect
<eihrul> well, the former can cause the latter
<Fufie> water: now that you have seen the term 'list-zummies' would you remember it?
<water> fufie: don't you mislead your supervisor about your research topic? :)
<Fufie> water: he would expect it anyway.. :)
<water> yes i would but it still doesn't tell me anything about the meaning
<Fufie> water: does list comprehensions tell you anything?
<water> and i'd doubt papers would get accepted if they had 'zummies' in the title
<water> of course not
<Fufie> worse words have been accepted as meaningful
<water> like what?
<Fufie> two seconds
<eihrul> water: yes, ivory tower nonsense... no reason a paper can't have the word 'zummy' in its title
<water> well the ivory towers are real to the people submitting papers
<water> even if just subconsciously
<water> would you hinge your phd thesis on the word 'zummy'?
<eihrul> sure!
<tcn> what aboud 'pointer swizzling'
<water> yeah well you're not even in college yet
<eihrul> it is just a symbol by which to convey an idea :)
<Fufie> creeping featuritis?
<Fufie> yuppie
<eihrul> it could be 3053 for all i care
<water> ok but swizzling is at least suggestive
<Fufie> zumming also makes sense
<water> zummy sounds like zoom though
<Fufie> you're not getting the sum (sigma) but you're getting out a result of e.g a MAP
<water> what concept would it denote?
<Fufie> I zummed the list for positive numbers
<Fufie> (map #'positive-p list)
<water> hm so zummy as in summy
<water> ok i can accept this all
<Fufie> I think I'll use that term in my thesis
<Fufie> :)
<water> and i think i'll just read the papers to figure out what comprehensions mean in general
<water> (they use a category-theoretic argument)
09:10am
<Fufie> category-theory seems to give many theorists orgasmic pleasure :)
<water> why not? it's both formal and general
<water> and anyone in fp should at least grok the basics, it seems
-:- ult [ult@user-38lcn4h.dialup.mindspring.com] has joined #Tunes
<Fufie> yes, knowing the basics is important.. a necessary evil
<water> anyway
-:- mathias-basket is now known as mathias
09:20am
<tcn> abi: coldforth?
<abi> tcn: i haven't a clue
<tcn> COLDFORTH is at www.forth.cx
<tcn> :)
<water> yeah i looked at it last week
09:40am
<tcn> heh.. the author (Charles Esson) is looking for Forth programmers to do "R&D on vision systems for the fresh fruit market" :)
<water> odd
<Fufie> sounds like a nice project
<Fufie> probably an embedded device they need to screen fruit and find the bad apples
<water> yeah
<eihrul> tcn: retro3 update?
<water> eek! i just made this weird connection between that project idea and fetuses
<water> s/weird/scary/
09:50am
<tcn> heh
<tcn> I'd be interested if it weren't on the other side of the world
<tcn> and if I were looking for a new job in the first place :)
<tcn> eihrul: not much new on retro
<Fufie> what do you do in your current job?
<tcn> database software
<tcn> or whatever we get :)
<tcn> heh.. 
-:- rares [rares@wtrb-sh1-port65.snet.net] has joined #tunes
-:- SignOff Fufie: #TUNES (Read error to Fufie[tunnel-44-26.vpn.uib.no]: EOF from client)
10:00am
<tcn> a provisioner I work told told me a story about a navy technician at an overseas base who ordered a microchip, but he wrote the part number wrong, and he got a locomotive instead
* eihrul/#tunes cackles.
<eihrul> locomotive probably costed less, no? :)
<tcn> no but that's what taxes are for
-:- SignOff eihrul: #TUNES (Read error to eihrul[usr5-ppp96.lvdi.net]: EOF from client)
<tcn> and they shipped it to a base in Saudi Arabia miles from any railroad
-:- eihrul [lee@usr5-ppp96.lvdi.net] has joined #tunes
<tcn> eih:  they shipped it to a base in Saudi Arabia miles from any railroad
<eihrul> muahaha
<tcn> oh and what else did I hear..
<water> that sounds familiar
<tcn> some dumbass fed himself to a jet engine last week
-:- Fufie [stig@tunnel-44-29.vpn.uib.no] has joined #tunes
<Fufie> unstable net today
<tcn> yeah
<water> that happens every year
<water> it doesn't help that the guys are working in 120dB areas
<tcn> right
<tcn> with double hearing protection
<water> it doesn't help
<tcn> I guess he got sucked right through
<water> those guys loose iq points by the month
<water> yes there's usually no trace of them
<eihrul> sublimination via jet engine? :)
<rares> you people are horrible :)
<tcn> does anyone get cooked by radars?
<tcn> I heard about a marine who was doing chin-ups on a waveguide..
<water> yes they do
<water> rares: i work next to this stuff
<tcn> I bet people get shocked alot too
<rares> I don't and I too find this grossly amusing
<tcn> how's 400volt/400hz compare to 120vac?
-:- XeF4 [xef4@194.255.106.103] has joined #tunes
<water> not by the radars they don't
<water> anything over 400volts has about a 70% chance of killing you
<water> although i don't remember the exact numbers
10:10am
<ult> dependso n the amperage
<water> well it's a chaotic phenonemenon
<tcn> man, we get some commercial projects where the people haven't even learned what military provisioners have been doing for centuries
<water> like fireballs
-:- hcf [nef@me-portland-us840.javanet.com] has joined #tunes
<water> there still isn't a complete theory of fireballs
<water> hey hcf
<hcf> hey water
<ult> hmm ok
<hcf> <water> apparently there are some for monads, and there's a few papers here about collections in general
<hcf> <eihrul> having hcf withdrawals? :)
<water> although i do remember a recent article (in wired?) about fireballs (ball lightning) and some causal link with the molecular level
<hcf> water: need stuff?
<water> hm
<water> not really. i have a lot to read
<hcf> k
<water> yes! i found a paper which (seems to) allow slate
<water> in its formalism, that is
<dalvarez> I have never seen #tunes as busy - is something special of which I missed the announciation taking place?
<water> lol
<eihrul> yes... the weekend
10:20am
<water> yeah, i'm here >:)
<tcn> haha
<rares> I don't think there should be a comp;lete theory of fireballs like Bill Joy says you'd have people burning each other's houses daily over spilled coffee
<water> well, it'll happen eventually
<water> but by then we'll have portable shield generators :)
<rares> how can you break into a house if you can't get in to turn off thbe shield unless they're wired to trhe Ministry of technology's Satellite
<rares> okay I'm rambling I'll go sit in a corner
<XeF4> rares: euhhh?
<ult> eh
<ult> i want a ball lightning generator
<rares> Xef4 it could happen
<water> ult: yeah so you can get yourself killed playing with it
<ult> heavily ionized air...massive potential differences inside the air...mmm
<ult> water: Thats it. Your house goes first.
<water> ult: i was stating that you wouldn't be able to control it :)
<ult> water: Duh. I'd leave it on your porch, ring the doorbell, run like hell =P
<rares> water we can contain antimatter can't we?
<ult> rares of course
<ult> we can contain anything basically
<rares> it's simple
<ult> very
<water> not indefinitely
<ult> though it requires incredible amounts of continous power
<rares> you don't contain it
<rares> you just don't arm it until the right moment
<rares> no conmtinuous power needed
<water> hm i think this goes down in the #tunes record book as the most useless conversation ever
<ult> yep
<rares> water the most fun so far
<water> well have it somewhere else please
<rares> so what do you want to talk about
-:- SignOff dalvarez: #TUNES (enough of computers for today)
<water> it isn't about that
<water> it's about what belongs on #tunes and what belongs on, say, #doplhin, #osdev, or #{}
<tcn> is #{} for real?
<tcn> guess so :)
<rares> well what specific things do you want to resolve?
10:30am
<eihrul> yes, Dolphin is the house of noise... inquire within
<water> in tunes/arrow/slate?
<eihrul> er #dolphin
<eihrul> :)
<rares> slate
<water> well i'm reading a bunch of papers right now, trolling for ideas
<rares> odd we could use brain implants about now but if we had them no one would use them
<water> i'm working towards completely specifying the default mo
<water> rares: who f***ing cares, then?
<water> it's the same issue with e-books and stephen king
<eihrul> how close are you to specifying? :)
<water> i'm working on it... there are a few questions left to answer
<rares> like?
<water> like how open the lookup and apply mechanisms should be at first glance
<eihrul> 'open'?
<water> yes at what level does the Top protocol allow you to fiddle with those mechanisms
<rares> like who gets to change and read objects
<rares> you could do it from a loosely defined purpose spec
<water> well, it may not be Top directly, i may make Top very primitive and unusable(?) and roll user-friendly stuff over it
10:40am
<water> especially now that i'm considering composing mo's
<rares> a short bit of XML that says what you want it to do
<water> why should i use xml?
<rares> sorry habit (xml has the potential of making things even more incompatible)
<tcn> let's go back to TXT
<rares> tcn amen to that
<water> forget this. back t othe point
<XeF4> tcn: or maybe unicode TXT
<rares> probably the uf layer is a good idea
<water> probably less noise would be a better idea
<eihrul> there's a difference between writing a spec and deciding what to write the spec in :)
<water> indeed
<rares> okay that one was not my fault :/ (but anyway... end of thread)
<water> s/not//
-:- SignOff XeF4: #TUNES (reboot)
<Fufie> is everybody a bit edgy or what?
<water> so the question is how would someone like to use the mo
<rares> hmm
<rares> probably call and go (gimme this behaviour on the double)
<water> we can't just say "make mutable" or "make immutable" without some refining
<water> rares: how much do you grok the slate docs?
<rares> pretty well
<water> well the bmo's already are supposed to act like that... that's the point of a dynamic language
<rares> right
<rares> I was thinking from the users perspective
<rares> now I'm thinking
<rares> You need someway to allocate behaviours to an admin level
<water> admin level?
<rares> you don't want people changing core properties
<water> you mean representing users and super-users as objects and giving those objects rights?
<rares> something like that
<rares> actually
<water> damn it i've already decided on that
<rares> no
10:50am
<water> i'm talking about how to get a really flexible framework based on some primitive ideas
* Fufie/#tunes whispers "ze curse of scheme.."
<water> i'm not building a feature list, i'm building a framework based on mo features
<rares> so not security? sorry...  lemme see
<water> grrrr
<water> no, security is implemented over the basic set of features
<rares> so you want the proverbial air, earth, fire, and water of the whole system? ...
<water> why do you have to throw in obscure metaphors?
<rares> habit
<water> i can't even comment on that
<tcn> water, where did you say what 'mo' is?
<rares> meta object
<water> yes it implements the lookup of messages and application of the messages
<rares> gah it's on the tip of my tongue
11:00am
<tcn> do you really need Mobius to do Slate?
<rares> you do
<water> it's just the self-hosted compiler
<water> which tunes definitely needs
<tcn> so Mobius takes syntax trees and compiles them, that's it?
<water> yep that's the current .plan
<tcn> and it can do a partial recompilation
<water> yes
<water> hopefully it'll be a rather general framework
<tcn> so Mobius could be as simple as converting a syntax tree to token-threaded code?
<water> sure
<tcn> ok
<tcn> Did you have a specific syntax in mind for Slate?
<water> um YES
<rares> read the page
<water> :)
<water> hell, read the posts to the tunes mlist
<water> and yes there's a specific syntax page
<tcn> ok lemme find it :)
<water> it's under "details"
<tcn> looking good
<water> yes its extremely simple
<rares> well so far you have the copying, info, and parenting primitives... what else is there? findQuickest_execution_route?
<water> heh
<water> actually parenting may not be primitive
11:10am
<water> most of the literal primitives should be there, but i'm rolling an abstraction system over it
<rares> you're right it isn't needed internally as much as the others
-:- SignOff ult: #TUNES (Leaving)
<water> it mostly is there to allow inspection of the delegation chain
<water> there's also object history which can be inspected
<rares> that could be useful for automatic learning
<water> which in the case of 'library' objects will include the process of taking root and incrementally adding all the features to it
<water> automatic learning? as in?
<rares> now what you need is a system to let hierarchially challenged people use the objects
<rares> automatic learning as in tunes could observe itself
<water> well i might not leave it as a hierarchy
<water> oh ok
<water> i've been considering the creation of other root objects
<water> (which may require mobius in the end)
<water> btw the hierarchy thing for namespaces is a bit of a facade, as the syntax page explains
<rares> like it could recognize operations and techniques used in other situations (and save as a stream to be applied unchanged to a diff context)
<water> hierarchy for inheritance just follows from the fact that everything is a modification of root
<tcn> you decided it'll be a graph, not a tree?
<water> yes and you can use macro-like expressions to share object histories
<water> tcn: it has to be in order to work
<tcn> 'k
<water> but i can make it look like a hierarchy for purposes of metaphor with file systems
<rares> right
<water> hm i should explain this better in syntax
<water> bleh, the namespace syntax hasn't been updated properly
<water> well the concept there is correct anyway
<tcn> what's different?
11:20am
-:- ult [ult@user-38lcn51.dialup.mindspring.com] has joined #Tunes
<water> not sure yet
<water> basically the syntax is non-compliant :)
<rares> to whateth?
<tcn>  so will Slate look like LISP with namespaces?
<water> "/ (.. foo)"
<water> yeah basically
<water> but the objects aren't lists, they're sets
<water> btw, adding slots is primitive
<rares> kinda hastobe
11:30am
-:- Tril [dem@bespin.dhs.org] has joined #TUNES
-:- mode/#tunes [+o Tril] by ChanServ
<Fufie> tcn: umm.. lisp has namespaces
<water> hey tril
<Tril> hi...
<Fufie> hi tril
<water> fufie: yeah but not every object in lisp is a namespace
<Fufie> water: no.. not every object
<water> and i don't count packages as proper namespaces
<Fufie> symbols are namespaces as well
<water> how so?
<Fufie> not true namespaces though.. but they're small containers of slots
<Fufie> you can add anything to a symbol
<water> what does that mean?
<Fufie> assume you have a symbol (or object if you want) 'foo
<Fufie> then you can add 'bar to it 
<water> oh wow :P
<Fufie> (setf (get 'foo 'bar) "some value")
<Fufie> and fetch it with (get 'foo 'bar)
<water> ok that's enough i don't care any more about lisp
<water> tril: what's up today?
* Fufie/#tunes smiles.. 
11:40am
<Fufie> it's kindof selfish when you look at it because the symbols nest nicely :)
<water> yeah whatever
-:- SignOff Fufie: #TUNES (Ping timeout for Fufie[tunnel-44-29.vpn.uib.no])
<Tril> reading today's log
<water> okay
11:50am
<Fare> Gakuk
<water> oh great
<water> hey fare, multi-methods in slate are simple
<water> (or did i already explain this to you?)
* Tril/#TUNES is away: (Auto-Away after 10 mins) [BX-MsgLog Off]
<water> bah
-:- SignOff Lothos_: #TUNES (Leaving)
* Tril/#TUNES is back from the dead. Gone 0 hrs 1 min 24 secs
<water> yes?
<Tril> ok, I am going to try to concentrate on this Zope thing today
<water> ok well as always i'm discussing language issues
12:00pm
-:- Downix [down@d-gnaps-4.ici.net] has joined #tunes
<rares> hey Downix
<hcf> abi: esoring is the Ring of Esoteric Programming Languages at http://lightning.prohosting.com/~kgaughan/esolang.html
<water> um it only references intercal
* Tril/#TUNES is away: (Auto-Away after 10 mins) [BX-MsgLog Off]
* Tril/#TUNES is back from the dead. Gone 0 hrs 0 min 6 secs
<Tril> argh
-:- tcn_ [r@cci-209150250148.clarityconnect.net] has joined #tunes
12:10pm
<Fare> sorry, 61453
-:- SignOff tcn: #TUNES (Ping timeout for tcn[cci-209150250174.clarityconnect.net])
<Tril> Fare: 61453 before, or now?
<water> as if it matters
<hcf> water: anything to not deal w/ the real issues
<water> no doubt
* Tril/#TUNES inputs his review database structure into postgres (overwriting lar1's version)
-:- tcn_ is now known as tcn
12:20pm
<water> i wonder if i can reconcile my method code's data-flow by slot model with the model of objects as method-hostories
<water> s/host/hist/
<rares> hostory inspires a solution
<water> explain
12:30pm
<rares> you have the history of stuff flowing through slots
<water> the slots of a host namespace?
<water> hmm
<water> of course this implies that most naturally one would show all objects' histories in such a 'host'
<rares> and you have the history of flows begun in alltimespace
<water> hm
<rares> you have two dimensions of history
<rares> that's all
<water> one problem may be that this namespace cross-cuts the usual namespace system
<rares> that's why i'm saying two dimensions
<water> not impossible, but it would have to be introduced to the user properly
<water> slate seems to have lots of dimensions :)
<rares> otherwise it would be a confusing crosslegged stream of history with a difficult task of untagling
<rares> untangling even
<water> right
* Fare/#Tunes is back
* water/#tunes reads a paper on a unified theory of inheritance mechanisms in major oo languages
<Fare> seen type inference algorithm for OO languages?
<water> of course
<Fare> well, if a computer can't infer the type, then the human cannot either.
<rares> i'm even beginning to doubt you need hierarchies
<water> rares: why?
<rares> well you have a task right?
<water> a "task"?
<Fare> I only believe in type frameworks simple instances of which have simple type inference algorithms
<rares> task = something to accomplish
<water> Fare: and i don't believe in France, but it exists just the same
<Fare> if the object systems are so horrid that no algorithm can infer the types, then the system is rubbish
<Fare> C++ object system is rubbish
<water> preaching to the choir, Fare
<rares> no kidding
<Fare> it's not a matter of opinion. It's a matter of arguments.
<water> *ahem* what's your point?
<rares> yeah but we're already convinced
<water> i'm addressing this in slate
<rares> now
12:40pm
<rares> tasks, tools, and behavior database
<water> rares: you were saying?
<rares> let the user mumble the thing he needs
<water> okay this reminds me of bottom-up programming
<rares> let the system connect the behaviours into different objects
<rares> the hierarchies should be stored models
<water> er
<water> what does this mean at the language level?
<rares> two things
<Fare> water: what about your slate multi-methods?
<Fare> how do they compare to CLOS, CECIL, Dylan, foo?
<water> Fare: they're just objects that act like methods on their fields and don't update their enclosing namespace's state
<rares> one: for automating the construction of objects you need heirarchies but that would be an internal language
<rares> two: the user just calls up things he needs done
<water> Fare: since all my objects have 'result' fields and are fully able to act as methods, the use of multi-methods is trivial
<rares> now there's a cool thing about this
<water> rares: you're arguing about UI it seems
<rares> partly
<water> the cool thing is... ?
<rares> one of the problems with C++ is that I have to connect objects in one and only one way the way the API specifies
<water> uh
<rares> cool thing: no m,ore 600 page app manuals
<water> :P
<water> i already know that
<water> try squeak sometime
<rares> i know
<rares> what i'm saying is you need a resource selection language
<water> yes but unless you can prove to me that such a thing is possible, i don't care
<water> and i don't see how it affects the slate language design
<rares> tell you what i'll work on it this week and report back to you:)
<water> ok
<rares> I have a few ideas that could kick ass
<rares> and the tech is already there
<water> er
<Fare> what's this "updating enclosing namespace state" thing?
<Fare> why should methods do that?
<water> heh
<water> because my objects are namespaces and vice versa
12:50pm
<Fare> sure, but why would method invocation modify namespaces?
<Fare> that seems wierd.
<water> because that's what it does in ordinary objects
<Fare> uh?
<water> otherwise there's no use in having methods
<Fare> objects that do that are EVIL (including C++ objects)
<water> sigh
<Fare> method dispatch is orthogonal to object state.
<water> an object's state consists of the returns of its accessor methods
<Fare> mingling the two is EVIL.
<water> it's illusory
<water> self even works this way
<water> besides slate has other language features that make the comparison to c++ or whatever you're thinking of moot
-:- SignOff ult: #TUNES (Ping timeout for ult[user-38lcn51.dialup.mindspring.com])
<water> and using the word "evil" only exposes an irrational bias
<Tril> Fare does have an irrational bias to anything with "message passing"... he flamed me out of using that in my first OS project.
<water> what os project was that?
<rares> I have an irrational bias to bloated scafforlding infrastructure ok scaffolding relationships conditionals get in the way of refglection
<Tril> it's what I was doing before I ran into TUNES
<Tril> tunes has a lot better ideas that made what I was doing seem silly
<water> oh
01:00pm
<water> btw tril i did an interview about tunes recently
<rares> w/ who?
<water> http://www.linux.com/interviews/
<Tril> job interview? news?
<water> open-source general interest
-:- hcf has changed the topic on channel #tunes to: Tunes http://www.tunes.org || Slate Language http://www.tunes.org/~water/slate-home.html || interview w/ water http://www.linux.com/interviews/
<Tril> how did you get to do this interview?
<eihrul> someone just popped into the channel...
<water> a newbie came into the channel with questions about tunes
<water> he thought it was interesting enough, even without code, to do an interview
<hcf> Tril: irc/2000.0321
<Tril> I think the interview should be posted on main tunes page, no?
<water> well he talked to me privately as well
<water> perhaps
<hcf> Tril: a news item
<water> i don't care much either way... i'd rather see code there
<Tril> hcf: the main page = the news page
<hcf> Tril: it should be, news page = news page
<Fare> message passing is great to simulation communication
<Fare> s/simulation/simulate/
<Fare> but to express simple transformations, it's bad.
<water> you can't build algebraic datatypes upon message-passing?
<Fare> it's too low-level
<rares> hcf yeah but it's silly to see a what's new link on the main page just tell me what;s new
<water> fare: get your head out of the ground
<Fare> water: sure, you can build anything on top of anything.
<hcf> i retract my last comment
<water> message-passing as a semantics feature does not have to be low level
<Fare> a purely message-passing calculus is the pi-calculus (or a better one, the join-calculus which happens to be isomorphic to good old actor languages)
<water> not necessarily
<water> you think message-passing is strictly in the time domain
<water> i disagree
<Fare> what does message passing "in" or "out" the time domain mean?
<water> you illustrate my point
<water> message-passing does not have to occur in real-time and situate events in real time
<water> and as i've already pointed out, message-passing in slate is illusory
01:10pm
* water/#tunes returns to reading
-:- FareWell [rideaufr@esmeralda.enst.fr] has joined #Tunes
-:- FareWell [rideaufr@esmeralda.enst.fr] has left #Tunes []
<water> hm where were we?
<rares> good question
<water> hm so are there any comments on the interview?
<eihrul> water: you were on method code data-flow :)
<water> oh yeah thx
<Tril> not really  .. it was fine. did it get newbies to start harassing us more often than usual?
<water> not that i've noticed
<water> and it's only been up for a few days yet
<water> if we have a namespace entry specifically for showing object histories, it should probably occupy a place uner whatever 'system' namespace we have
<water> which reminds me that i don't have a satisfactory idea of how to organize the major namespaces
<water> so far, my only ideas are to unify squeak or self's ideas with those of unix
<Fare> hey, my fav comic book series is online! jp-petit.com
<water> s/unix/posix/
<Fare> posix? yuck
<water> whatever
<water> the general concept used in the usual file system of the same name
01:30pm
<water> Fare: how do you propose to organize a Tunes system?
<water> Tril: see just how useful Fare is?
<Tril> organize what?
<water> tunes objects
<eihrul> water: so what is the dilemma with method data-flow right now?
<water> i wonder why i put up with this some days
<Tril> well, it's sort of a vague question, water
<water> not much, just finding a way to introduce it into the namespace system
<Tril> first of all, I don't know what all tunes objects there will be, so how can I know how to organize them?
<water> tril: well i can't aska more specific question until i get answers
<eihrul> water: could you elaborate on 'introduce into namespace system' :)
<water> eih: where does it fit since it crosscuts other namespaces?
<eihrul> rmm, how so?
<Tril> Secondly, I dont think "organize" is even applicable in such a system.  
<water> tril: well have you looked at smalltalk or self's organization?
<water> it definitely does apply
<water> supose you have millions of objects and you want to learn how the system works
<Tril> Well, each system might be organized differently. Since each user can change it..
<water> no kidding
<water> but it still helps to have a way to conceptualize the system
<Tril> and there would be multiple paths to every object, I would hope so that all objects were "nearby" objects theat are trelated in some way.
<eihrul> water: so where does it cross-cut?
<Tril> I have not looked at SELF because I can't frigging run it, thank you very much.   Smalltalk, well, is organized hierarchically I believe
<water> eih: well object history cross namespaces
<water> self runs on macs
<water> and you can get the programmer's reference on line
01:40pm
<water> well smalltalk demonstrates some concepts wrt organization
<eihrul> eh, i thought the object history was just consumed elements of a stream kept on the object :)
<water> eih: it is but in general its a graph of expressions evaluated to create the desired set of objects
<Tril> OK, I will clarify what I said about organization not applying
<eihrul> water: and the problem is that these expressions may not mean the same thing in different environments?
<Tril> a particular "organization" is not forced on teh system by us as designers. Multiple organizations can simultaneously apply. Therefore, we don't have to "Decide" on "the" organization per se.
<eihrul> water: er namespaces
<water> Tril: besides if you intend on having any sort of scoping in the tunes language then you already have implemented an organization method!
<Tril> We can choose multiple organizations, or none at all, and let someone else, perhaps a user, decide
<water> eih: among other things, yes
<water> tril: preaching to the choir
<Tril> given that, I still think that I don't know what I'm organizing
<water> abi pttc is Preaching to the Choir
<eihrul> water: well, i'm thinking closures here... the namespaces must be encapsulated in the expressions themselves...
<water> eih: k
<water> tril: given what?
<Tril> given that you just basically said you already know everything I just said
* water/#tunes sighes
<Tril> but you still insitst on asking the question, my only answer is, WHAT am I supposed to organize
<water> well consider that i'd like to have different kinds of numbering systems
<Tril> to organize arbitrary something, I like to pick orthogonal categories and stick everyhitng into every category that applies. OK?
<water> grrrrr
<water> my point is that we need a meta-system for that
<eihrul> or indirection :)
<Tril> that = multiple numer systems, or organizing things in categories?
<water> even Tril is plagued by vagueness when discussing Tunes :(
<water> both
<eihrul> you get out what you put in, water...
<eihrul> the question was vague :)
<water> eih: i don't have anything else to put in!!!
<Tril> why?
<eihrul> because slate is young
<water> because i'm trying to construct the subject matter by question-and-answer
<water> which is called the dialectic method iirc
<water> 2000 years old!
<rares> Marx used it so beware
<water> marx is a person
<Tril> water: we need an organization system that scales
<water> absolutely
<water> what about my number system example?
01:50pm
<Tril> do I know anything about multiple number systems?
<water> how do you help the user notice where the existing code is that handles this?
<water> huh?
<eihrul> grep...
<water> eih: yeah but we don't always use ascii in tunes
<eihrul> water: as applied to objects :)
<water> eih: in fact, rarely so
<water> grep doesn't scale
<rares> that's why I suggested building objects at some time before running let behaviours be in the main  database until then
<eihrul> well, yes... searching the whole tree of objects would be quite slow and complicated
<Tril> I make all dependencies explicit. If it's using a particular "number system" (whatever) then there is a function somewhere that has an argument of "which number system" and it has been partially evaluated on a particular one.  Using reflection the user can see which argfument the function was Pe'ed on and thus get a reference to the number system in question. 
<water> bah you're missing the point
<Tril> all configuration options are specified by incorporating them into the generic function framework
<water> i have a set of digits somehwhere which are used variously by different number systems
<water> it's this ambiguity i'm trying to address
<water> and partial-eval of functions has limits when used that way
<Tril> if the point revolves around some details of the example multiple number systems, I don't know what multiple number systems are, please clarify
<water> witness the limitations of intuitionist logic
<Tril> Do you mean like base-10, base-2 base-16?
<water> not just that
<Tril> Roman?
<water> number systems where iteration produces 0,1,2,3,0,1,2,3,...
<water> or number systems of general group theory
<water> or distinguishing rationals from integers from naturals
<water> or the algebra of limits
<Tril> this may be off topic, but do you have any idea how to implement the real number system?
<water> yes and there are papers on it
<Tril> most languages have it, but at a really low level so I'm not sure they need to deal with the issues that we would?
<water> there are papers on high-level real numbers
<Tril> where?
<water> hold on
<Tril> and where are papers on that other-type of reflection you said you were using in Slate?
<water> (i have about a 1000 research papers by now, it's hard to track down urls')
<water> check researchindex
<water> abi ri?
<abi> ri is, like, researchindex at http://citeseer.nj.nec.com/cs?dd=2
<hcf> and cora
<water> http://citeseer.nj.nec.com/dony-semantics.html
<water> note the citations as well
<Tril> that's not about real numbers
<Tril> it's about organization
<water> i know i'm looking through the 68 returned references for a particular one
02:00pm
<water> i do a ton of trolling through research archives, so Back Off!
<Tril> so, what percentage of papers written are never published on the internet due to scientific journals taking exclusive publishing authority
<water> i don't know and don't care
<Tril> are there good repositories of research papers that you can read the entire text of?
* Downix/#tunes  shakes his head, no
<water> yes, ri and cora
<Downix> other than at universities
<eihrul> water: ri?
<Tril> scroll up dude
<water> eih: researchindex
<eihrul> oh, ok
<hcf> abi: tell eihrul about ri
<eihrul> more paper indexes are always good
<water> well unless you guys have some useful feedback on my questions, i'm going to be leaving in a bit
-:- Ghyll [karltk@mp-217-229-10.daxnet.no] has joined #tunes
<water> Tril: if you want papers on real numbers, search ri and cora for them and pick the most intuitive returns
<water> otherwise, perhaps hcf would assist you
<Tril> so what is the question
<Tril> if I say "0" what do I mean?
<water> sure
<water> or how do you mechanically determine what the user means when he says "0" in a way that supports tunes
<water> and if you say pe i'm leaving right now
<water> because you need more than pe
<eihrul> well, isn't that issue resolved by the compiler + evaluator?
<water> i have taken the scoping idea and extended it as much as i can in the slate language
<water> no kidding
<Tril> I dont really like type inference. So I'd have my tunes configured to make me tell the system which 0 I meant every time. Until it got annoying.
<water> it will get very annoying i assure you
<Tril>  I also don't like LISP's dynamic typing
<water> besides, i don't mean type inference
02:10pm
<Fare> I don't like LISP's _forced_ dynamic typing
<water> fare: how would you suggest to organize tunes objects?
<water> (i ask yet again)
<Tril> you're already assuming a layer of UI above the pure semantics of the system. Like it has to interpret what I am saying or something.
<water> no i'm not
<Fare> object spaces, where each object is described by name, type, intentions, etc
<water> sigh
<water> thanks, nice and vague
<Fare> then you can query a space for contained objects that satisfy some property
<Tril> If I'm saying "0" to the system, that implies I typed to the system the ambiguous character 0, instead of selecting the number out of a particular number system.
<water> no it doesn't
<Fare> the most basic userland spaces would be purely namespaces
<water> you don't have to do it that way
<Tril> Why can't I just write "binary 0"  or  binary(  0 1 0 11 0 0) 
-:- SignOff hcf: #TUNES (Leaving)
<Tril> or "binary mode on" and then make it implicit afterward
<Fare> more advanced ones would be associated to AI search mechanisms
<water> because that's very limiting
-:- hcf [nef@me-portland-us847.javanet.com] has joined #tunes
<water> bah, mmore hacks
<Tril> water: do you know Emacs? I don't, but from what I can tell it has modes for everything.  Can't users just select what mode the interpreter is in?
<water> grrrrrrrrrrrrrrrrrrrr
<water> what rock do you guys hide under?
<water> not everyone using tunes will be a programmer
<water> the subject matter of all number systems in tunes will not necessarily be computations per se
<Tril> Your question is asking what is the general solution to finding the semantics of a token. Hasn't someone been working on context calculi (if such thing exists) and so on to help us?
<water> that's Low Leve
* Downix/#tunes nods to water
<water> Level
<water> get a clue, tril
<water> context calculi are limiting
<Tril> water: I dont know that. I've never seen one.
<water> they're based on intuititionistic logic
<water> i have definitely seen one and worked through the ideas
<Tril> which one is that
<Tril> I dont know much logic, either!
<water> gah
<Tril> Fare seems to like that one right?
<Tril> Yes, you can be frustrated at me
<water> yes he likes it quite a bit more than its worth
<water> i just got a book on substructural logics that explores lots of logical ideas
<water> it extends what i was looking at in "exploring logical dynamics" and "arrow logics and multi-modal logics"
<Tril> what book
<water> hold on
<water> it got published just recently
<water> btw intuitionistic logic has been around since the '30's
<water> damn it my cookies keep disappearing
02:20pm
<Tril> COOKIE?
<water> yeah my web browser cookies
<Tril> I make my cookies disappear on purpose
<water> well lets not talk politics
<Tril> did you +R cookies.txt and then forget you had?
<water> no
<hcf> http://www.phil.mq.edu.au/isl/
<water> yeah that's the one
<hcf> it was in one of the review@ posts i did
<hcf> i guess Tril didnt pay attention
<water> you should take it as an assumption that tunes members don't pay attention
<Tril> right
<Downix> ok, see you all later, time to snooze
* Downix/#tunes waves
-:- SignOff Downix: #TUNES (BitchX: now in non-drowsy formula too!)
<Tril> Anything addressed to review@ is for archive purposes, and for core to add to the page.
<hcf> water: i know, thats one reason i stopped posting
<Tril> I wish I could read every link, but, how can you expect that?
<water> well i read everything
<Tril> like I said, how can you expect me to aspire to Water's greatness
<water> bah
<water> i'm obsessed
-:- SignOff hcf: #TUNES (Excess Flood)
-:- hcf [nef@me-portland-us847.javanet.com] has joined #tunes
<Tril> anyway, I was referring to an abstract notion of "context calculus", not any particular one you saw that was unacceptable. I'm assuming something suitable could be invented to meet the need of whatever you propose
02:30pm
<water> yeah well context calculus is a real thing that someone actually looked into
<water> it's based on block-structured functionality
<Tril> I'm not trying to put off your question. I agree that the context thing is not clear enough in tunes.
<water> which is the status quo of programming language ideas
<water> well what can be done with my namespace idea in slate?
<water> i actually plan to have namespaces with the numbers in them, btw
<water> that's one of the points about my objects being sets instead of lists
<Tril> I haven't lookde at your name space idea in slate yet
<water> well my namespaces are just objects and vice versa
<water> check the semantics and syntax pages
<Tril> what's a name? A string, or an absract "unique value"?
<water> that's a good question
<water> i'd like it to be the second one
<water> in general it's anything unique up to isomorphism as the tunes spec so aptly puts it
<water> :)
<water> but beyond that i'm still working out the ideas
<water> for instance, naming is an interface notion
<water> also, internationalizing the language's names may be very difficult, as jecel found with smalltalk
<water> what do you think?
<Tril> well, the "set of abstract unique values" is very close to my notion of type
<water> i'm sure it is
<water> but the subject matter is slate
<Tril> well, maybe you will want to discuss it later with me then
02:40pm
<water> oh of course since our discussion so far has been so fruitful
<Tril> right. because I'm doing other stuff, and haven't read the slate info completely!
<water> i suppose you're going away for a bit, then?
<water> alright then
<Tril> I should leave IRC because of that I guess
<water> i'll go out on the town
<Tril> ok
<water> all: cya later tonight
-:- water [water@tnt-9-186.tscnet.net] has left #tunes []
<Tril> I'll try to read slate stuff soon.
-:- SignOff Tril: #TUNES (Tril has no reason)
02:50pm
-:- SignOff tcn: #TUNES (ircII EPIC4pre2   cLIeNUX. Can you say that?)
-:- SignOff hcf: #TUNES (Leaving)
-:- XeF4 [xef4@194.19.185.227] has joined #tunes
-:- SignOff rares: #TUNES (Read error to rares[wtrb-sh1-port65.snet.net]: Connection reset by peer)
-:- SignOff Ghyll: #TUNES (Ping timeout for Ghyll[mp-217-229-10.daxnet.no])
-:- SignOff eihrul: #TUNES (Ping timeout for eihrul[usr5-ppp96.lvdi.net])
-:- eihrul [lee@usr5-ppp79.lvdi.net] has joined #tunes
-:- ult [ult@user-38lcn5t.dialup.mindspring.com] has joined #Tunes
-:- hcf [nef@me-portland-us819.javanet.com] has joined #tunes
-:- Kaufmann [newbie@200.224.105.144] has joined #tunes
<Kaufmann> Foo
<eihrul> quux
<eihrul> (quux! 'kauffie)
<Kaufmann> urgh
* Kaufmann/#tunes is mutated once again
<Kaufmann> (note to self: avoid Lee in order to preserve state)
06:40pm
<eihrul> well, in an unpure world, purely immutable beings are going to have a problem...
<Kaufmann> yeah
<Kaufmann> that's why God is dead
<Kaufmann> that whole "eternal" stuff couldn't stand up against evolution
<eihrul> good point
<Kaufmann> "Omigod!" "What?" "Look, Tom! Isn't that..." "Yes, that's right, Tiffany! That's Brian Rice of Slate fame!!!" "Omigod, I think I'm gonna faint..."
<eihrul> huh?
<Kaufmann> (re. interview w/ Brian on Linux.com)
<eihrul> uh... huh, whatever :)
<Kaufmann> wouldn't it be great if one of those wannabe-"with it" Linux-heavy venture capitalists read the interview and decided to hire us all to work on his new startup company, eTunesent?
<Kaufmann> LOL
<eihrul> ugh
<Kaufmann> knowing Brian, he'll keep me, Fufie and the rest out because we're "noisy"
<eihrul> "Uh, sorry... Brian gave us orders not to let noisy people. eTunesent police, it's our job."
<eihrul> and with the navy connections he has :)
<eihrul> s/people/people in
<Kaufmann> heh
<Kaufmann> he'll hire a bunch of his Navy buddies to run the company, and organise it like his private army
06:50pm
<Kaufmann> "Gen. Rice, CEO"
<Kaufmann> spooky
<eihrul> then he'll IPO...
<Kaufmann> oh man
<Kaufmann> I forgot that part
<Kaufmann> the name alone will earn a 5000% rise
<eihrul> then Microsoft will buy out the company...
<Kaufmann> too late, they've already gone public
<Kaufmann> can you buy out a public company?
-:- mathias [math@france2.hemmet.chalmers.se] has left #tunes []
<Kaufmann> although I must admit, MS Arrow 2000 _would_ be an interesting sight
<eihrul> well, you can buy all the stock :)
<Kaufmann> You have performed an illegal operation: Model-level reflection. Please crosscut ontologies or contact Microsoft Tech Support at www.microsoft.com/arrow.
* eihrul/#tunes rofls.
07:00pm
<eihrul> "You have reflected upon your mouse's position. Please reify the system for changes to take effect." :)
<Kaufmann> egads
* Kaufmann/#tunes looks around to double-check that water isn't here
<Kaufmann> last time we joked around about that company from whom he received an email beating him to it and releasing "eArrow 1.0 for Linux", he got pissed off
<eihrul> doh
07:10pm
<Kaufmann> "with XML support, even"
<Kaufmann> *chuckle*
<Kaufmann> I'm submitting it to Slashdot
<Kaufmann> just to increase our chances with aforementioned VC
<eihrul> man, don't abuse your karma...
<Kaufmann> heh
07:20pm
<Kaufmann> but on an entirely different topic, have you ever used ETH Oberon?
<eihrul> nope
<Kaufmann> urhm
<Kaufmann> oke
<Kaufmann> i
-:- SignOff Kaufmann: #TUNES (A stranger is just an asshole you haven't met yet. -- Meet the Fascist Moderator at www.osopinion.com)
07:30pm
-:- SignOff XeF4: #TUNES (Ping timeout for XeF4[194.19.185.227])
-:- SignOff hcf: #TUNES (Leaving)
-:- hcf [nef@me-portland-us728.javanet.com] has joined #tunes
-:- water [water@tnt-9-95.tscnet.net] has joined #tunes
<water> hey all
10:50pm
-:- SignOff ult: #TUNES (Ping timeout for ult[user-38lcn5t.dialup.mindspring.com])
-:- ult [ult@user-37kba0k.dialup.mindspring.com] has joined #Tunes
-:- TFox [tfox@dialupW227.sttl.uswest.net] has joined #tunes
<TFox> hey
<water> hey
<water> are you familiar with tunes?
<TFox> Draugath says you're working on an operating system, that's as much as I know
<water> heh
<water> yes it's a kind of operating system
<TFox> explain?
<water> have you read the faq or looked aroung the website?
<water> i can answer lots of questions btw
<TFox> I browsed the website a little, but I didn't get much of a feel for it
<water> hm
<TFox> it's a GUI, right?
<water> check the interview linked to in the channel topic
<water> um no
<TFox> no?
<water> no but it will include a gui
<water> it's not based on low-level languages
<TFox> right... what OS doesn't right? well, except MAC OS
<water> sorry the difference with tunes is not what you think it is
<TFox> ?
<water> what languages have you programmed in?
<TFox> none
<TFox> I'm not a programmer
<water> well then much of the tunes argument would be lost on you without much explanation
<water> but i am patient :)
<TFox> I'm a service tech/consultent
<water> in most operating systems, the end user receives a black box, right?
<water> heh i did that in high school
11:50pm
<TFox> I'm familiar with operating systems and most program structures, I just don't know the code for programming
<water> well with linux for example, you can recompile your kernel, right?
<TFox> a command line text. yes
<TFox> right
<water> that's ok, i'm explaining
<water> command line??
<TFox> command shit
<water> the command line is irrelevant
<TFox> command line interface?
<TFox> yeah, anyway
<water> anyway, the c language shapes how you can customize your kernel and other services
<TFox> ok
<water> most oses are coded at a low level like c or assembly, so the services they provide are low-level, too
<water> furthermore, the nature of languages like these is that the end-user can
<water> oops
<TFox> I apologize, I'm not really interested in the programming of it, more things like stability, compatibility, and interface
<water> can't modify the kernel functionality to his or her own needs
<water> well the programming of it definitely affects those things
<TFox> yes, but if you could summarize it all for me.
<water> for instance, tunes is designed to be a programming language
<water> like i said, refer to the interview link in the topic
<TFox> I'm not familiar with the programming language's relevance
<water> well then you won't grok tunes
<water> have you heard of squeak?
<TFox> no
<water> abi: tell tfox about squeak
<water> squeak is a smalltalk-based os that gets hosted on other oses or bare hardware
<water> everything in the os is an object that can be understood and explored by end-users
<TFox> my interest for a new operating system at this point is stability, graphic quality (3d would be nice
<water> this is from the Xerox PARC labs of the '70s
<water> although squeak handles all the things you mention
<water> tunes will be like squeak in that respect
<TFox> and I have a bitch about the fact that No one can make a GUI indipendant of a base OS
<water> sigh
<TFox> indeed
<water> you obviously have no idea what i'm talking about then
<water> squeak runs identically on 12+ oses
<TFox> I'm waiting for a 3-dimentional voice recognitive OS, but that is far from reality
<water> and it hosts its own gui with 3d acceleration anti-aliasing, etc
<water> lol
<water> no it isn't
<water> the squeak team has small prototypes of those things already
<TFox> nice... 
<water> included in the standard distro
<TFox> I'd be all over that
<TFox> how can I check it out?
<water> one problem with squeak though is that optimizations are not complete, and a lot of code is experimental
<water> download the particular version for your os at http://www.squeak.org/
<water> it runs fairly fast, but the JIT compiler is not quite ready for production use
<TFox> I'm running win98 right now, with linux on a 400M
<water> and the experimental code is not nearly well-optimized yet
[msg(TUNES)] newlog 2000.0326
IRC log ended Sun Mar 26 00:00:01 2000