IRC log started Sat Aug 14 00:00:01 1999
[msg(TUNES)] permlog 1999.0814
-:- ruiner [nate@ppp137.wi.centuryinter.net] has joined #tunes
-:- SignOff ruiner: #TUNES (Leaving)
<liar> what code?
<eihrul> the threading code
<liar> what threading code?
<abi> threading code is not really finished yet, though...
<liar> forget threading code
<abi> liar: I forgot threading code
<eihrul> the code that implements threads in brix :)
<liar> all of it?
<eihrul> if you'd be so kind
<liar> like suspend/sleep/die/kill/fork/call/resume?
<eihrul> yessir
12:50am
<liar> let me think on that
<liar> remind me tomorrow night and its yers
<liar> i gotta sleep now
<eihrul> okay
<eihrul> 'night
<liar> cya
-:- SignOff liar: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep)
01:00am
-:- SignOff eihrul: #TUNES (Leaving)
-:- fare [fare@quatramaran.ens.fr] has joined #Tunes
-:- SignOff Mr_Wrong: #TUNES (Ping timeout for Mr_Wrong[adsl-63-192-213-146.dsl.snfc21.pacbell.net])
-:- core [core@core.suntech.fr] has joined #tunes
<core> people hi
<core> fare: t'es dans le coin?
<core> abi: fare?
<abi> somebody said fare was sometimes looking at another screen and interruptible with a beep (/msg Fare ^G) or at http://www.tunes.org/~fare/ or connected through a crappy ISP (cybercable.fr) or pronounced Fahree
05:10am
-:- Mr_Wrong [seanl@adsl-63-192-213-146.dsl.snfc21.pacbell.net] has joined #tunes
-:- SignOff fare: #TUNES (Connection reset by pear)
-:- SignOff core: #TUNES (thanks fare)
-:- water [water@ppp-tnt-233.tscnet.net] has joined #tunes
-:- ProGuy [Paul@p421-097.ppp.get2net.dk] has joined #tunes
<water> hey
<ProGuy> hey
<abi> hello, ProGuy
<water> anything on your mind?
<ProGuy> Not right now... Just looked in to see if something interesting was happening
<water> if fare were here, i'm sure that there would be
<ProGuy> I still haven't read the tunes or arrows description... but I will soon do
06:00am
-:- Miellaby [sgarden@St-Brieuc-4-142.club-internet.fr] has joined #TUNES
<water> hello
-:- Miel [sgarden@Rennes-Villejean-4-143.club-internet.fr] has joined #TUNES
<water> hello again
-:- SignOff Miellaby: #TUNES (Ping timeout for Miellaby[St-Brieuc-4-142.club-internet.fr])
-:- SignOff Miel: #TUNES (Leaving)
-:- Miellaby [sgarden@Rennes-Villejean-4-143.club-internet.fr] has joined #TUNES
<water> ?
<Miellaby> Sorry, lot's of time I didn't use IRC.
<water> k
<water> you from france?
<Miellaby> Ye,s I am
<water> do you know Fare?
<abi> Fare is sometimes looking at another screen and interruptible with a beep (/msg Fare ^G) or at http://www.tunes.org/~fare/ or connected through a crappy ISP (cybercable.fr) or pronounced Fahree
<water> abi: shut up
<Miellaby> I found home pages of TUNES project some days. I doon't know Fare personnaly, soryy.
<water> ok
<water> what brings you here? anything on your mind?
<Miellaby> Perhaps I could bring some ideas. I dont think I'm as much clever as you, unfortunnately. Maybe I could help for some low level design.
<water> that's fine
06:30am
<water> it took me years of reading and learning to understand certain things
<Miellaby> I think about a revolutionary computing system from more than one year ... I dont have a large global view of the problem as Fare seems to have.
<water> well, the global stuff just extends the philosophy of tunes beyond single machines
<water> i.e. trying to make the internet work better for people
<Miellaby> My pb is that TUNES members don't seem to like OO-stuff a lot. I'll have to learn to forget my dreams...
<water> pb?
<abi> i think pb is processbase
<water> abi forget pb
<abi> water: I forgot pb
<Miellaby> I'm not very fond of Lambda calculus neither. 
<water> oo-stuff tends to be c++ or java or some other lame version of object-management, where the programmer has to explicitly specify everything
<water> btw, the web site reflects Fare's opinions, not the group's as a whole
<Miellaby> water: have you an example of over-specified program ?
<water> well, sure. any program that explicitly handles low-level details when it doesn't need to is over-specified
* ProGuy/#tunes wonders what people have against C++
<water> hehe
<water> proguy: don't get me started :)
<Miellaby> Ok, but sometimes, we must help computers to organize their behaviours, their data, and so one. I don't think those things can be completly automated.
<water> no, not completely. but today, almost none of that is automated, and you can't change what is automated and what isn't
-:- ProGuy_DK [Paul@p432-054.ppp.get2net.dk] has joined #tunes
06:40am
<ProGuy_DK> water: Hmm...
<Miellaby> Each time I saw someone using High level functionnalities to design something, the result was very poor by quality.
-:- SignOff ProGuy: #TUNES (Ping timeout for ProGuy[p421-097.ppp.get2net.dk])
-:- ProGuy_DK is now known as ProGuy
<water> miel: what tools were they using?
<Miellaby> - Fat... redondant ... full of dead part of code ... I think about Java.
* ProGuy/#tunes has done AHLL in C++....
<water> blah!
<ProGuy> I like classes, but I guess it is possible to do things with structs instead
<water> you guys should look at some real languages sometime
<water> java is a bloated design from the start
<water> of course, c++ is as well, only more low-level
<ProGuy> Java is crap... 
<Miellaby> One day, I worked on a JAVA application comptletly designed through Collections (even If data were 1000 unindexed unitary-bits).
<ProGuy> I can see the problem with C+
<ProGuy> +
<Miellaby> One minute to launch on a recent pentium...
<water> that's due to the jvm, though
<Miellaby> Not only.
<water> although java definitely does not allow specifying hardware optimizations
<water> what's the point of this argument, though?
<ProGuy> Doesn't some cell-phone company use java in their cell-phones??
<water> nokia, i think
<Miellaby> Designer must be sensibilised to complexity of each hidden algorithm there is behind a function or a method/feature or whatever you want.
<ProGuy> That isn't the brightest thing to do...
<water> what?
<ProGuy> use Java in cell-phones...
<water> miel: what? how would that be accomplished?
<water> miel: why not just make the language better?
<Miellaby> I'm very prudent when I must argue about HLLs.
<water> miel: what are you talking about?
<Miellaby> Every HLL I know has contributed to the degradation of the efficiency of programs.
<Miellaby> (I hope my english isn't too poor)
<ProGuy> I generally don't like HLL's.
<water> then you don't know of a good HLL
<ProGuy> Miellaby: Exactly
06:50am
<water> you guys hold these opinions based on a lack of experience
<ProGuy> water: Probably
<water> self is just as fast as c++ and is much higher-level, and much simpler in design and use
<Miellaby> I think we can include some very High level concepts into a very low language, the nearest one from the kernel.
<ProGuy> water: I can only base my opinion on my own experience
<Miellaby> 'self' is a kind of Lisp ?
<water> self is a kind of smalltalk without classes or enforced variables
<Miellaby> It seems interesting.
<water> abi self?
<abi> self is a prototype-based object system or at http://self.sunlabs.com/
<water> java actually inherited its lame compilers from self's much better ones
<Miellaby> How can we forgive classes ? I though they were necessary.
<ProGuy> What about python? 
<water> there's also the oberon language, where code is loaded and stored as abstract syntax trees. code-generation and run-time optimization allow incredible speed-ups
<water> miel: cloning replaces both subclass instantiation and instance creation
<Miellaby> Are we agrry on the fact that's inheritance is a major relationship between entities ?
<water> no, it is minor
<water> inheritance has nothing to do with object-orientation
<Miellaby> I think that's the real progress.
<water> what's the real progress?
<Miellaby> component-building programmation exists from years. Building an application with some Black-box entities is a good thing. But carried/carrier relationship dont avoid completly the amount of redondant data. Only inheritance does.
* Miellaby/#TUNES is searching for an example.
07:00am
<water> black-boxs are bad!
<water> inheritance is just as bad as code-inclusion
<Miellaby> about black-boxes.  I though about the old "System" model uses in the 70'.
<water> huh?
<water> ever looked at modula-3? everything in the language is modular and relatively safe, but still open
* ProGuy/#tunes has to learn some more languages....
<water> i haven't mentioned a single language yet that isn't on Tunes' language review page
<Miellaby> System modeling : Just a way to describe applications through 'filter likes black box'. Ada is a good exemple of modular-langage either.
<water> modula is a much better example, because it doesn't suffer from feature bloat and complexity
<water> oberon is also a good example, because the intermediate code is trivial to check for bugs, and still is portable across architectures while being simple to optimize at run-time
<Miellaby> OK. Now, could you sump up all these languages for me ;)
<Miellaby> Seriously. What's more important thant inheritance ?
<ProGuy> Maybe I should check those langauges, and learn them...
<water> inheritance is inflexible... if you want to develop your code, you either specialize a class or you're out of luck
07:10am
<water> and you can only re-use code from a superclass, which is messy anyway, especially if the implementation changes
<Miellaby> Those problems dont seems related to inheritance as for me.
<water> what?
<water> not related?
<water> inheritance forms a hierarchy for code-reuse and conceptual structure
<water> it actually limits how you can re-use code
<Miellaby> hmmm well inheritance is the only way to structurate entities.
<water> ever tried using smalltalk?
<water> hell no!
<Miellaby> Just a minute. I think we dont't understand together. What do you mean by inheritance ?
<Miellaby> smalltalk is based upon inheritance relationship, isn't it ?
<water> yes, and it helps illustrate why inheritanc eis bad
<Miellaby> wow!
<water> miel: read the glossary entry on inheritance
<Miellaby> I'll do it.
<Miellaby> OK. I forgive. I'm going to read more precisly TUNES pages.
07:20am
* Miellaby/#TUNES is depressed.
<Miellaby> Well, I feer I dont understand Fare's explanation.
<water> hmm
<Miellaby> How TUNES project intends to factored data ?
<water> in ways that depend on the actual data under consideration
<Miellaby> Could wa take an exemple.
<water> in other words, no single scheme for everything
<Miellaby> From a low level point of view, we are obliged to choose a system of knowledge representation.
<water> basically, tunes should allow you to modify data in the traditional way, which is the way of prototyping. but it also allows reflection to modify the data in ways that can't ordinarily be done
<water> sort of
<water> actually, my solution to it is like that
<water> (knowledge-representation style)
<water> but kr is presently very limited. my idea is much more general
07:40am
-:- Miel [sgarden@Rennes-Villejean-4-143.club-internet.fr] has joined #TUNES
-:- SignOff Miellaby: #TUNES (Ping timeout for Miellaby[Rennes-Villejean-4-143.club-internet.fr])
<water> wb
<Miel> ...
<Miel> sorry
<Miel> .
<Miel> Please water. You was so nice with me. Could you still answer an additionnal question ?
<water> sure
<Miel> Let's consider some differents kinds of entities...
<Miel> well, like : dogs, cats, and so one.
<water> not computational, but that's ok
<Miel> How, in tunes, do you think we can gather the common points of those entities.
<water> kr, i suppose, to state the relationships
<water> prototyping to model the taxonomy
<Miel> yes.
<Miel> When we define types, We can compare type/object relation with inheritance, can't we?
07:50am
<water> i guess
<Miel> ok. Then I'm happy.
<water> i'm not
<Miel> why?
<water> i don't like the type/object relation at all
<Miel> what do you mean by "prototyping"?
<water> it introduces conceptual complexity that eventually will show up elsewhere
<water> prototyping takes an object as a bunch of slots and just modifies it to make a new one
<Miel> like clonig?
<Miel> cloning
<water> yep. prototype = clone and modify
<Miel> do you keep a relation bettween the 'old object' and the 'new object'
<Miel> ?
<water> you can, but you can also separate the new object from its parent
<water> so that it only inherits from 'root' or something
<Miel> and then, what would you do with the case of cats/dogs animal entities?
<water> whatever i want
<Miel> would you keep a relation with a factored-entity called 'furry animal'.
<water> if it were useful, yes
<water> keep in mind that this is not my approach to tunes at all
<water> btw, prototyping assumes a "dynamic" language
<water> not a static one like java
<Miel> whatever the approach. aren't you agrree on the face that it's necessary to keep a relation between entities who are equals by their structures?
<water> yes, but i wouldn't define things based on their structures, i would define structures based on structures :)
08:00am
<water> so, the relation is not between the things, but the structures themselves
<Miel> It's difficult to imagine.
<water> ever looked at hyper-programming?
<Miel> or meta-programming ?
<water> text-identifiers replaced by direct links
<water> what about meta-programming?
<Miel> I think I can see what hyper-programming looks.
<water> imo, hyper-programming should include language keywords in its link system
<water> that would allow some interesting reflection
<water> unfortunately, the Napier language family hasn't looked into it
<Miel> I understand what's kind of approach you follow. I suppose you see a very dynamical computational system where the user always defines news concepts, deletes somes and clone other.
<water> sure, a system that evolves into supporting new applications at the language level
<Miel> you dont do a difference any more bettwen the developper and the final user.
<water> right, which is a tunes goal
<water> i mean, a Tunes Goal (TM)
<water> :)
<Miel> ok,ok, let's suppose TUNEs already exists and runs from years.
<Miel> To my mind...
<water> ...?
<Miel> without factoring, each system instance would have becomen a bunch of unindexed data like the world wide web.
<water> ok, but factoring doesn't have to be hierarchical or even defined from a single perspective
<water> imo
08:10am
* Miel/#TUNES : what means imo ?
* Miel/#TUNES I remind.
<water> in my opinion
<Miel> ok. :)
<Miel> I have to go. I'll try to better understand reflection.
<water> k
<Miel> thanks a lot
<Miel> bye.
<water> no prob
-:- SignOff Miel: #TUNES (Leaving)
<water> bye all
-:- water [water@ppp-tnt-233.tscnet.net] has left #tunes []
-:- hcf [nef@me-portland-us301.javanet.com] has joined #tunes
08:20am
-:- SignOff ProGuy: #TUNES (Ping timeout for ProGuy[p432-054.ppp.get2net.dk])
-:- ProGuy [Paul@p432-054.ppp.get2net.dk] has joined #tunes
-:- smkl [sami@MXLVII.rdyn.saunalahti.fi] has joined #tunes
-:- ShaolinMonk1 [Whyme@1Cust38.tnt41.nyc3.da.uu.net] has joined #tunes
-:- ShaolinMonk1 [Whyme@1Cust38.tnt41.nyc3.da.uu.net] has left #tunes []
-:- SignOff smkl: #TUNES (Ping timeout for smkl[MXLVII.rdyn.saunalahti.fi])
-:- water [water@ppp-tnt-46.tscnet.net] has joined #tunes
-:- smkl [sami@CMLIV.rdyn.saunalahti.fi] has joined #tunes
-:- SignOff hcf: #TUNES (Leaving)
-:- hcf [nef@me-portland-us245.javanet.com] has joined #tunes
<water> wb
10:40am
<water> abi: omega is a self-like language with static typing at http://infosoft.soft.uni-linz.ac.at/Info/Omega.html
<hcf> odd, that should of been in there, my bad
<water> no prob
<hcf> water: gonna write up a review of it? ;)
<water> i'm thinking of writing a small programming system for the aplm pilot
<water> s/aplm/palm
<hcf> smalltalkish?
<water> sure, i'll write up a review
<water> yeah, i'll probably re-use a lot of squeak code
<water> or self code
<water> the major thing is to support the stylus and make it small
<water> on the other hand, perhaps i should stick to working on arrow :)
11:00am
<hcf> water: any idea for a name of ur palm squeak?
<hcf> i suggest, chirp
<water> PipSqueak? :)
<hcf> thats good
<water> i'd better post to the squeak mailing list
<hcf> i tend to make names 5 letters or less
<hcf> pip works
<water> yeah
<water> it would be nice to have a self, though
<water> it would ssave memory and be faster, but take more work to make
11:10am
<hcf> a small self called, mini-me ;)
<water> haha
<hcf> pip clashes w/ the CP/M, RSX-11, RSTS/E, TOPS-10, and OS/8 copy command in the program-name namespace, no clashes in the proglang-name namespace
<ProGuy> CP/M rules....
<water> how do you search for that?
<hcf> water: just w/ foldoc
<water> abi foldoc?
<hcf> water: which includes the langlist
<abi> i heard foldoc was greatly outdated. All lisps have been statically scoped since 1984
<hcf> abi: foldoc for pip
<abi> pip may be sought in foldoc at http://wombat.doc.ic.ac.uk/foldoc/foldoc.cgi?query=pip
<hcf> YA self name, ego
-:- SignOff water: #TUNES (Ping timeout for water[ppp-tnt-46.tscnet.net])
11:20am
-:- water [water@ppp-tnt-215.tscnet.net] has joined #tunes
<hcf> is there a smalltalk called babytalk?
<water> nope
<water> pipsqueak _is a company
<water> not related to computers, though
<hcf> company-name clashes are nearly unavoidable
<water> not sure if i like the "babytalk" name
<hcf> u could use goo & gah instead of foo & bar ;)
11:30am
<hcf> whats wrong w/ just 'pip'?
<water> oh, pip is good
<hcf> brb
-:- water has changed the topic on channel #Tunes to: Reflective Computing System <http://www.tunes.org> | http://osdevnew.listbot.com | http://www.researchindex.com | http://www.tsia.org
-:- SignOff ProGuy: #TUNES (Ping timeout for ProGuy[p432-054.ppp.get2net.dk])
<hcf> back
<water> k
<water> i'm still waiting for the squeak mailing list subscription to return
<hcf> wheres the smalltalk mlist archive?
<water> you can get to it from the homepage
11:40am
<hcf> which is where? /me being lazy (a rare occasion w/ urls)
<water> http://squeak.cs.uiuc.edu/mail/squeak/
<hcf> hmm, smalltalk.com is up for sale
<water> wow
<water> i've always been planning to start a software company/group
11:50am
<hcf> cool
<hcf> for doing what?
<water> based, of course, on tunes/arrow ideas.
<water> "technology for the mind"
<water> :)
<water> toys (software and otherwise) and tools
<hcf> inet ready "barbie" dolls and gi joes?
<water> all designed to diversify the market
<water> lol
<water> more along the lines of puzzles that teach by being embedded into an internet/arrow style information system
<water> of course, i'd also like to sell high-end unified expert systems with "arrow smarts"
<water> to, say, automate (not aid) engineering and the information processes involved
-:- Crimson_ [crimson@chaosdev.org] has joined #tunes
<water> hopefully, i could run the company forever with 30 or less people, and if it grew, the policy would be to split the group
12:00pm
-:- SignOff Crimson: #TUNES (http://www.mistik.net)
-:- SignOff Crimson_: #TUNES (http://www.mistik.net)
-:- Crimson [crimson@chaosdev.org] has joined #tunes
<Crimson> Yo!
<water> wow. you're actually here!
<Crimson> Who?
<water> you
* Plundis/#tunes too :)
<Crimson> thats right. =)
<water> geez, the statues come to life.
<water> you guys lurk too much
<water> so, what's up?
<Plundis> storm 0.2.0 is out
<water> what's that?
<Plundis> eihrul has written som IPC stuff.
<water> cool
<Plundis> water: check out www.chasodev.org
<water> k
<water> oh, the chaos uK
<Plundis> yes. :)
<Plundis> of course.
<Plundis> haven't you checked it out?
<water> not recently
<water> uK's aren't related to my work
12:10pm
<Plundis> oksu..
<Plundis> okay even.
<Crimson> wherer is sadasdasdasda is eihrul?
<Plundis> =)
<Crimson> hehe..
<Crimson> buggy keyboard...
* Crimson/#tunes is sitting in a very small room together with plundis doing hard work on the storm.
<water> hey, you guys have heard about the motherboard design ibm released for ppc g3 and g4, right?
<Crimson> no.
<water> it's on slashdot
<water> the article, that is
<water> the design is completely open, and ibm wants companies to make linux boxes
<Crimson> havent been online at all lately...
* Plundis/#tunes haven't been online for more than a week..
<Plundis> water: cool. but Linux sucks.
<water> hello?
<water> port your system to it! (eventually)
<Plundis> of course.
<Plundis> we will, be sure.
<Plundis> but we concentrate on Intel-clones for a start.
<water> sure
<water> how integrated will tornado be to the system?
12:20pm
-:- SignOff water: #TUNES (Ping timeout for water[ppp-tnt-215.tscnet.net])
-:- water [water@ppp-tnt-25.tscnet.net] has joined #tunes
-:- SignOff Crimson: #TUNES (http://www.mistik.net)
<Plundis> water: it will be a natural part of the system.
<Plundis> but not required.
-:- Crimson [crimson@chaosdev.org] has joined #tunes
<water> so, it won't be gui-based
<Plundis> well, the chaos system will be gui based.
<Plundis> but it will be possible to use the system without a GUI, if you want to.
<Plundis> like in Linux, but with a better GUI integration.
<Crimson> but most our programs will be designed to use tornado.
<water> ok
<Plundis> yeah.. the installation procedure will probably be graphical, for example.
<hcf> w/ a tetris game too? ;)
<Crimson> is there anyone here who knows much about the vga-hardware?
<Crimson> hcf: yeah.. =)
<Plundis> hcf: well.. =)
<Plundis> hcf: that's just for fun. :)
<Crimson> no, for testing.
<Plundis> Crimson: why not?
<Plundis> I WANT EIHRUL!!!
<Plundis> seen eihrul
<Plundis> isn't there a bot in here?
<hcf> abi: seen eihrul
<abi> eihrul was last seen saying something on IRC Sat Aug 14 00:52:10 1999
<Crimson> We have a running chaos-system in this room right now. It rules! =)
<Plundis> oh..
<hcf> thats like bad data
<Plundis> well, we'd better write the message system ourselves.. it's not very complicated.
<hcf> s/like/likely/
12:30pm
<water> k, so your ipc is the usual uK messaging
<water> darn it! i'm still not subscribed to the squeak list
<Plundis> squak?
<hcf> abi: sqeak?
<abi> wish i knew, hcf
<water> lol
<hcf> abi: squeak?
<abi> squeak is a cool language descended from Smalltalk, at http://squeak.cs.uiuc.edu/ or at http://www.squeak.org/
-:- SignOff Crimson: #TUNES (http://www.mistik.net)
-:- Crimson [crimson@chaosdev.org] has joined #tunes
<Crimson> ej ull
12:40pm
<Plundis> no wool?
<Crimson> argh! found a bug in the keyboard server. MUST be fixed noW!!!
<Crimson> where is eihrul?
<abi> it has been said that eihrul is making YAMK :)
<Crimson> cool...
<Crimson> abi: crimson?
<abi> it has been said that crimson is currently working on the ide-driver for my os. or very strange
<Crimson> abi: forget crimson
<abi> Crimson: I forgot crimson
<Crimson> crimson is working on storm, the microkernel in a chaos system.
<Crimson> plundis?
<abi> plundis is the ruler of the known world. I control this world at the top of Plundis's fingertips.
<water> lol
<Plundis> =)
<Crimson> =)
-:- Crimson is now known as Hans_Oulu
12:50pm
<Hans_Oulu> water: are you having trouble with your internet-connection?
<water> actually, there's a hardware conflict that occasionally causes this machine (a friend's) to lock up so i have to reboot
<Hans_Oulu> water: bad thing. os?
<water> hehe... win98r2
<Hans_Oulu> =)
<Plundis> water: are you kidding??
<Plundis> win2k ruuulez! ;)
<Hans_Oulu> japp.
<water> i only have a laptop myself, and am looking to get a desktop soon
<water> hopefully i can get one with the beos or linux installed
<Hans_Oulu> what type of connection are you using?
<water> 56k modem
<Hans_Oulu> cool. we too,
<water> via usb port
<Hans_Oulu> elite.
<water> (it's one of the things causing the lock-ups)
<water> (the other is a rage128)
<Hans_Oulu> wanna write a usb-server for chaos?
<water> no thanks, the only os i ever wrote was a hacked-together pmode programming system
* Hans_Oulu/#tunes is mIRC:ing on a laptop. =)
<water> i work on cs theory, mostly
<Hans_Oulu> what?
<water> huh?
<Hans_Oulu> cs theory?
<water> computer science
<Hans_Oulu> ahhh...
<water> actually, information theory and mathematics
<water> heard of my Arrow system idea?
<Hans_Oulu> cool. I am currently writing a scheduler for storm.
<Hans_Oulu> No, no Arrow system...
<water> it's actually more abstract than tunes, if that's possible, but it does have a formal theory
<Hans_Oulu> so it is an os?
<water> no, but it eventually will be... i'm starting from the top down
<Hans_Oulu> i have never understood what tunes actually is.
<water> most c-coders don't get it
01:00pm
<water> it's basically a linux-style os with a language so good and powerful that it can re-compile any part of the system, including the os, dynamically
<water> (tunes, that is)
<Plundis> but how is that supposed to work?
<Plundis> will the operating system be running under an interpreter like Emacs or something?
<water> hehe.. that's what all of the tunes site and papers are about
<water> yech
<water> sort of
<water> but you won't be limited to text at all
<Hans_Oulu> ok... i am concentrating on my own project primarly.
<water> sure, no prob
<Hans_Oulu> that wont be very fast, will it+
<water> yes, because reflection allows optimizations to be made without changing source code
01:10pm
<water> there's even a specialized form of reflection call aspect-oriented programming that handles issue-oriented optimizations
<water> s/call/called
<Hans_Oulu> yeah yeah, will it be multitasking at all?
<water> well, separate memory spaces will be used only for emulations and untrusted code
<water> we hope that tunes will allow all of the high-level, provable code to run as threads within a single process, along with device driver objects, etc
<water> also coroutines
<Hans_Oulu> that sounds fast, but unstable
<water> well, the key words are high-level and trusted (provable)
<Hans_Oulu> water. is there a patch for grub to boot over a network?
<water> dunno
<Hans_Oulu> ok..
<Hans_Oulu> is going to change back to my usual nick. =)
-:- Hans_Oulu is now known as Crimson
01:20pm
<Plundis> anybody here who would like to mirror www.chaosdev.org abroad?
01:30pm
-:- core [core@core.suntech.fr] has joined #tunes
<core> people hi
<water> hi core
<core> water! our island of intelligence. hi :)
<water> hehe
<core> still not afloat?
<water> i only go to sea for three weeks in september. after that, i will be working on land for a year or two
<core> ahh, ok. we won't deprived from communication with you for too long then
<water> nope
<core> good :)
<water> what's new for you?
-:- [AnimaL] [something@pm2-19.wmbg.widomaker.com] has joined #tunes
<water> hello, animal
<core> hmm.. well, still working hard on clementine, i almost have a fully functional system now besides low-level support for persistent store. and i guess my life is good lately, nothing terribly exciting tho :) what about you ?
<[AnimaL]> hello
<abi> bonjour, [AnimaL]
<core> plundis: are you intending to make your repository work sometime?
<[AnimaL]> hey
<water> i'm still working on my theories, but my laptop is very sick lately
<core> water: hmm. time to get a Vaio :)
<water> i have had to use this computer (a friend's) to get anything done
<[AnimaL]> where can i get information on os development?
<water> os web?
<water> the os-ring?
<[AnimaL]> url?
<core> animal: *shrug* all over the place, you can check http://www.tunes.org/Review/OSes.html for a bunch of pointers.
<[AnimaL]> yea, i saaw that
<[AnimaL]> saw even
<core> well, there are tons of pointers there.
<[AnimaL]> cool
<[AnimaL]> are there any os's written for x86 in pure assembler?
<water> hehe... boy, are there
<core> animal: more than enough
<[AnimaL]> cool
<[AnimaL]> pmode?
<abi> well, pmode is *fun* :)
<[AnimaL]> heh
<core> abi: forget pmode
<abi> core: I forgot pmode
<[AnimaL]> what are the names of some of them?
<core> animal: yes, i guess it's more fun to eat razorblades than do real-mode OS work.
<[AnimaL]> hehe
<core> animal: check out Retro, a potential prototypes for tunes
<core> animal: brix is still vaporware..
<[AnimaL]> do u have the url?
<core> animal: and i don't know, there are a lot on that page.
<core> animal: it's in the os review, that's why we made that page
<core> retro?
<abi> retro is at http://bespin.cx/~tcn/retro.html
<core> brix?
<abi> brix is an OS at http://www.qzx.com/brix or still not released yet or still open binary or erotically stimulating to some
<[AnimaL]> ok
<[AnimaL]> i saw brix
01:40pm
<water> http://osdevnew.listbot.com
<water> http://www.500mhz.net/resources.html
<water> actually, http://www.500mhz.net/
<[AnimaL]> ok
<[AnimaL]> retro is in pure asm?
<core> animal: yeah; it's growing a forth interpreter on top of it, but it's asm as well.
<[AnimaL]> ok
<smkl> does L4 have sources released?
<[AnimaL]> what's cvs?
<abi> cvs is probably Concurrent Version System at http://www.cyclic.com or documented at http://www.loria.fr/~molli/cvs/doc/cvs_toc.html or at http://www.gnu.org/software/cvs/cvs.html
<[AnimaL]> hmm
<core> animal: a version control and collaborative development system.
<[AnimaL]> err, what's it do and how does it work?
<core> animal: basically it maintains a remote server ("repository") of all the sourcecode of a project, and any number of people can sync with it (downloading changes) and update it (commiting, ie. making changes).
<Crimson> cool! lots of people... =)
<core> animal: it keeps track of all changes made since day one too, so you can "roll back" in time and ask the version of, say, august 3 1988 :)
<[AnimaL]> coo
<Crimson> core. hi, pal...
<core> crimson: i still can't access your cvs repository
<core> crimson: and it's not that i care, but the THANKS file seems devoid of my name :P
<core> crimson: hi :)
<Crimson> anonymous access to the cvs dont work. =)
<[AnimaL]> so it uploads the src files to a server?
<core> crimson: no shit :)
<core> animal: yes, it uploads differences.
<Crimson> core: you can download the source from our download site now.
<core> crimson: ah?
<[AnimaL]> that's kinda cool
<core> crimson: and, umm, where? :P
<Crimson> www.chaosdev,org/files/chaos-current.tar.bz2
<Crimson> updated daily by a cron-script.
<core> crimson: ah, so it'll be updated tomorrow :P
<Crimson> it is now.
<[AnimaL]> cvs is for nix?
<core> animal: you have 'doze clients as well.
<water> there's also wincvs, but it's pretty buggy
<[AnimaL]> ok
01:50pm
<Crimson> core: adding you to te THANkS file...
<core> crimson: they all say that ;)
<Plundis> Crimson: you are psychic? :)
* Plundis/#tunes thinks Crimson eavesdropped his monitor.. :)
<core> crimson: add "fix the cvs repository" to the TODO list :)
-:- eihrul [lee@usr5-ppp194.lvdi.net] has joined #tunes
<Crimson> core: plundis is working on it rght now.
<Plundis> eihrul!
<Crimson> eihrul: WOW! 09
<core> crimson: it's okay really.. :)
<eihrul> crim: eh?
<Crimson> hehe... broken shift-key. =)
<Crimson> eihrul: chaos works almost perfectly now.
<eihrul> ALMOST :)
<eihrul> perfection being unattainable
<Crimson> eihrul: the ide server and the 3c509 server is working.
<Plundis> eihrul: we need messaging.
<core> can you run programs from the shell yet? :P
<Crimson> core: just the builtin tetris. ;)
<core> crimson: hehe
<Plundis> eihrul: can you fix it now so that we can continue coding? :)
<core> crimson, plundis: i updated the OS review btw.
<Plundis> core: great!
<core> seems like your most direct "competitor" is openblt.
<Plundis> core: is it good?
<Crimson> core: COOL!
02:00pm
<core> plundis: well, it's also an uK, it seems quite nice even for an uK, seems quite organised, and the people keep a low profile.
<core> plundis: and they are mostly where you are at (tell me if i'm mistaken), ie, a functional microkernel, short of filesystem support.
<core> (well, they have message passing.)
<core> eihrul: does your message system support fragments?
<core> (or whoever message system it is)
<eihrul> it should
<core> eihrul: it "should"? :)
<core> eihrul: fragments as in multipart messages.
<eihrul> yep
<eihrul> because it doesn't know what a message boundary is
<core> eihrul: so how do you know how much memory to map or copy?
<eihrul> the way i've thought it out... you can read one message in multiple parts
<eihrul> min (sender_length, receiver_length)
<core> is it copied or just shared memory?
<Crimson> core: the review page ruled! =)
<core> crimson: does it? thanks
<eihrul> core: i'm having trouble deciding how to do the inter address space copies right now
<eihrul> i'm kind of torn between mapping on demand, or disabling paging
<core> eihrul: ok, rule of thumb: "copy large amounts of data" == "bad". "remap little amount of data" == "bad".
<core> eihrul: er? _disabling_ paging?
<Plundis> core: I think you will be able to checkout now..
<eihrul> well, for small messages... you can just use the registers
<core> eihrul: yeah, like L4.
<eihrul> but how do you do messages not large enough to fit in there?
<core> plundis: ok, let me try :)
<core> eihrul: well, either copy, or shared memory. copying a ton of memory is bad.
<core> eihrul: clementine components do not do explicit message passing, but the RMI uses shared memory when done locally across different address spaces (Fare, it's not in cvs yet :P)
<eihrul> core: well, my kernel is providing real shared memory as well
<core> plundis: cvs server: failed to create lock directory in repository `/usr/local/cvs/chaos': Permission denied
<eihrul> but that doesn't mean i want to totally disallow messages > 16 bytes
<core> eihrul: i don't understand what shared memory has to do with messages < 16 bytes.
<core> plundis: it doesn't work :/
<core> plundis: try it yourself from a clean directory?
<eihrul> core: it doesn't... grrr
<eihrul> core: so how do i handle such messages? :)
<core> eihrul: well, copying less than 16 bytes is very fast, or it can be passed in registers i guess.
<core> eihrul: if you support thread migration for synchronous IPC, there is little problem with that actually.
<eihrul> you can't copy directly between address spaces easily though
<core> eihrul: well, again, for synchronous IPC it's easy; for large messages or async IPC, you need to either do two copies (source->kernelspace kernelspace->destination) or support shared memory.
<Plundis> core: it takes too long over a modem line.. :)
<core> plundis: well, it'll fail quickly, even over a modem line :)
02:10pm
<eihrul> core: i provide shared memory in addition to message passing :)
<core> eihrul: so you can provide a "it's going to be shared" attribute for the source thread that allocates memory, so you can directly map the message memory into the destination, and there you go, they'll have a lot of children and be happy?
<core> well, if you have memory allocation, that is. i don't remember if Storm has that :)
<Plundis> core: the problem is this: the chaos group is able to read/write to the cvs repository. but guest isn't a member of this group, and cvs seems to want to create lock files in each directory..
<core> plundis: right
<core> plundis: so add guest to the chaos group. and set it in the "readers" file so cvs won't let guest write anything.
<Plundis> how do I do? i'd better try to turn locking off.. or something.
<core> plundis: locking is mandatory
<core> plundis: edit "readers" in your CVSROOT and add guest on one seperate line to it
<core> plundis: then add guest to the chaos group
!core:*! how long does it take the average user to get the hint from nickserv, usually?
<core> chaos: we're setting up a box dedicated to opensource projects, at suntech, this month; we'll be able to host services for chaos by then, if you need to; mailing lists and whatnot (i assume we'll at least host a mirror of Tunes too).
<core> plundis: there, it works!
<core> ah, nevermind :(
<core> does chaos support thread migration?
02:20pm
<water> ok, i've sent the Squeak project an e-mail with my proposals
<water> oops.
<water> my machine will have to reboot soon
<water> brb
-:- SignOff water: #TUNES (Ping timeout for water[ppp-tnt-25.tscnet.net])
02:30pm
-:- hcf_ [nef@me-portland-us308.javanet.com] has joined #tunes
-:- SignOff hcf: #TUNES (Leaving)
-:- hcf_ is now known as hcf
-:- Tril [dem@bespin.dhs.org] has joined #TUNES
-:- mode/#Tunes [+o Tril] by ChanServ
<hcf> hoy Tril
<Tril> hi
<abi> hello, Tril
!HyrlikW:*! .
<core> hey tril :)
<Tril> yo
<core> how's your lisp work? :)
<Tril> i'm not really working on lisp code right now..
<Tril> i'm flushing out some design
<core> fleshing or flushing? ;)
<Tril> fleshing maybe is the word
<core> hehe.. freudian slip
<core> and I'M the non-native speaker ;)
02:40pm
<core> what are you designing anyway?
<Tril> a framework for interoperating language paradigms
<Tril> declarative/functional/procedural, including OOP feature of poly
<Tril> then every known language can be added as an extension of the framework
<Tril> and share one persistent store 
<Tril> does that make sense?
02:50pm
<core> sorry, i was away, girl issue ;) yeah, it does, although i have no idea HOW you would do that :)
<Tril> there will probably be a short paper on it ...
<core> i'm really interested to know how you would do that
<core> seriously 
<Tril> well , it basically goes like this
<core> that's a problem that will happen when, say, our very own fare starts writing his compiler finally and wants to operate it with clementine.
* core/#tunes listens
<Tril> functional can be turned into procedural by explicitly saving arguments into variables. so instead f(g(x)) write a=g(x), then in sequence f(a)
<Tril> the reverse , procedural into functional, can be done by some complicated tracking of assignment statements (static analysis) as long as you allow functions that return more than one arg, accept more than one, etc
<core> okay.. you only need to do that explictly, at a boundary between two paradigms, though?
<core> hmm.. returning more than one argument might be bad in procedural languages
<core> the complexity is a compile-time issue, right?
<Tril> to answer the first question, i'm not sure when you NEED to do that, it's for when you want to convert between wto languages
03:10pm
<Tril> for the compile time question, i'm talking about source level changes so it's all pre-compile time
<core> ohhh. okay, good :)
<Tril> I mean, before compile time, not at precompile
<core> yes i understand
<core> i didn't know if you were standing at meta-sourcecode to sourcecode generation level, compilation level, or execution level ):
<core> :)
<Tril> declarative is more interesting. the idea is to declare what you want so specifically that the functions to get there and their combinationms can be automatically determined. But if the computer can't figure it out, the user can help
<core> well, for the "when", i guess you need to do it when mixing two different paradigms
<core> yeah.. you want to declare as much as needed to map to all possible paradigms ?
<Tril> dunno
<Tril> so it's not really just converting paradigms, the framework is reflective, it's the tunes HLL
<core> well, not converting, more like smoothening edges between them :)
<Tril> glue
<core> right
<Tril> sure
<core> that assumes that the whole system can comply to the tunes HLL tho
<Tril> brb
<core> what about temporary runtime glue? ie. to start running tunes HLL on top of retro or clementine, with a low-level calling convention? that was fare's preoccupation iirc. of course more and more of the system would be replaced by HLL until none is left and the convention can go.
<Tril> back
<Tril> the problem is always emulating a precise high level specification using some limited lower layer. 
<core> yeah
<core> it doesn't have to be performant though, it's temporary
03:20pm
<Tril> well, it's better to have HLL interact directly with hardware. 
<core> yeah i know, but assuming we start writing HLL on top of some system like retro, clementine or whatnot
<Tril> but it will always be an inexact match, there's always room to have multiple implementations for different situations
<core> for some time (ie, until the system is all HLL) it'll need a runtime glue
<Tril> oh that
<core> so that, say, components written in HLL can use code still written in C components
<core> until the system has completely been borgified :)
<Tril> something like that. im not too worried about implementation stuff. i'm trying to focus on what the system will be like after it's bootstrapped
<Tril> then maybe even someone else can do the implementation like fare
<core> yeah i understand
<core> well in this case your work is important for the future then :)
<Tril> right
<Tril> see you later..
* Tril/#TUNES is away: (afk) [BX-MsgLog Off]
03:30pm
-:- SignOff core: #TUNES (night :))
-:- vanillaca [user6876@209.138.160.210] has joined #Tunes
<vanillaca> hello
<vanillaca> Hi Mr_Wrong
-:- water [water@ppp-tnt-243.tscnet.net] has joined #tunes
-:- vanillaca [user6876@209.138.160.210] has left #Tunes []
<water> hey all
<eihrul> lo
04:40pm
<water> darn it, i missed tril
<hcf> water: read the log
<water> doing that now
<water> that's why i regret missing him
* Tril/#TUNES is back from the dead. Gone 1 hrs 17 min 39 secs
<Tril> darn it what do you want
<water> sorry, didn't mean to bother
<Tril> i just feel guilt for not doing more work on tunes :)
<water> i was just interested in what you were discussing
<water> as do i
<Tril> you have question?
<water> did you read the arrow logic intro?
<Tril> no
<Tril> is it long?
<water> er...34 pages
<water> but its pretty easy to read
<water> imo
<water> btw, my approach is similar to what you were discussing
<water> except that i'm not concerned with "3 paradigms" of computation specifically
<water> i was thinking along the lines of a continuum of computing models
<water> build from arrows ;)
<water> s/build/built
04:50pm
<Tril> hm
<water> anyway, that intro covers a lot of the material of the logic and its relatives, mostly for conceptual explanation
<Tril> ok, i can't read postscript on this screen, i'll read it sometime
<water> hm.. ok thanks
<water> btw, i posted to the squeak mlist about making some changes to their vm
<Tril> did you make changes?
<water> not yet, but i intend to
<water> it might be possible to persuade them to convert to prototype-based object systems
<water> or a VM based on syntax trees and hyper-code vice bytecodes
<Tril> did you look at poplog? i didn't download yet
<water> a long time ago... it looked pretty complicated for me
<hcf> 'hyper-code vice bytecodes', vice?
<water> vs.
<water> hmm
<water> what about your type/object idea?
05:00pm
<Tril> obviously i'm trying to get a spec clear enough to implement
<water> k
<water> my laptop still won't play... it's getting really frustrating
<water> with all my new work on real paper, it's probably a good time to convert my writing process to TeX
<Tril> i cant give any sggestion, i havent dived into tex yet
<water> i guess i should mention that i've been battling boredom lately
<water> it's becoming really difficult to focus on my work
<water> however, i've been putting together lots of research, and hopefully i can find someone at a university who will co-write some papers with me
-:- liar [brand@p0wer.qzx.com] has joined #tunes
<water> hey liar
<liar> hey water
<eihrul> liar: hey :)
<water> tril: well anyway, i've at least managed to prove that all the papers i plan to write will be unique and will have sufficient sources to draw from
05:10pm
<water> tril: as for implementation, the squeak code should be enough.  i'm trying to bootstrap a useful environment from it
<water> does anyone have something to discuss?
-:- SignOff Tril: #TUNES (changing servers)
-:- Tril [dem@bespin.dhs.org] has joined #TUNES
-:- mode/#Tunes [+o Tril] by ChanServ
<Tril> am I still here?
>>> Tril [dem@bespin.dhs.org] requested PING 934676170 648332 from #TUNES
<water> hehe
<water> i'm a little lagged
<water> well, it does seem that my arrow quantifier concept should be sufficient for bootstrapping, given also the basic arrow operators
<water> in other words, if i have the squeak system answer deterministic queries about the arrow store, the quantifier introduces something(?) that allows for modelling anything else
-:- SignOff [AnimaL]: #TUNES (Read error to [AnimaL][pm2-19.wmbg.widomaker.com]: Connection reset by peer)
05:20pm
<water> (that's my current hypothesis)
<Tril> ok
<Tril> you tried to explain the quantifier yesterday, right?
<water> well, i tried to explain it a while ago
<water> i don't remember discussing it yesterday
<Tril> with the stuff about modal logic?
<water> oh yeah, but that's a different quantifier
<water> or not
<water> darn, i don't remember
* water/#tunes looks at the logs
<water> bleh... too much to look through
<water> it must have been a few days ago
<liar> wtf is a platinum mastercard?
<water> its a credit card with a _really_ high spending limit
<liar> kewl, i was just upgraded to one
<liar> it has a $1 million life insurance policy on it too
<liar> i just need to figure out howto cash that policy while im still alive
05:30pm
<Tril> grep
<Tril> dont worry about it, though :)
<water> i'm not :)
<water> anyway, the quantifier just returns a graph of all possible arrows satisfying the given on according to the 'current' ontology
<water> and yes, that's a really damn general purpose quantifier
<water> it would take a lot of work to make an ontology whose quantifer returned just a finite computable set
<hcf> water: the modal logic discussion was on #modtunes wasnt it?
<water> hmm
<Tril> sounds dirty to me, what is the current ontology? In my system what you describe would mean that the quantifier is actually a different quantifier for each ontology, and that you need 2 arrows one to say 'quantifier' and one to select the ontology (like an argument to the first arrow)
<Tril> but i'm using functions not arrows, so it couldb e completely different
<water> the ontology _is supposed to be implicit, but reifying it would introduce a _lot_ of complexity, so yes it is dirty. i think the point is to encapsulate all the dirty work in that one idea.
-:- SignOff smkl: #TUNES (sleeping ...)
<Tril> complexity is good when you are dealing with something complex
<Tril> dont overtsimplify (like ms windows)
<water> sure, have the formalism fit its subject
<water> one reason why i chose arrows as epistemic constructs
<Tril> it doesnt fit any subject?
<water> lol
<water> yeah, for the most part
<Tril> do you think i should make up new words, instead of using 'object', 'type,' function'?
<water> it might help to think of it in a different light, so yes
<Tril> like garf, blub, and varb
<water> i feel that those words probably make me assume too much about what they should mean
<water> hehe.. maybe something with a little more meaning :)
<Tril> oh, i thought you were referring to those
05:40pm
<Tril> if those have too much meaning, i guess i'll have to draw pictures with no labels :)
<water> do it with arrows, hint, hint :)
<water> but seriously, something with more meaning than "garf"
<hcf> abi: rot13 tunes
<abi> gharf
<water> ?
<Tril> hmm
<Tril> aib: rot13 arrow
<abi> neebj
<water> wtf?
<Tril> water, you dont know what ROT13 is?
* water/#tunes puts on the dunce cap.
<water> no
<Tril> you havent read Hackers by steven levy
<Tril> it's a simple encryption used by hackers
<water> probably not
<Tril> just split the alphabet in half and translate
<water> ah
<Tril> i.e. rot13 twice gets you back the way you started
<hcf> abi: rot13 is at http://ucsub.colorado.edu/~kominek/rot13/
<abi> vf ng uggc://hpfho.pbybenqb.rqh/~xbzvarx/ebg13/
<hcf> abi: \rot13 is at http://ucsub.colorado.edu/~kominek/rot13/
<Tril> uggc://, that's a good one
05:50pm
<hcf> does anyone have links for any xxx in multiple proglangs where xxx is not 99 beers, acm's hello world, or the rot13 url above?
<hcf> abi: quine is at http://www.nyx.net/~gthompso/quine.htm
<hcf> and not that one
<water> cool
<Tril> huh?
06:00pm
<water> i like the one for false that is only 34 bytes
<water> anyway, any ideas, tril?
<Tril> i have no idea what we are talking about
<Tril> hcf is just babbling nonsense
<water> the url is for self-printing programs in various langs. he was looking for other programs in multiple langs
<Tril> programs that are self printing in multiple languages? it seems hard enough to get a non-self-printing program to compile in multiple langs.
<water> no, each program in its own language
<Tril> nope, don't know any :)
06:10pm
<water> back to important stuff, then
<hcf> abi: forget hcf
<abi> hcf: I forgot hcf
<hcf> water: need palm links? ;)
<water> hcf: no thanks, but if you can find roms, that'd be cool
<Tril> what kind of roms
<water> palm pilot roms
<Tril> havent' got any of those either
* water/#tunes slaps Tril around a bit with a large trout
<water> the arrow logic intro uses the term 'frame' a lot
<water> it's basically the same idea as my graph notion
<water> the term 'frame' comes from modal logics, where the current group of logical objects has some relationships axiomatized over its members
<water> the logic occurs within these various worlds of objects
06:20pm
-:- pyro [tcn@cci-209150250093.clarityconnect.net] has joined #tunes
<water> hi pyro
<pyro> hey
06:30pm
-:- water has changed the topic on channel #Tunes to: Free Reflective Computing System at http://www.tunes.org/ | http://www.recordingwebsite.com/language.html
-:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us308.javanet.com])
-:- water has changed the topic on channel #Tunes to: Free Reflective Computing System at http://www.tunes.org/
<Tril> trying to read this arrow paper but there are too many symbols
<water> don't read too much into it
<Tril> I'm lost at Definition 3
<water> 1.3?
<Tril> page 3.  definition 3. But I'm going to eat , back in a bit
06:40pm
-:- overfien [overfiend@m226.telcomplus.net] has joined #tunes
<overfien> y0h
-:- hcf [nef@me-portland-us308.javanet.com] has joined #tunes
<water> wb
<water> hey fiend
<overfien> heh ;)
-:- depag [depag@phila-dialup552.nni.com] has joined #tunes
-:- depag is now known as sr
07:00pm
<hcf> abi: pocket smalltalk is a Smalltalk System for the PalmPilot at http://www.tiac.net/users/ajb/pocketst/
<water> whoa
<water> oh
<sr> water: do you own a palmpilot?
<water> not yet
<sr> acquire one :)
<water> darn it, the ide is for a pc
<water> but it could be ported
<hcf> abi: pocket smalltalk is also at http://www.pocketsmalltalk.com
<abi> okay, hcf.
<water> i could use pst and the squeak library to re-create the VM on the palm pilot!
07:10pm
<water> i'm sold
-:- ]AnimaL] [jayabdal@98CEEE3E.ipt.aol.com] has joined #tunes
<water> wb
<]AnimaL]> y0
07:30pm
<sr> hey animal
<sr> what's up?
-:- SignOff abi: #TUNES (Ping timeout for abi[bespin.dhs.org])
-:- abi [nef@bespin.dhs.org] has joined #tunes
<]AnimaL]> nothin
<]AnimaL]> printin out pentium man's
07:50pm
-:- SignOff pyro: #TUNES (pyro has no reason)
<water> wow. pst is much better than squeak for pda's
<hcf> water: note, u didnt ask for any searching to be done, i did a little anyway, and thot u had done a little, and at least had the pst link
08:00pm
<hcf> but doubted that it was in abi, so i added it
<water> thanks
<water> no, for some reason the sites i looked at didn't have pst
<hcf> set the sites right
<hcf> they've been wrong for far too long
<water> hehe
<water> yeah
<hcf> (lyrics from something)
<water> i think that i can extend pst's image to support a very small version of squeak
<water> but honestly i won't need much
<sr> water: will you ever switch to something different than squeak?
<water> sure, if self or something like it ever got good support, or if a really good lisp were made
<water> (of course, if tunes or arrow showed up, i'd switch in a flash)
<sr> what about C++
<water> bwahahhahaa
<water> that's pretty funny :)
08:10pm
<water> maybe straight c or better yet, c--
<sr> what about Forth?
<water> forth doesn't work well in anything besides a text interface
<sr> 3dForth
<water> forth with access to 3d api's?
<water> why?
<sr> forth represented in 3d
<water> whoa. that's different
<water> got a url?
<water> stackflow?
<sr> it's my idea
<sr> i'm just rambling
<water> actually, there's this stackflow project that uses a 3d representation of forth
<water> http://www.taygeta.com/forth_intro/stackflo.html
08:20pm
<water> hmm. it's just a visual representation... no implementation
-:- SignOff ]AnimaL]: #TUNES (Read error to ]AnimaL][98CEEE3E.ipt.aol.com]: Connection reset by peer)
<water> sr: sounds like a good opportunity for a project ;)
-:- sr is now known as tao
* tao/#tunes is an expression of water
-:- tao is now known as sr
<water> lol
<sr> water: you should work on a project...
<sr> a graphics engine, or a game
<water> hehe
<water> not until arrow theory is done and i have a good, useful implementation to let others improve
08:30pm
-:- SignOff water: #TUNES (Ping timeout for water[ppp-tnt-243.tscnet.net])
-:- water [water@ppp-tnt-73.tscnet.net] has joined #tunes
<water> damn, it's quiet
<water> time for a good rant, i think
<water> anyone want to discuss some details of what reflection is?
<hcf> ok
<water> wow. an answer
<water> here's something i've been tossing around in my head:
<hcf> which will likely be over my head
<water> we'd like to optimize things in a software-automated way
<water> we want the compiler to take our high-level code and notice places where optimizations can happen
<water> hcf: :P
<water> which means that the compiler must be able to get the necessary info from the programmer about the code or data
09:00pm
<water> like whether or not a sin function is better to implement in a straight-forward way or as a table of values
<water> someone or something has to decide on this. traditionally, either the programmer does this manually (and statically) and the software accepts the decision mechanically
<water> (ignore "either" from the previous)
<water> to change this situation, the software needs (more) information upon which to base its decisions
<water> the function specification for "sin" is declarative in any language. it is part of the declarative paradigm (why this is not called declarative programming, i'll never know)
<water> but this declaration gives the software the wrong (less useful) kind of information
<water> it's "low-level" - it declares that a specific sequence of instructions fulfills "what the user wants"
<water> the problem is that "what the user wants" should be central - that information should regard it rather than the machine. users persist; hardware (os's, api versions, ...) does not.
09:10pm
<water> my proposal? extend the declarative mechanisms in ordinary programs to encompass things that aren't algorithmically limited.
<water> of course, the immediate response is that the best i can do is to make a declarative language like Prolog or Mercury or Maude
<water> these of course are relatively weak programming languages, but programming is not the issue. the goal is to re-use information about user interests in general
<water> for instance, our "sin" function could be subject to various constraints of space and time. perhaps there is a memory limit for the look-up table or a time penalty for the calculation
<water> these constraints on implementation can easily be declared, but the first thing that comes to mind is that these declarations will be static
<water> to make this really general, we would have to be able to abstract ourselves from these situations
<water> we would go from "predicate(s) foo apply to x" to "predicate(s) bar imply that predicate(s) foo apply to x per this formula: ..."
<water> does anyone follow that so far?
<eihrul> sort of
* water/#tunes has obviously got everyone to ignore their channel windows :)
<eihrul> but of course, i'm not really paying attention
<water> oh
09:20pm
<eihrul> but the gyst of what your saying is the programmer should tell the compiler what it wants rather than what specifically might implement it? :)
<water> sort of... the programmer should also use a unified framework to the greatest extent possinle
<eihrul> my mind has wandered to that grey area before
<hcf> what about the user?
<water> same thing
<water> it's not a grey area for me... i'm studying ways to unify computation models with undecidable logics of all kinds in a single continuum
<eihrul> until someone does that, it's grey area :)
<water> hehe... use the arrow system :)
<water> (ok, so it's not very usable yet, but it's getting there)
<eihrul> for now... i'm stuck with C/C++/assembly and all the other yucky stuff :)
-:- SignOff overfien: #TUNES (Leaving)
09:30pm
-:- _water [water@ppp-tnt-189.tscnet.net] has joined #tunes
<_water> :P
-:- SignOff water: #TUNES (Ping timeout for water[ppp-tnt-73.tscnet.net])
-:- _water is now known as water
<eihrul> mmm... timeouts
09:40pm
<water> mmm... silence :)
<eihrul> mmm... interrupting silence
<liar> is there an easy way to append characters to alot of filenames?
<water> please do
<water> questions like that make me want tunes to exist today
<eihrul> in unix?
<liar> in brix i would just add a keyword to all the filenames
<water> the beos would probalbe handle that with a script language
<liar> eihrul: ya
<water> a c-shell script should do it
10:00pm
<water> i wonder if palm pilots can be over-clocked }:)
<water> i'll ask the experts
<hcf> abi: palmiga is at http://www.koeln.netsurf.de/~Richard.Koerber/palmiga/whatis.html
-:- SignOff sr: #TUNES (Ping timeout for sr[phila-dialup552.nni.com])
10:10pm
-:- SignOff water: #TUNES (Read error to water[ppp-tnt-189.tscnet.net]: Connection reset by peer)
-:- water [water@ppp-tnt-189.tscnet.net] has joined #tunes
-:- SignOff eihrul: #TUNES (Ping timeout for eihrul[usr5-ppp194.lvdi.net])
-:- eihrul [lee@usr5-ppp194.lvdi.net] has joined #tunes
<water> well, good night all
-:- water [water@ppp-tnt-189.tscnet.net] has left #tunes []
-:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us308.javanet.com])
11:00pm
<liar> im reading this ddj article and NT is really unsafe
<liar> a driver has access to everything
11:20pm
<eihrul> liar: could i check that code? :)
<eihrul> or were you just lying?
11:40pm
<liar> heh
<liar> wanna check some io code?
<liar> i have a bug in my io allocation code
<liar> and its funny cuz it should work since its based on the same code that my ram and memory allocator uses
<eihrul> sure
<liar> dword [offSYS+d_sys_used_nodes_io] = 3
<liar> and the offIOLIST = 0xf000
<liar> 0xf000 <bogus+0>:       0x000003b6      0x0000000f      0x00000044      0x00000004
<liar> 0xf010 <bogus+16>:      0x0000004c      0x00000368      0x00000000      0x00000000
<liar> i am attempting to allocate 2 ports at 3b4h
<eihrul> er ack
<liar> what
<eihrul> now someone else is bugging me :)
<liar> eh?
<eihrul> so you'll have to wait a few :)
<liar> so man u gotta make them wait :)
<eihrul> but i started talking to them before you
<liar> so
[msg(TUNES)] newlog 1999.0815
IRC log ended Sun Aug 15 00:00:00 1999