IRC log started Tue Jan 18 00:00:02 2000
[msg(TUNES)] permlog 2000.0118
-:- mikew [mike@TK137014.telekabel.at] has joined #tunes
12:10am
-:- hcf_ [nef@me-portland-us814.javanet.com] has joined #tunes
-:- mikew [mike@TK137014.telekabel.at] has left #tunes []
-:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us303.javanet.com])
-:- hcf_ is now known as hcf
<lar1> Who feels philosophical?
12:20am
<abi> i do
<lar1> Besides you abi
<eihrul> abi: how does one optimize a dynamically typed language such that it is as fast as the fastest statically typed language?
<abi> eihrul: i haven't a clue
<eihrul> gah
<lar1> eihrul:  Simple
<lar1> eihrul:  You can't
<eihrul> hah!
<eihrul> don't think so, man
<eihrul> it'll happen... some day
<lar1> Think about it, whenever you introduce overhead, you loose speed
<lar1> just like e=ir
<lar1> err
<lar1> forget that 2nd line ;)
<eihrul> umm...
<eihrul> no, you just introduce new challenges to be optimized away
<eihrul> you're just dislocating the work
<eihrul> away from the human
<eihrul> and towards the system
<lar1> Well
<lar1> If you design the language
<lar1> to take existing overhead, and move it
<lar1> then it can be optimized in the manner you desire
<lar1> The ultimate, of course
<eihrul> that statement was so vague it hurt
<lar1> s/vague/general
<eihrul> nah, vague... perhaps ambiguous
<lar1> Meaning you wish for me to elaborate or shut up?
<eihrul> or perhaps rephrase the statement to make sense :)
12:30am
<lar1> Ok, I'll try
<lar1> If one must jump through hoops becuase of static types in language x then dynamic types will be faster dispite the introduced overhead.  You remove the hoops.  Does that make anysense?
-:- ne [dtr@18-11.CampusNet.ucl.ac.be] has joined #tunes
<ne> ne1 alive here?
<lar1> Yes, do you need somthing?
<ne> was coming here to have a discussion about the post on slashdot
<lar1> Heh, thats what the influx of people is all about :)
<ne> yep
<eihrul> lar1: well, dynamic typing helps preserve genericity... so the challenge becomes to make it perform like static typing
<lar1> Well I think fare is idel at the moment if that is who you seek.
<eihrul> abi: fare?
<abi> hmmm... 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 or a comp.sci student (for phd)
<ne> i think it's legal here in belgium to release binary only inside a corp
<lar1> Inside an incorperation?  Thats leagal in the US, no?
<ne> i mean if that bin is coming from modified GPL'ed code
<lar1> eihrul:  Hmm, interesting
-:- NeBulOuS [hidden@sl4lh.dorms.usu.edu] has joined #tunes
<eihrul> eh... what's so interesting about saying dynamic typing needs to be as fast as static typing? :)
<air> is it true that bill gates is resigning?
<eihrul> he did
<lar1> NeBulOuS: Here to discuss the /. post?
<air> damn kewl
<lar1> air:  As, CEO, yes
<eihrul> air: he's still president or sumfin like that....
<NeBulOuS> nope, just looking for some asm gurus to help me with a mindlessly easy first assignment
<lar1> He is the cheif software acrhitect and still has a controling intrest
<NeBulOuS> which I WOULD do myself... had the book arrived by now
<eihrul> air: but microsoft is an evil all its own, bill gates is just a figurehead :)
<hcf> lar1: do NeBulOuS's homework for him
<NeBulOuS> hehe
<NeBulOuS> eihrul: true, true.
12:40am
<NeBulOuS> I said help, not do it for me :)
<lar1> NeBulOuS:  Well ask away
<ne> NeBulOuS is it x86 assembler?
<eihrul> well, if its mindlessly easy, why?
<NeBulOuS> 1.  Write a MIPS assembly program to add the four 32-bit word values at memory locations 1256, 1260, 1264, and 1268 then place the sum in memory word location 1272
* NeBulOuS/#tunes stares blankly
<eihrul> umm, just do 4 loads
<NeBulOuS> yeah.
<eihrul> add them all into an accumulator
<eihrul> then store
<NeBulOuS> yeah.
<NeBulOuS>  2. Translate your assembly program into binary and give the hexidecimal equivalent
<NeBulOuS> ...?
<abi> zero
<eihrul> NeBulOuS: uh, just assemble it.... print out the binary
<eihrul> or just get an opcode map
<air> NeBulOuS: abi just gave u yer answer :)
<eihrul> and do it in yer head
<NeBulOuS> any idea where I can get an opcode map?
<ne> www.eecs.harvard.edu/~nr/toolkit/specs/mips.html
<NeBulOuS> thanks bro :)
<NeBulOuS> "Chief Software Architect"... what's that all about?
<NeBulOuS> I'm sure Microsoft's software will be stable and bug-free from now on, with Gates in that position.
<lar1> NeBulOuS:  To cover his ass because of the Antitrust
<ne> NeBulOuS I think http://www.ee.duke.edu/~jab/ee152/toc.html is better than the former one, download the pdf there
<NeBulOuS> ok, thanks
<ne> isn't it Chief Software Acquierer
<NeBulOuS> holy... 800 pages?
<air> no, he put himself in that position so he will run the windows m$, ballmer will run the applications m$ and richard belluzo will run msn when the doj breaks m$ up
<ne> see page A-181
<air> NeBulOuS: hahahahhaha
<air> er, ne: hahahahhaha
<NeBulOuS> good thing I gots a 10baseT connection to download it with :)
12:50am
<ne> yeah i'm on t3 too
-:- SignOff ne: #TUNES (if fare's not here i shall go (gotta study Linear Circuits Ananlysis-- exam tomorrow))
* NeBulOuS/#tunes turns on the printer and prepares to print the book
<NeBulOuS> hehe
<NeBulOuS> "MIPS online documentation (800 page pdf file!! Don't print it, view it!!) "
<air> damnit, m$ settled with caldera
<NeBulOuS> yeah
<air> now i hate caldera
<lar1> old news, air
<air> why the hell would they settle
<lar1> time for bed
<lar1> later all
<NeBulOuS> keep source code secret?  not open up other possible lawsuits? who knows
-:- lar1 [larman@sdn-ar-005casfrMP093.dialsprint.net] has left #tunes []
<air> NeBulOuS: i was talking about caldera
<NeBulOuS> oh
<NeBulOuS> they must have liked the settlement terms :P
<NeBulOuS> Weird timing for the case, though, it was win95 that caused DR-DOS problems? DR-DOS used to be owned by novell, right? and novell would never sue MS, would they? 
* NeBulOuS/#tunes shrugs
01:00am
<air> caldera bought drdos from novell so they could sue m$
<NeBulOuS> DR-DOS isn't really a part of Caldera's business anymore, anyway.  I think they were happy with what they got for it :)
<air> well they should have done the right thing
<NeBulOuS> ?
<air> but instead they bent over and let m$ shove gobs of cash up their asses
<NeBulOuS> the right thing being...?
<air> not to settle
<air> m$ settled because they knew they couldnt win and that lose would hurt them in their other cases
<air> loss
<air> anyway im off to bed, cya
-:- SignOff air: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep)
<NeBulOuS> eh..
<NeBulOuS> nite.
<NeBulOuS> BILL GATES IS THE DEVIL!
<NeBulOuS> aha... now I feel better.
<NeBulOuS> try it guys
01:10am
<hcf> eihrul: here?
-:- NetSplit: varley.openprojects.net split from irc.linux.com [01:19am]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [varley.openprojects.net]
01:20am
<NeBulOuS> lets see
<NeBulOuS> Format:
<NeBulOuS> LW rt, offset(base)
01:40am
<eihrul> hcf: now i am
<eihrul> hmm
<eihrul> split
<NeBulOuS> so to load a value at location 1256 I would have to... what, lw $t0, 1256(?)
<NeBulOuS> help me out eih? :P
* eihrul/#tunes doesn't know the specifics of mips assembly.
<eihrul> why don't you know how to load a value from memory yourself?
<eihrul> assuming its a homework assignment
<NeBulOuS> because... I wasn't born knowing and the book is still in shipment
<NeBulOuS> sigh.
<eihrul> uh, why do you have an assignment on a topic you know nothing of?
01:50am
<NeBulOuS> because it's the first week of class.
<NeBulOuS> heh, that's why :)
<eihrul> eh, so why do you get assignments on something that you can't do?
<eihrul> there is some missing link here
-:- chriseb [chriseb@vortex.ukshells.co.uk] has joined #tunes
-:- chriseb is now known as cbz
<NeBulOuS> yes, there is.
<cbz> fare > just tried to send you an email but it bounced.
>>> hcf [nef@me-portland-us814.javanet.com] requested PING 948189495 from TUNES
-:- hcf [nef@me-portland-us814.javanet.com] has joined #tunes
<eihrul> hcf: what'd you need?
<hcf> eihrul: was just gonna suggest using the slate mlist for ur dynamic typing thing
<eihrul> just a random, inane statement brought up by the subject matter i'm reading
<eihrul> wishful thinking
<hcf> get the mlist rolling w/ something anyway
<eihrul> well, i believe all these papers will answer my questions :)
<eihrul> so before i get into a long mlist discussion its best i finish reading them
02:00am
* NeBulOuS/#tunes dies
<NeBulOuS> if I... only... had the book...
<eihrul> man, its in the manual!
<NeBulOuS> and so sue me for paying $47 instead of $85
<eihrul> this is not the right channel to come with assembly programming questions, necessarily
<NeBulOuS> yep.
<NeBulOuS> I actually just wanted to ask air... but anyway...
<eihrul> perhaps try #asm on efnet?
<NeBulOuS> yeah, I did... nobody there much
02:10am
-:- SignOff hcf: #TUNES (Leaving)
-:- hcf [nef@me-portland-us814.javanet.com] has joined #tunes
<eihrul> hcf: slate has a mlist? :)
<hcf> does now
<eihrul> off of tunes mlist page?
<hcf> yes the collab page
<hcf> listed last, above the irc blurb
-:- abi [nef@bespin.dhs.org] has joined #tunes
<eihrul> abi: wb
<abi> well, wb is welcome back
02:30am
* Fufie/#tunes wakes up
* eihrul/#tunes kicks Fufie and knocks him unconscious
-:- SignOff eihrul: #TUNES ([x]chat)
<Fufie> ouch
02:40am
-:- eihrul [lee@usr5-ppp58.lvdi.net] has joined #tunes
-:- SignOff eihrul: #TUNES ([x]chat)
-:- SignOff cbz: #TUNES (using sirc version 2.2+ssfe)
<hcf> abi: laml is Lisp Abstracted Markup Language at http://www.cs.auc.dk/~normark/laml/
03:00am
-:- SignOff hcf: #TUNES (Leaving)
<Fufie> hmm fare is linked from /.
03:10am
<Fufie> augh.. /. is bad
03:20am
<NeBulOuS> is it now?
<NeBulOuS> well... I need to reboot
-:- SignOff NeBulOuS: #TUNES (guess why :)
03:40am
-:- AlonzoTG [alangrimes@216-164-130-238.s619.tnt2.lnhva.md.dialup.rcn.com] has joined #tunes
-:- Kaufmann_ [kaufmann@jobim.fluid.impa.br] has joined #tunes
<Kaufmann_> Foo!
<Kaufmann_> anyone...?
<smkl> hmm
<Kaufmann_> hmm what?
<abi> hmm what is the url to the arrow's paper
<Kaufmann_> lol
<Fufie> who?
<Fufie> bah..
06:30am
<Kaufmann_> bah who?
<Fufie> nothing
06:40am
-:- hcf [nef@me-portland-us830.javanet.com] has joined #tunes
-:- hcf has changed the topic on channel #tunes to: TUNES, Free Reflective Computing System <http://www.tunes.org> || slate <http://www.tunes.org/~water/slate-home.html>
<Kaufmann_> ho
<Kaufmann_> if it isn't the mighty hcf :)
<hcf> mighty?
<Kaufmann_> dunno
<Kaufmann_> you're an op, aren'tcha?
<hcf> Kaufmann_: have u made web pages for _any_ of ur projects yet?
<Kaufmann_> hcf, not quite
<Kaufmann_> I really really hate HTML
<hcf> well
<hcf> u need not use html
<Kaufmann_> I know
<Kaufmann_> I guess I'm just lazy :)
<hcf> just put up a bunch of files
<hcf> hows that orthogonal ide coming?
<Kaufmann_> just fine
<Kaufmann_> I just have to finish writing the callbacks
07:10am
<Kaufmann_> which, of course, will take me forever :)
<Kaufmann_> eventually, I'll also write some code for compiling Orthogonal code to Perl, and then to perlcc
07:20am
-:- binEng [Anders@j141.ryd.student.liu.se] has joined #tunes
<Kaufmann_> *yawn*
<hcf> *hickup*
07:50am
<abi> *croak*
-:- SignOff abi: #TUNES (dying by hcf's request)
-:- abi [nef@bespin.dhs.org] has joined #tunes
<Kaufmann_> how amusing
-:- hcf [nef@me-portland-us830.javanet.com] has left #tunes []
-:- Downix [down@d-gnaps-170.ici.net] has joined #tunes
<Kaufmann_> hey Down
<Kaufmann_> so you work for Amiga Corp?
08:00am
<Downix> No
<Kaufmann_> hrrm
<Kaufmann_> weird
-:- thomas [thomas@193.217.63.152] has joined #tunes
* Downix/#tunes works for himself, why do you ask?
<Kaufmann_> I could have sworn it
<thomas> hi
<Kaufmann_> hello thomas
<Downix> Kaufmann_:  There's an easy way to tell I don't... I know what an Amiga is
<Kaufmann_> Downix, okie
-:- nef [nef@me-portland-us830.javanet.com] has joined #tunes
<Downix> Kaufmann_:  I find Amiga's better than SInclairs and Atari's tho
<Kaufmann_> what do you mean?
<Downix> Kaufmann_:  I collect unusual and innovative computer designs
<Kaufmann_> I mean, what did you mean by "I know what an Amiga is"?
<Downix> Kaufmann_:  The new Amiga Corp says and acts like the Amiga is 100% software
<Downix> ignoring the careful balance
<Kaufmann_> oic
<Downix> End result is no better than Windows
<thomas> end result for amiga in 89 was that of windows 9x
<thomas> but, sadly, not much has been done since
<Kaufmann_> Downix, wasn't it you that posted this? http://slashdot.org/comments.pl?sid=00/01/04/1357209&cid=144
<Downix> Kaufmann_:  What was that topic
<Downix> thomas:  Agreed.  Rather sad
<Kaufmann_> something about 1984
<Downix> Kaufmann_:  That was me
<Kaufmann_> for some reason, you used "we" there
<Downix> could you cut n paste the sentance?
<Kaufmann_> urgh, too late
<Downix> it's ojk
<Downix> I believe I was either referring to "we" as in the Amiga community or "we" as in my company
08:10am
<Downix> Anyways
<Downix> what did you think of my posting Kauf?
08:20am
-:- wilwauk [wilwauk@193.217.204.225] has joined #tunes
-:- wilwauk [wilwauk@193.217.204.225] has left #tunes []
<Kaufmann_> back
<Kaufmann_> anyway, I don't really care about Amigas :P
<Downix> well, do you see my view tho?
<Kaufmann_> yeah
<Downix> ok, that's all I asked
<Kaufmann_> Woo-hoo! BeOS R5 will be beer-free!
<Downix> 8)
<Downix> I noticed
* Fufie/#tunes thinks beer should be free _and_ gratis
<Fufie> coke as well
<Kaufmann_> Now I've got no excuse left not to install VirtualPC on my iWhack
<Kaufmann_> I'm going to make it a dual-boot... BeOS/FreeBSD
<Kaufmann_> and if I fsck it up (as I always do), it's just a matter of removing the HD image and doing a reinstall
-:- nef is now known as hcf
<hcf> abi: laml
<abi> rumour has it laml is Lisp Abstracted Markup Language at http://www.cs.auc.dk/~normark/laml/
-:- Jay1 [user5389@p14-nsv2-cardiff3.tch.dtn.ntl.com] has joined #tunes
-:- Jay1 [user5389@p14-nsv2-cardiff3.tch.dtn.ntl.com] has left #tunes []
09:00am
<Kaufmann_> This has really made my day!
<AlonzoTG> beer is free, just heavily regulated. (see appropriate section of the US federal code)
<Fufie> the recipe for the beer is usually not free? it's probably a trade-secret
<Fufie> just like dvd
<Fufie> and coke
<Kaufmann_> heh
<Kaufmann_> I can make beer. Problem is, it'll suck.
<Fufie> but if you open source the recipe
<Fufie> we can help you
<Fufie> fix bugs
<Fufie> update documentation
<Fufie> write howto
09:10am
<Fufie> patch the kernel
<Fufie> :)
<Kaufmann_> yeah right
<Kaufmann_> in two weeks, there'll be sixty seven mildly differing variants of the same recipe, all with crappy docs :)
<Downix> lol
<Kaufmann_> you can get a beer recipe anywhere. The catch is in making it properly
<Kaufmann_> what are the axes in that graph?
-:- SignOff Downix: #TUNES (BitchX: so real, you'll wet yourself!)
09:20am
<thomas> yepp, nomatter how good your recipe is, some sucker is likely to interpret it incorrectly
<Kaufmann_> urgh
<Fufie> kauf: you made one kind of beer.. I make it work with icecubes.. 
<Fufie> someone else adds support for mixing other stuff in
<Fufie> some people are concerned about alcohol level
<thomas> and microsoft sell pain killers :)
<Fufie> no.. microsoft seels a sour version of it but with a dazingly cool label and market it really well
-:- SignOff smkl: #TUNES (changing servers)
<Fufie> and soon everyone think bill invented beer
<Fufie> s/seels/sells/
09:30am
-:- Downix [down@d-gnaps-255.ici.net] has joined #tunes
<thomas> fufie, how is your thesis going?
<Fare> Gakuk
09:40am
<hcf> abi: tell Fare about laml
-:- SignOff Downix: #TUNES (Ping timeout for Downix[d-gnaps-255.ici.net])
<Kaufmann_> good grief, Fare
<Kaufmann_> how's the traffic on cybernethics, by the way? :)
<Fare> Grief?
<Fare> Kaufmann: aren't you subscribed?
<Fare> Kaufie: subscription more than doubled in one day.
<Kaufmann_> yeah, I'm subscribed
-:- Downix [down@d-gnaps-202.ici.net] has joined #tunes
<Kaufmann_> I just can't read it from work
09:50am
<Fare> not even through the web? all messages are archived.
-:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes
10:00am
-:- SignOff Kaufmann_: #TUNES (Read error to Kaufmann_[jobim.fluid.impa.br]: EOF from client)
-:- jdl [jiml@ultra1.inconnect.com] has joined #tunes
<jdl> Anybody awake?
<Fare> no
<Fare> rrr zzzzzz
<jdl> oh, okay then   ;)
<jdl> set beep_when_away 5
<Fare> what's that?
<jdl> set beep_when_away 1
<binEng> hello jim
<jdl> Hi bin
<binEng> long time ago I saw you here
<jdl> I don't come by very often....
10:20am
<jdl> But I recently lost my job so now I have time on my hands.  :)
<binEng> lost your job?? what happened?
<jdl> No biggie  :)  I'm a consultant.
<binEng> um, ok. What is you're doing?
<jdl> ?
<binEng> what's your.. speciality?
<jdl> Oh, web, OO, Java
<jdl> bin: Did you ever see my glossary response on the Prism list?  It never came back to me.
<binEng> yes, I saw it.
<binEng> but I find it hard to reply to
<jdl> why?
<Fare> glossary response?
<jdl> Fare: bin's working on a glossary
<binEng> jdl: code/metacode/metamodel/model.. it's a little confusing
<jdl> bin: Actually, yeah, I wanted to talk to you about that.  
<binEng> sure
<jdl> I'd like to figure out what your definition of metamodel is so we both know what
<jdl> the other person is talking about.  :)
<jdl> Fare: See the Prism list archive for the working docs
<jdl> bin: So what do you mean when you say "metamodel?"
<binEng> Meybe there's little point in going on on the mlist when we're the only participants
<jdl> bin: The lack of participation can be frustrating for sure.
<binEng> Here's the entry for it so far:
<binEng> : Metamodel :metamodels
<binEng> A metamodel is the "implementation" of an [idea], a representation of a [code]. It defines a [syntax] for some [semantics].
<binEng> Could also be seen as §format§ or §type§.
<binEng> Metamodels define sets of valid [models]. They are in fact models themselves and instances of a [meta-metamodel], [implicit] or [explicit].
<jdl> But as long as we're on the list, other people can listen and maybe eventually
<jdl> feel like participating.  Also, the list is archived, which is nice.  Everything stays together.
10:30am
<binEng> Btw, in case I sounded like I had only written what I showed in the "abstract/concrete" mail, I must say it's not that way
* jdl/#tunes is thinking about the definition
<binEng> (I have some 25K text)
<jdl> bin: no, I thought not.  You just pulled out a related chunk (or so it seemed)
<binEng> yes exactly
<jdl> bin: Did you get BOTH responses I sent?  First was "Re: Glossary" second was "Re: Glossary (abstract/concrete)"
<jdl> It was the second that I never got a copy from the list; I'm thinking of resending.
<binEng> Yes I got both. From the mlist too
<jdl> bin: okay, good.
* jdl/#tunes is still thinking about def
<jdl> Okay, here's my definition, tell me where we disagree:
<jdl> A *code* is the concept of how other concepts are encoded into models and decoded
<jdl> from models.  Given a model, we use a code to decode that model and determine
<jdl> what concept the model represents.
<jdl> (more)
<jdl> A *metamodel* is a model of a code.
<jdl> (that's it)
<binEng> hm
<binEng> I'd only say a code is a concept, without other restrictions than it is immaterial
<jdl> What do you mean "without other restrictions?"  I think I agree but I'm not sure why you brought it up
<jdl> .
<binEng> Well, you said "the concept of how other concepts are encoded...". That's a restriction on the definition.
<jdl> I still don't fully understand... so you're saying the definition should be
<jdl> A *code* is a concept.
<jdl> ?
<binEng> seems so. I'll check at that semiotics page
<jdl> The problem I have with that definition is that it doesn't add any value.  There's
10:40am
<jdl> no difference between code and concept.
* binEng/#tunes is pondering
<jdl> Here's the relevent portion of one semiotics tutorial:
<jdl> (But keep in mind that I adapted my terms from a number of sources, including semiotics,
<jdl> and that they won't match up exactly.)
<jdl> he meaning of a sign depends on the
<jdl>                   code within which it is situated; codes provide a framework within which signs make sense.
<binEng> You say that in order to decode a model, we need the code. But each "concrete concept" can be described by multiple metamodels/models. Therefore it seems to me that the code alone is unsufficient for decoding.
<jdl> (Oops, sorry for the formatting)
<jdl> I don't know what you mean by "concrete concept," but it's true that a concept can
<jdl> be described by multiple models.  A code IS sufficient for decoding a model, but in
<binEng> From the text you pasted from that tutorial it seems like code is something like I would call metamodel
<jdl> order to *properly* decode it you must use the proper code.
<binEng> (sorry, with 'concrete concept' I meant concrete stuff like a number, in contrast to the set of all numbers)
<jdl> yes.  And originally, I used "metamodel" instead of "code" which might be where some
* Downix/#tunes is totally lost now
<jdl> of the confusion is coming from.  But I split the term "metamodel" into both "metamodel" AND
<binEng> Downix: So are we, kinda :P
<Downix> lol
<jdl> "code" to emphasize the distinction between concept and model.  "Code" is the concept, "metamodel" is the model.
<jdl> Downix: :)
* Downix/#tunes keeps working on his register codes
<binEng> I agree a metamodel is a model of a code.
<jdl> okay
<binEng> I see a code as the meaning, while the metamodel defines a syntax that maps to those immaterial ideas
<jdl> Ah... okay.
<binEng> ofc, this is just my understanding of the term. might be wrong
* jdl/#tunes is thinking
10:50am
<jdl> Okay, before we talk about metamodel, do we agree on code?  How would you define it now?
<Downix> OH man!  miRC just got it's 10th registered user today!
<binEng> jdl: it's not my task to define 'code'.. I've told how I see it, but from the tutorial it seems like it means metamodel (what I see as metamodel)
<binEng> Downix: ?
<jdl> bin: I'm just trying to make sure that we understand each other
<Downix> binEng:  Someone just registered the 10th copy of mICQ
<Downix> Erm, mIRC
<binEng> but mIRC isn't directly new in the game?
<Downix> No, but only 10 people have registered it
<Downix> And the 10th happened today
<binEng> you sure? that sounds amazing
* Downix/#tunes rocks to his new MP3
<Downix> yes I'm sure
<Downix> check the micq homepage
<Downix> http://www.ircnews.com/mirc.html
-:- SignOff hcf: #TUNES (Ping timeout for hcf[me-portland-us830.javanet.com])
<binEng> is that a joke?
<jdl> bin: ?
<Downix> Nope
<Downix> I double checked
<Downix> only 10 people worldwide have ever registered miRC
<binEng> jdl: sorry, did I miss anything?
<binEng> jdl: I thought you were up to say something next
* jdl/#tunes grins... oops
11:00am
<jdl> bin: I wanted you to tell me what your understanding of "code" was so we're on the same page.
<binEng> I'm a little confused over this code thing now (as if I weren't before). It seems like I indeed has misinterpreted the meaning
<binEng> The tutorial uses the word not as the idea but the physical occurances
<jdl> bin: I'm pretty sure that your "metamodel" is similar to my "code."
<jdl> Maybe an example would help:
<binEng> But you changed from metamodel to code. How do you use metamodel now?
<jdl> The C language is a code.  The C language *specification* is a metamodel.
<jdl> I.e., "C language" is a concept we carry around in our head and use to decode C models.
<binEng> Can't we say the code is the concept, and the metamodel the mapping between physical representation and the code?
<jdl> The C language specification is a *model* of that code that allows us to learn the code.
<thomas> code is interpreted by metamodels - AND - metamodels are defined with code
<thomas> is there an example of code not representing a metamodel ?
<binEng> thomas: Aren't you using it backwards? The code is supposed to be the concept..
<jdl> thomas: ? It's the other way around - metamodels represent codes
<jdl> thomas: Maybe you knew this already, but "code" doesn't mean "source code"...
<binEng> not in this [context] at least ;)
<jdl> ;D
<binEng> Can't we say the code is the concept, and the metamodel the mapping between physical representation and the code?
<abi> i already had it that way, binEng.
<binEng> abi: shut up
<jdl> bin: I'm not sure what you meant by "mapping between physical representation and the code"
<jdl> :)
<jdl> It seems like you want to separate semantics and syntax so that code<->semantics and metamodel<->syntax ..?
<binEng> something like that, yes
<jdl> Okay
* jdl/#tunes is thinking
<Downix> k
<binEng> But as I said, that seems not be the tutorials take on it
<binEng> btw, I don't know what source you refer to as the tutorial, I look at 'Semiotics for Beginners'
11:10am
<jdl> bin: yep... my approach has been that the *code* maps between physical representation (syntax) and meaning.  I.e., the "ANSI C" code tells you that this model: "printf("Hello World");" equals this concept: "Display a greeting".
<jdl> bin: yep, that's what I meant.  I'm sure there's others, but that's what I've been using.
* jdl/#tunes knows very little about semiotics
<binEng> but it can't be that way, if several metamodels can describe the same code.
<binEng> me neither
<jdl> bin: why not?  Let's say you have the "C Language specification" in English, German, and Swedish.  Wouldn't each of those be a different metamodel (because they are all different) that describe the same code?
<binEng> I'd say the specification is a metamodel that tells you how to interpret printf(...) as belonging to the C code/"paradigm"
<binEng> Would they be different metamodels? Why, the models that can be created with them are exactly the same, so I don't see it that way.
<jdl> Exactly, that's what a metamodel is.  It describes a code... after reading the specification (= decoding the metamodel) you understand the code.  Thus you can decode C models.
<binEng> For a real difference between metamodels, take integers which you yourself examplified under different metamodels
<jdl> bin: They're different metamodels because their contents are different.  The concept they represent is the same -- ANSI C.
<binEng> yes yes, I agree on what you say, but obviously we're missing the point, as our views differ
<binEng> " Let's say you have the "C Language specification" in English, German, and Swedish.  Wouldn't each of those be a different metamodel (because they are all different) that describe the same code?" <-- To this I say no.
<jdl> Which part do you say no to?  They aren't different metamodels?  Or they don't describe the same code?
<binEng> And then I argue that those "different" metamodels are not different, as they use the same set of models. In difference to those integer metamodels.
<binEng> Different metamodels (they're not, that is)
<binEng> (Sorry if this get repetetive or so..)
<jdl> Okay, let's go back to the integer example.  You agree that we can have one concept -- the number 42 -- represented by different models?
<binEng> Indeed.
<jdl> (No biggie... these concepts are core to Prism, and the more we discuss the easier it will be to explain to others.  Also, I'm not saying my way is the only way.)
11:20am
<binEng> And different metamodels must be used to decode those models.
<jdl> So, for example, we could model '42' in English, German, and Swedish.  The models would be different but the concept would be the same.
<binEng> wait a sec, depends on *how* you would model them. The digits are the same, for example..
<jdl> Let's say we use words  :)  "fourty-two," etc.
<binEng> ..but not if you spell the number out
<binEng> yes
<binEng> I agree.
<jdl> Okay, by analogy then... (keep in mind my definition of metamodel is different from yours):
<binEng> fyrtiotvå, zwei und vierzig :)
<jdl> 42:fourty-two::C:C spec
<Fufie> førtito
* Fufie/#tunes smiles.  *SMILE*
<binEng> thank you Fufie
<jdl> Just as we can model '42' in multiple human languages, we can *metamodel* C in multiple human languages.
<binEng> what did you mean with that 42:four... line?
<Fufie> hi jdl :)
<jdl> Hi Fufie... Stig, right?
<Fufie> yup
<jdl> bin: not important... it ended up being confusing.  "42 (the concept) is to "fourty-two" (the model) as C (the code) is to C specification (the metamodel).
<binEng> jdl: um, yes I think I could agree with that
<binEng> jdl: But I don't see how metamodeling C in multiple human languages would be or differ.
-:- iepos [root@d44.k1-2.tecinfo.com] has joined #TUNES
<binEng> iepos: Hello. How're your web pages doing?
<iepos> heh heh
<iepos> oh
<iepos> i've been working on them
<iepos> a bit
<iepos> hmm ...
<iepos> maybe i should ftp the new version
<binEng> :)
<jdl> bin: C wouldn't change, of course, just the way you described it.  Like I said in a previous email, metamodels are kind of like a bridge between codes.  If you know Swedish, then you need a Swedish C spec in order to understand C.  If you speak
<jdl> English, then you need an English C spec.  Etc.
* Fufie/#tunes is slightly away.. reading logs on meta*
<iepos> binEng, justa sec...
<binEng> heh, imagine metalogs ;)
* jdl/#tunes can't even begin to imagine what that would be  :)  A log of log file activity?  :)
<jdl> Or a metaglossary :)
<binEng> yes, maybe keeping track of several logs files or something
<binEng> yik
<binEng> "What /is/ a glossary" :)
<jdl> lol
11:30am
<thomas> jdl, I thought metamodel represented the _way_ code is interpreted ( syntax to semantics mapping ) while the code just represents itself - code , doing something or describing something - what I like to call a 'thing' :)
<thomas> the important thing is that a metamodel is _also_ a 'thing' .
<binEng> thomas: Are you on the Prism mlist? (just wondering)
-:- SignOff Downix: #TUNES ([BX] The name's X. Bitch X.)
<thomas> bineng, nope
<iepos> binEng, the ftp'd the new stuff ... it's at tunes.org/~iepos/new
<binEng> iepos: ok
<Fufie> jim: from what I understand your project seems to be a very brave attempt at something which is almost an AI-complete problem..
<jdl> t: Well, when you say "code" do you mean "source code" (printf, etc.)?  Because when we say "code" (right now) we actually means something like "language".
<thomas> defining things this way, ends in endless recursion, but we don't care and call it bootstrapping
<binEng> Fufie: I don't see it that heroic..
<binEng> or code == convention
<jdl> Fufie: It could be, but I'm taking a lot of short cuts, so as bin said, it's not nearly that heroic.
<Fufie> bin: unifying languages in a meta-system is an ai-complete problem
<thomas> fufie, why?
<binEng> Fufie: do you have a definition on "ai-complete"?
* binEng/#tunes also wonder why
<thomas> binEng: easy for us - difficult as hell for computers !
<binEng> :P
<binEng> and what in the whole world is a metasystem?
<jdl> bin: The whole world, of course!  ;)
* binEng/#tunes is taking notes
<binEng> I didn't have 'metasystem'. Weell?
* jdl/#tunes doesn't actually know what fufie meant, but can never pass up the opportunity for smart-ass remarks.  ;)
<Fufie> jim: I still think it's heroic if I have gotten the docs right, attempting the full problem is foolish
* binEng/#tunes kicks jdl's smart ass ;)
<Fufie> bineng: the jargon file defines ai-complete I think
<thomas> fufie, do you consider the terms 'ia-compliete' and 'np-complete' related ?
<binEng> Fufie: you're right
<Fufie> jim: languages have subtle nuances which are next to impossible to capture in a spec or formal languuage.. you cannot translate e.g a C language spec from english to swedish without losing some info and you will probably gain some too.. despite swedish not being as sophisticated as norwegian
<binEng> I'd be happy to find out what np-complete is also
<Fufie> thomas: see the jargon file..
<Fufie> thomas: ai-complete is a loose term though
<thomas> though?
<binEng> Fufie: Actually, it is quite the contrary. But, we didn't really expect *you* to know ;)
<Fufie> jim: the nuances also exist in programming language and by going to model language you lose lots of info
<jdl> fufie: Ah, okay.  The big shortcut I'm taking is not requiring formal metamodels (specifications).  That's why it's not that heroic.  :)
<jdl> fufie: My metamodels are defined in natural language, not in a formal language (that's what I meant by not requireing formal metamodels).
<Fufie> thomas: 'though' as in comparison to 'np-complete' which is more well-defined in use
<thomas> jdl, it's a top-down approach ! cool
<jdl> thomas: ? You lost me
11:40am
* jdl/#tunes goes and looks up ai-complete in the jargon file
<thomas> jdl, instead of building upwards ( to higher abstraction layers )  from a formally defined fundament, you go from top to bottom.
<Fufie> jim: a lot of the problem encountered with CSF is that it is a trade-off between completeness and usefulness
<Fufie> jim: I expect that you will also encounter this for your meta-languages
<thomas> fufie, I suspect the same
<jdl> AI-Complete: ...ndicate that the solution presupposes a solution to the `strong AI problem' (that is, the synthesis of a human-level
<jdl> intelligence). A problem that is AI-complete is, in other words, just too hard. 
<binEng> I don't see any such problems here..?
<thomas> jdl, so you'll have objects which are described in, for example english?
<thomas> what is the strong AI problem?
<thomas> categorization ?
<binEng> jdl: Let's continue
<jdl> fufie: I think the reason you encountered that problem is because CSF attempts to capture information from several different languages at once.  That's not necessarily the case for Prism... I'm not trying to come up with one format that captures
<Fufie> bineng: building a complete framework atop natural languages and programming languages for converting between them is ai-complete
<jdl> every language's semantics... rather, I've come up with a common way of defining *multiple* formats in a standardized way
<binEng> jdl: Should we postpone the metamodel vs code discussion for now?
<jdl> thus allowing one format per language (or more, or less, whatever you want).
<jdl> bin: oh, I thought we had come to an agreement.
<Fufie> jim: ok, then I might have misunderstood.. how is this different e.g from SGML then, which is also a meta-format?
<binEng> jdl: I'm afraid not. But there's other stuff I'd like to ask
>>> binEng [Anders@j141.ryd.student.liu.se] requested PING 948224844 121926458 from #tunes
<jdl> fufie: In my terminology, SGML is a "metacode".  (So is XML.)  The answer is that it's not different -- Prism also revolves around a metacode.  The difference is in that Prism is an abstract metacode designed specifically to focus on semantics
<jdl> rather than syntax.
<Fufie> ahh
<jdl> bin: Okay, fire away
<binEng> jdl: Ah yes, 'metacode' is another word you need to explain. Is it necessary?
<jdl> thomas: check out my web site -- http://www.teleport.com/~sphere
11:50am
<jdl> bin: I'm not sure if I can explain metacode because it's based on my definition of code and metamodel, which we haven't quite resolved yet.
<binEng> jdl: OK, next issue then. Reification. I disagree with your definition, see http://www.tunes.org/papers/Glossary/index.html#reification
<jdl> bin: But... concepts are represented by models.  Models are interpreted with codes.  Codes are concepts, represented by metamodels.  Metamodels are interpreted with codes called "metacodes."  Metacodes are concepts, represented by metamodels.
<jdl> bin: That definition came straight out of my dictionary.
<binEng> "Models are interpreted with codes" <-- No, this is what you have metamodels for.
<thomas> jdl, there is no goal in itself to create alot of jargon and split things into many cases. Why not 'things' ?
<jdl> bin: That's where we disagree.  I use "code" where you use "metamodel."
<binEng> jdl: This is what the link says: "Reification is an operation by which something that was previously implicit, unexpressed and possibly unexpressible is explicitly formulated and made available to conceptual (logical or computational) manipulation."
* Fufie/#tunes gives thomas a bop over the head :)
<binEng> jdl: Seems so. But then we at least know where we disagree
<thomas> you disagree because the spitting into code and metamodel is "incorrect"
<jdl> thomas: Sorry, I'm confused again.
<thomas> fufie, thanks :)
<binEng> If you have a model, it is useless even if you have the code, unless you know the canonical metamodel with which to decode it into that code (see example on integers)
<Fufie> the word I dislike the most is 'metamodel' because a model is already meta..  (meta)^2  might do the trick :) (very serious)
<thomas> jdl, why do you need Codes, concepts, metamodels, codes AND metacodes
-:- Ghyll [karltk@mp-217-229-140.daxnet.no] has joined #tunes
<binEng> Fufie: a model isn't "meta" automatically, is it? Say, "42". Nothing meta there.
<Fufie> hi ghyll
<Ghyll> hi fuf(f)ie
<binEng> thomas: oh you have heard nothing yet.. seriously
<binEng> haha, fuffie
<Fufie> bineng: in normal scientific jargon, a model is a meta-explanation of a concept, preferrably seen in real life
<binEng> Fufie: well we screw up normal jargon here. A model is something concrete, a "thing" as thomas wants to put it
<jdl> thomas: normally you just need codes, concepts, and models.  But when you talk about those things, you need metacodes and metamodels.
<thomas> jdl, NO
<thomas> (sorry)
<Fufie> bineng: obviously, because in the jargon I am used to, a model is _not_ concrete.. it's meta.. 
<binEng> Fufie: sorry for that
<binEng> Maybe we should use something else instead of 'model' here?
<thomas> a model is some thing concrete or not describing the meaning of relations between (possibly) some other things.
<binEng> Or does semiotics use that word?
12:00pm
<jdl> bin: No, I got model out of my dictionary -- not the scientific meaning of the word, just the everday meaning:
<jdl> A thing that represents another thing.
<thomas> jdl, you just need a way to express what you mean, implemented by, as you say, a model/metamode/interpreter/vm
<binEng> This is what I say in the glossary:
<binEng> : Model :models
<binEng> A model is a representation of something. This something can be a single trivial [concept] such as the number 42, as well as a set of other models, [metamodels].
<binEng> Models are associated with [syntax].
<Fufie> jim: your dictionary implicitly says it is meta
<jdl> bin's definition is right on target.  That's what I mean when I say "model."
<jdl> fufie: Well, if 'model' is meta, then metamodel is meta-meta.  :)  But (meta)^2-model would be meta-meta-meta.  ;)
<thomas> fufie, everything is meta. That something is meta to you, is dependent on how you view it. 
<binEng> context!
<thomas> context!
<Fufie> jim: yes.. get rid of the meta-word.. it's hyped up and overused :)
<jdl> context!
* jdl/#tunes jumps on the bandwagon (not exactly sure why, but happy to be a sheep)  :D
<binEng> :P
<jdl> lol
<binEng> Well as long as I were first
* thomas/#tunes guess he thinks it's funny too
<jdl> thomas: sorry, didn't meant that as anything bad towards you :)
<thomas> no offence taken - whatsoever :)
<binEng> Beck to the topic, gentlemen
<jdl> fufie: "meta" may be overused, but I think it has a fairly-well understood meaning that's appropriate in this case.
<Fufie> jim: even when we have meta-model which is (meta)^3 ;)
<Fufie> or was it 2
<Fufie> impossible to keep track of
<binEng> When you're not around to crew up, it's clear enough for me
<Fufie> and metaprograms, metacodes, metalanguages, metacompilers, meta object protocols and meta-somethings
<binEng>  s/crew/screw/
<Fufie> meta loses meaning when too much has a meta-prefix
<jdl> fufie: Okay, so you're saying a "model" (as we use the term) is the same as a "meta-concept" -- the concept of a concept?  Hmm.  I don't think I agree.
<Fufie> bineng: :p
<thomas> fufie, good meta-observation !
<jdl> metaprogram: a program that operates on programs.
<jdl> metacode: a code that describes codes.
<thomas> input: program
<thomas> metaprogram: program
<jdl> As for the rest, I don't use them.
<jdl> (Well, I used "metacompiler" once but I shouldn't have.  Bad Jim!)
<binEng> hehe
<binEng> now, here's what i write about reification:
<Fufie> I think you use metalanguages as well
<binEng> : Model :models
<binEng> A model is a representation of something. This something can be a single trivial [concept] such as the number 42, as well as a set of other models, [metamodels].
<binEng> Models are associated with [syntax].
-:- jdl is now known as meta-jdl
* meta-jdl/#tunes scolds jdl
-:- meta-jdl is now known as jdl
-:- thomas is now known as meta
* jdl/#tunes is chastened
<jdl> ;)
12:10pm
<meta> hehehe
<meta> I can be anyone!
<meta> but still, allways, at evey time - I am just me :
<meta> )
-:- meta is now known as thomas
<Fare> meta: you're no more the "me" you used to be.
* jdl/#tunes has gotten lost
<jdl> Where were we?
* binEng/#tunes notices he pasted the wrong glossary entry :P
<iepos> fare, gakuk!
* jdl/#tunes was wondering why that was posted
<AlonzoTG> are metapersons unique?
<binEng> reification!
<binEng> : Reification :reify
<binEng> For computers to [understand] [information] containing [implicit] parts require them to acquire necessary parts of human [context] and implicit [knowledge]. The more the information can be placed in an [explicit] context for the computer, the more it can make out of it. The process of doing this is §reification§, something very important for [flexibility] and [automation].
<binEng> From [Tunes]: "Reification is an operation by which something that was previously implicit, unexpressed and possibly unexpressible is explicitly formulated and made available to conceptual (logical or computational) manipulation."
<AlonzoTG> are metapersons unique?
<thomas> they, may not be unique persons, but they are unique things
<Fufie> are metapersons the same as cannibals, ie they take persons as input? :)
<thomas> :D
<binEng> good point
<jdl> From my dictionary: "To regard or treat (an abstraction) as if it had concrete or material existence."
<Fare> Fufie: cannibals are not devoid of output (although depending on the point of view, you may consider that a side-effect)
<jdl> lol
<Fufie> fare: and cannibals which eat cannibals are meta-meta-persons? :)
<jdl> bin: I don't see how the dictionary definition relates to the two definitions you posted.  This isn't a tech dictionary, though, so maybe the definition is incomplete.
<binEng> jdl: In a way your definition could be right.. but it's pretty far-fetched
<Fufie> jim: would it possible if you built stuff from scientific terms which should be more familiar?
<jdl> bin: I have no opinion on the word "reify" but I'm not inclined to think of the dictionary definition as "far-fetched"  :)
<binEng> far-fetched in this context
<binEng> context!
* thomas/#tunes is not going down that lane again !
<binEng> argh
<jdl> Fufie: I use the terms I know... if you have terms which are more appropriate, I'm definitely happy to use those instead.  (Seriously)
<thomas> binEng, sorry
<binEng> jdl: I'll go with the definition I have for 'reification' then
<jdl> binEng: What is the context for "reify," then?  The only person I know who uses that term is water.
<binEng> jdl: well, I guess the dictionary did not specifically have computers in mind..
12:20pm
<jdl> bin, fufie: I do agree that it's important not to change the meaning of existing terms, particularly if they are established in CS.
<Fufie> at least as I know the term model, it is a (simplified) explanation or a formula for a more complex system
<Fufie> ie it is meta already, describing something more complex
<jdl> Fufie: Aka a representation of that system.  (Almost always simplified, but not necessarily so.)
<thomas> fufie, almost all "models" are simplified, but in CS, it's important and possible to make the model complete
<jdl> Fufie: I disagree that 'meta' means describing something more complex.
<Fufie> if it isn't simplified it is the system itself
<Fufie> jim: no.. it is not that alone, but it also has that meaning
<jdl> fufie: No, A CAD drawing is a model of an actual thing, not simplified, but still not the system itself.
<Fufie> jim: a CAD drawing is a model of a more complex actual thing
<jdl> fufie: Not necessarily -- you could have a CAD drawing of a nail.  Is the actual nail more complex?
<Fufie> if the system described with a model isn't more complex you have the system itself
<Fufie> the actual nail is more complex, yes
<thomas> fufie, mabe we want the system itself
<Fufie> thomas: yes, but then you don't call it a model, do you?
<jdl> fufie: How is it more complex?
<binEng> jdl: About the syntax in the glossary... I did have /italics/ at first, but that created some problems. How about \italics\?
<Fufie> jim: a nail usually is a complex object consisting of various forms of metal melted together and the molecules are in a certain structure
<jdl> bin: Sure - as long as I can type it on my keyboard I don't care.  :)
<Fufie> jim: the CAD drawing is a simplification of the actual nail used
<binEng> fine then, I'll change that
<thomas> fufie, how about matematical models?
<Fufie> thomas: mathematics often describes simplified physical models.. 
<jdl> fufie: okay, I'll grant that anything in "real life" is more complex than its model because we can't model the complexity of the universe.  But why does it matter?
<Fufie> thomas: you know from physics that the models used aren't the same as real life
<thomas> fufie, who cares, in CS, matematics alone are important
<Fufie> jim: because model is a term used that way already.. 
<Fufie> thomas: the mathetical models in CS is still simplified.. the calculi are very limited
<binEng> jdl: Another thing, do you think I should continue show a few entries a time, or should I "publish" it all as it is? Maybe the whole thing might be a little overwhelming?
<jdl> fufie: how is your usage different than mine?  I don't attempt to model electrical behavior of the computer, so my models are just as simplified... I've lost track of the argument.
<thomas> fufie, simplified compared to what?
<Fufie> thomas: to the practical/dark side of CS
12:30pm
<Fufie> thomas: or do you usually try to prove your programs or use mathemathical models.. no you do the real thing..
<jdl> bin: Put the whole thing up on a web page, then solicit feedback for specific entries on the list.
<thomas> fufie, the side who's problems we don't discuss - right!, I forgot
<jdl> bin: (Sadly, you're not likely to get much response no matter how you do it :( )
<binEng> k
<binEng> I'll put it up now and see what you think
<Fufie> jim: by using e.g 42 as an example which for all practical purposes left being a model several thousand years ago..
<Fufie> jim: 42 is about as atomary as you get
<Fufie> 42 is 42 is 42
<thomas> the whole 42 and nothing but the 42 - complete and consistent
<jdl> fufie: Okay, I need a way to distinguish between the *concept* of the number 42 and the multitude of possible representations.  I've been calling those representations "models."  What word would you recommend I use instead?
<Fufie> representation
<binEng> http://www.tunes.org/~bineng/glossary.html
<jdl> (Keep in mind that it has to work for a representation of any concept, not just 42.)
<binEng> No index yet, though
<Fare> Fufie: no, "42 is 42" is not 42
* Fufie/#tunes gives fare a fwap
<jdl> AND -- "42" is not 42!  (Seriously.)
* Fare/#Tunes goes 3053
<jdl> "42" is a sequence of two ASCII characters *denoting* the concept of the number fourty-two.
<binEng> according to a metamodel :)
<jdl> bin: No, according to a code :)
<thomas> a code representing a model
<binEng> jdl: but that code can't describe every way 42 can be encoded, eh?
<jdl> bin: no, it describes one way of encoding '42' (and similar concepts).  Another code describes 'fourty-two' and another one describes 'fiertiotva,' etc.
<binEng> oh so that's the way you see it. So the concept of 42 is something else altogether then?
12:40pm
<jdl> bin: yes, I think.
<jdl> (If I understand you correctly.)
<jdl> fufie: I think I'm going to keep using "model" for now.  It's slightly different than the mathematical/scientific definition, but it is consistent with it, and it's also consistent with the commonly-accepted definition and the definition used in
<jdl> software engineering.
<Ghyll> jdl: emfbi. I assume you define your term 'model' there anyhow ?
<jdl> g: yes, in my essay "The Will and the Word."  I don't have a glossary yet - that's what we're working on.  :)  (bin is coordinating it.)
* Fufie/#tunes goes into Fuffie-mode.. 
-:- Fufie is now known as Fuffie
<binEng> Fuffie: what's that?
* jdl/#tunes wonders what 'emfbi' means
<binEng> jdl: maybe you'd like to look at it then.. http://www.tunes.org/~bineng/glossary.html as I said
<Ghyll> jdl: 'excuse me for butting in'
<jdl> g: oh.  no problem!
* jdl/#tunes is looking at the glossary
<Fuffie> bin: it's a mode where one is somewhat frustrated... I also go into Fuffie mode when people try to kill each other over when the millennium starts and stops
<Ghyll> Fuffie: all the time.
<Ghyll> Fuffie: any millenium moves forward in leaps and bounds.. 
* Ghyll/#tunes covers his head..
<Fuffie> bin: it is very useful when people try to be pedants over fictional stuff
* Fuffie/#tunes ignores Ghyll
12:50pm
* Fuffie/#tunes goes and grabs some food
<jdl> bin: still here?
<binEng> I'm here
<jdl> I looked at the glossary -- nice layout, by the way, and I also like the idea of modeling authors and related projects so you can view subsets of the glossary.
<jdl> Two comments:
<binEng> As I mention in the example at the bottom?
<Fuffie> bin: url?
<jdl> yes
<binEng> ok, it's not really finished..
<binEng> (the example, that is)
<binEng> Fuffie: http://www.tunes.org/~bineng/glossary.html
01:00pm
<jdl> 1) Some entries are pretty subjective.  "ultimate goals" comes to mind.  The subjectivity/comments added in to Fare's glossary is its biggest shortcoming, IMHO -- I'd like to avoid that.  Perhaps we could model "definitions" separately fro
<jdl> from "comments?"  (But have both be associated to a single entry, with subsets with and without comments?)
<Fuffie> bin: as for artifical code you do have genetic programming which is evolved but still readable and at least machine-readable
<jdl> 2) Our disagreement on code vs. metamodel is pretty pervasive throughout the glossary, causing me to disagree with several of the entries.
<Ghyll> see you later.
-:- SignOff Ghyll: #TUNES (foo)
<binEng> I do want to keep the entries objective, although reflecting the subject
<binEng> but yes, I try to show something of the philosophy
<jdl> bin: regarding "code" -- perhaps "code" is a bad term.  It tends to get confused with source code.  Perhaps I should use "language" instead of "code?"
<binEng> On the 2nd comment, yes that's right and is a problem
<binEng> perhaps, yes
<jdl> brb
<jdl> back
<binEng> wb
<jdl> bin: Okay, regarding code/metamodel: Do you not understand my definition, or do you not agree with my definition?
<binEng> neither, I guess
<binEng> what do you see metamodel as?
<jdl> An explanation (model) of a code.
<binEng> uh??
01:10pm
<jdl> Okay, not explanation -- description.  A model of a code.
<jdl> What about that doesn't make sense?
<AlonzoTG> om
<binEng> but a code is already.. how should I put it..
<jdl> Perhaps another point of view...?  Fufie, care to define it?  (Even if I disagree with your definition, it will help me see where I'm explaining poorly.)
<binEng> but a code is already a concrete description of a concept
<binEng> I'd appreciate if you could point out subjective parts in the glossary..
<jdl> How is a code a concrete description of a concept?  It's like a language -- it's a concept that you use to encode/decode models.
<jdl> bin: okay, but let's resolve this code/metamodel thing first.
<binEng> sure
<jdl> (Then I'll go through and make comments entry by entry.)
<binEng> very kind of you
<jdl> np.  So, how is a code a concrete description of a concept?
01:20pm
<binEng> if it describes how 42 is written as "42"..
<binEng> (or fyrtiotvå)
<AlonzoTG> eye don't think it iz.
<jdl> bin: Well, the code doesn't "describe" how 42 is written as "fyrtiotva" -- rather, it "knows" that 42 is written as fyrtiotva (and vice versa).
<binEng> if it's a *code* that's needed to decode a model
<jdl> In other words, if you know the code, you know 42<->fyrtiotva.
<binEng> :P
<jdl> ?
<binEng> So a metamodel would define what models are legal, but a code knows what models map to what concepts?
<jdl> Metamodels don't enter into it yet.  A code knows what models map to what concepts, and in so knowing, defines what models are legal within that code.  (I.e., in the Swedish code, "feriotva" ;) maps to no concept (I hope) so it's not a legal model.
* AlonzoTG/#tunes deteckts a blik.
<binEng> So what is a metamodel?
<jdl> Well, first, are we in agreement on "code?"
* AlonzoTG/#tunes is nihilistick and doesn't believe in the eggzistance of knocepts.
* binEng/#tunes nullifies AlonzoTG
<jdl> What a concept!
<binEng> fight implicity
<jdl> ?
<binEng> concepts exist, it's just a matter of observing them. reification :)
<binEng> jdl: I know your definition of code, yes
01:30pm
<jdl> bin: Okay, so by my definition, a code is a concept, right?
<binEng> no
<jdl> why not?
<binEng> "42" and "fortytwo" are different codes
<binEng> but same concepts
<jdl> Right, but the code isn't the concept of 42.  It's the concept of how models made up of 1234567890 (for example) map to integers.
<AlonzoTG> bineng: you are delusional.
<AlonzoTG> Model a toothpick.
<binEng> AlonzoTG: sure, in what respect?
<jdl> Also, "42" and "fourty-two" aren't actually different codes... they're different models, representing the same concept, *encoded* with different codes.
<AlonzoTG> any way you like. you can pick any symbol in the set of real numbers.
<jdl> atg: "toothpick"
<jdl> atg: "------------------"
<AlonzoTG> =)
<AlonzoTG> I like that second one.
<binEng> AlonzoTG: yourself
* AlonzoTG/#tunes hurls a laser guided flaming woodchuck at binEng
<jdl> atg: Anything can be a model of anything, given the right code.
<AlonzoTG> =\
<binEng> jdl: look, different codes or encoded with different codes - that's splitting hairs
<jdl> bin: I knew you probably meant "encoded with different codes" not "different codes" -- that's why I originally agreeed.  But I wanted to make sure we were on the same page, because the two phrases mean different things.
-:- SignOff thomas: #TUNES (Read error to thomas[193.217.63.152]: EOF from client)
<jdl> bin: still there?
<binEng> yes I'm here
<binEng> what do you want me to say..
<jdl> We were discussing whether codes were concepts.
<binEng> yes. Do you agree that under your definition, there are different codes for "42" and "fortytwo"?
<binEng> And can you also agree that they stand for the same concept?
01:40pm
<jdl> Yes.  "42" is the mathematical code, "fourty-two" is the English code.  And yes, those models both represent the concept of the integer 42.
<binEng> Two codes, one concept; codes!=concepts, QED
<jdl> (Note: The English code includes "thirteen" and "hello" and "flower" and "toothpick" as well as "fourty-two".
<binEng> so you're saying numbers are a subset of English? Come on
<jdl> The flaw in your logic above is that codes do not represent the concept '42.'  They represent the concept 'fourty-two is a model of 42'.  (along with lots more mappings.)
<binEng> yes! but that's not what you said
<jdl> No, numbers aren't a subset of English -- what I'm saying is that English includes mappings of models to concepts for much more than just 'fourty-two<-->42'.
<binEng> I don't see where we're trying to get here
<jdl> The point is that codes are concepts... the concept of how a set of models maps to a set of concepts.
<binEng> hair-splitting!!
<binEng> enough of it
<jdl> No... it's very important.  You see, if you agree that codes are concepts, then:
<jdl> metamodels are models of codes.
<binEng> I still think it's hairsplitting, as the same implementations of one code on different platforms could result in different metamodels with your overly strict definition
<binEng> the implementation details are not important
01:50pm
* Fuffie/#tunes ponders on Fufffie mode
<jdl> Exactly!  One code can have multiple, different metamodels!  Thus the metamodel can change independently of the code, provided it describes the same code.  This allows metamodels to be improved and changed even when codes do not.
<binEng> the models doesn't change, the code doesn't change - I see absolutely no reason to differentiate different implementations that fulfills that
<jdl> For example, one metamodel might be written in natural language, and thus only be useful to humans.  But another might be written in a machine-readable form, and thus be useful to computers!  Same code, different metamodels.
<binEng> whatever!!!!!!!!!!!!!!
<binEng> I can't take this anymore
<Fuffie> bin: Fufffie mode is still available :)
-:- binEng is now known as Fufffie
* jdl/#tunes shrugs
* Fufffie/#tunes needs some food
-:- Fufffie is now known as binEng
-:- SignOff jdl: #TUNES (Leaving)
-:- Fuffie is now known as Fufie
<binEng> now what
* Fufie/#tunes shrugs.. many new terms and concepts in one day
<binEng> why did he leave
<Fufie> but there'll come new days when the intention and meaning is easier to grasp
<Fufie> I don't know why he left.. hopefully it wasn't something I said.. I never meant to be rude with the Fuf*ie-modes
<Fufie> I got somewhat confused with the terms and the pedantic picking on my statements about 42
<binEng> oh well, how long was he here, 3-4 hours. That's some time.. but I needed to eat already before that :P
02:00pm
* binEng/#tunes is away: food
-:- NetSplit: irc.linux.com split from lewis.openprojects.net [02:05pm]
-:- BitchX+Deb1an: Press Ctrl-F to see who left Ctrl-E to change to [irc.linux.com]
-:- Netjoined: irc.linux.com lewis.openprojects.net
-:- Fare [fare@quatramaran.ens.fr] has joined #tunes
-:- Fufie [stig@tunnel-44-22.vpn.uib.no] has joined #tunes
-:- AlonzoTG [alangrimes@216-164-130-238.s619.tnt2.lnhva.md.dialup.rcn.com] has joined #tunes
-:- binEng [Anders@j141.ryd.student.liu.se] has joined #tunes
-:- iepos [root@d44.k1-2.tecinfo.com] has joined #tunes
-:- smkl [sami@glubimox.yok.utu.fi] has joined #tunes
-:- jdl [jiml@ultra1.inconnect.com] has joined #tunes
<jdl> Sorry for leaving abruptly like that... nothing personal, I was just getting frustrated.  The conversation seemed to be going in circles.
<AlonzoTG> =)
* jdl/#tunes is away (msg me to get my attention)
02:20pm
-:- lispbliss [lispbliss@cont01p06.ont.micron.net] has joined #tunes
<iepos> hello, lispbliss.
<lispbliss> hi
<lispbliss> anyone here knowledgable about mach?
<lispbliss> (the kernel not the speed)
<iepos> not me
-:- SignOff lispbliss: #TUNES (Ping timeout for lispbliss[cont01p06.ont.micron.net])
02:30pm
-:- SignOff iepos: #TUNES ([BX] I'm out like a light...)
-:- lispbliss [lispbliss@cont01p62.ont.micron.net] has joined #tunes
* Fufie/#tunes sighs..
* AlonzoTG/#tunes left clicks on lispbliss
-:- abi [nef@bespin.dhs.org] has joined #tunes
-:- pyro [tcn@cci-209150250077.clarityconnect.net] has joined #tunes
<lispbliss> left clicks?
<lispbliss> heh
02:50pm
<Fare> Gakuk
* Fare/#Tunes unclicks on lonzoTG
-:- eihrul [lee@usr5-ppp95.lvdi.net] has joined #tunes
03:00pm
<Fufie> (def-timeconsuming 'bootstrapping)
<AlonzoTG> om
<AlonzoTG> yeah I need to do that to my OS =\
<eihrul> Fufie: what are you 'bootstrapping?
<Fufie> my hypermedia-system
<Fufie> moving hard-coded objects out into the DSL 
03:10pm
-:- SignOff lispbliss: #TUNES ()
<Fare> DSL?
<abi> well, DSL is domain-specific language or digital subscriber line
<Fufie> I have a lisp-based dsl tailoured for making pluggable objects with distributed linking capabilities, database-support, indexing for search-engines, presentation to xml and html, etc.
<Fufie> and to prove that the dsl is not a toy language.. it has to be bootstrapped
<Fare> is the dsl embedded in lisp (with ,(forms) and stuff)
<Fare> ?
<Fufie> yes
-:- air [brand@p0wer.qzx.com] has joined #tunes
03:20pm
<Fufie> which means I don't need to add all the trivial base-functionality for those that need extra power
<Fufie> but can concentrate on adding the features needed
<Fufie> so I have e.g  (with-db (db) body)  which is weaved with the necessary db aspect code
* Fare/#Tunes is making a backup of bespin
<air> Fare: why would u wanna do that? :)
* binEng/#tunes is back
<binEng> jdl!
<Fare> air: It's my prayer to Baah-Kup, the God of Data Recovery
<Fare> aren't you a believer?
<Fare> when I pray bespin-backup-20000118.tar.bz2
<Fufie> I thought Bru and Tar were the only true Gods of Data Recovery
<binEng> Fare: Is that a backup of the whole site?
<binEng> jdl: Are you there?
<Fufie> bin: msg him
<Fare> then the mighty Baah-Kup listens to my prayers and I get a copy of the contents of bespin on said day
<Fare> binEng: less /root/bin/backup
<Fare> binEng: if you want stuff added in there, tell me.
<eihrul> Fare: what about Bith-Roth, the God of Data Corruption?
<binEng> ok
<Fufie> fare: permission denied
<Fare> Bru is an evil proprietary god
<jdl> Hey, bin
<binEng> jdl: Sorry for loosing my temper like that, I've had a bad day and had got little food.
<Fare> permission denied?
<Fufie> /root/bin/backup: Permission denied
<jdl> bin: No biggie, I think we were all getting a little frustrated.
<Fare> oops. Try again. (wrong permissions on /root)
<binEng> jdl: I thought about another thing, you said the 'ultimate goals' thing was very political.. I don't see that..?
<jdl> Not so much political as subjective.
<Fufie> fare: ok.. thanks for taking backup of /cvs 
<Fufie> fare: I have 31M there I am quite fond of :-)
<binEng> jdl: But surely there can be recognized certain reasons for which computers are used?
<jdl> bin: Sure, but does it really belong in a glossary?  The glossary should say what "ultimate goals" *means,* not what they *are.*
03:30pm
<Fare> Fufie: they account for less than a 2M increase in the .tar.bz2 backup since 19991116
<jdl> bin: Anyway, that's my take on it.  It's not that big of a deal.
<binEng> I see your point
<Fufie> fare: it's 4.3M gzipped so bzip must do a good job with it
<Fare> and it's co-redundant with the existing CVS tree, too
<binEng> OTOH, I didn't intend it to be a strict glossary, but more a discussion around the terms that [should] float around
<Fufie> co-redundant?
<Fare> i.e. K[x;y] < K[x] + K[y]
-:- AlonzoTG [alangrimes@216-164-130-238.s619.tnt2.lnhva.md.dialup.rcn.com] has left #tunes []
<Fufie> yes :)
<Fufie> one day I need to clean the tree though..
<jdl> bin: okay, like I said, no biggie
<binEng> jdl: Do you agree with my direction for the "glossary"?
<Fufie> the structure isn't the same as in 98 anymore
<binEng> jdl: Giving each node an own page would reduce the possible annoyance of such nodes :)
<Fufie> fare: btw, what qualifies as a collaborative tunes project anyway?
<binEng> jdl: How about renaming it to something that reflects its purpose?
<binEng> jdl: Hello?
<abi> i guess Hello is a multi-threaded operating system written in Standard ML and can be found at http://www.ics.hawaii.edu/~esb/prof/proj/hello/
* binEng/#tunes spanks abi
<jdl> bin: sorry, I was on another desktop
<jdl> bin: I'm personally most interested in a completely objective definition of terms.  Specifically, terms used by Prism that are not generally understood or which are used somewhat differently in Prism.
<binEng> with the empasize on 'definition' I guess
<Fare> NSI sent me an invoice. I intend not to pay them.
<jdl> bin: But don't let that stop you from following your vision.
<binEng> jdl: Above all, I should have a purpose of this, do something that is not done yet and there's a need for
<binEng> Fare: What's that again..?
<jdl> bin: Well, my needs right now are just for a page I can point people to when they say "what's an Xxx?"  Prism introduces/modifies a lot of terms, and I'm concerned that it confuses people.  That's why I'm interested in pure definition.
<Fare> networksolutions.com -- part of the internet's problem set
03:40pm
<binEng> Fare: Ah yes, but what if you won't pay them?
<Fare> abi: NSI is networksolutions.com -- part of the internet's problem set
<Fare> bE: I'll pay someone else to be my registrar instead!
<binEng> alright
<binEng> jdl: Do you feel discussion is inappropriate?
<pyro> what did they bill you for?
<Fare> $35 per year
<Fare> and _lousy_ service
<Fufie> you got service at all?
<binEng> Fare: Have you considered removing or separating the politics from the Tunes glossary?
<Fare> bE: I've considered it.
<jdl> bin: Depends on what you're trying to accomplish.  For a pure glossary, yes.  IMHO, a glossary's definitions should be unambiguous and (relatively) difficult to argue with.
<Fare> bE: did you read my original reply to Prof. Shapiro?
<binEng> Fare: I think I did, but can't remember it. What was the spirit of it?
<jdl> fare: I remember seeing that, but don't remember what you said.
<pyro> Fare: what domain.. tunes.org?
<Fare> pyro: yup. 2 years already. And nothing done :(
<binEng> jdl: yes.. as I said, it depends on what goal I choose
<Fare> bE: that I do want the political stuff to be in it, yet recognize that it should be clearly separated as comments; which spearation should ideally be done by proper technological means
<pyro> I discovered NSI a couple weeks ago when I was tracking down some spam.. it was from NSI
<Fare> NSI should be RBL'ed
<jdl> bin: One option is to define your data format such that definitions and opinion are related but separate, allowing the comments to be compiled out in some editions.
<jdl> :) I think Fare and I just said the same thing.
<binEng> sounds like we need the same thing, then
<binEng> jdl: But now you use the word 'opinion'. Is that what you intended?
<jdl> s/definitions and opinion/definitions and non-definitions (comments, opinion, politics, ...)/
<binEng> How about a joint effort? Or maybe there already exist some software for this?
03:50pm
<jdl> bin: I don't know if software exists.  As for a joint effort, if I were to work on it I would want to do it as a Prism code.  :)
<binEng> Like "Everything" but filterable text
<binEng> Pieces of text (not just nodes) could be tagged with (dynamic) tags and filtered/chosen/formatted from that.
<binEng> jdl: Would it be usable in such case?
<binEng> (as for 'joint effort', I was thinking of Tunes that apparently is in need of the same system)
<binEng> With such a thing, I see no reason why both my small glossary and Tunes' one could be merged.
<jdl> bin: I think that may be too fine-grained.  I would be more inclined to have several, um, "chunks" associated with each word.  Each chunk could be associated with a type GUID indicating whether it were a definition, comment, etc.
<binEng> chunks?
<Fufie> chunks?
<binEng> context!
<binEng> come on now guys
<jdl> bin: But don't lose track of the fact that you're trying to create a glossary, not a cool glossary-maintenance program.
<jdl> chunks!
<jdl> glossary: blew chunks, core dumped
<Fufie> bE: put chunks in your glossary as well :-)
<jdl> ;)
<binEng> :P
<Fufie> bE: right next to metamodel and reification :)
<jdl> :b
<Fare> that will all begin to sound like xanadu
<Fufie> bE: if you can wait a few weeks I can let you experiment with such in my hypermedia-system :)
<binEng> really?
<jdl> bin: Actually, XML probably would be a good choice for this application.  (Assuming you don't use fufie's system.)
<Fare> XML is not a good choice for anything but hype
<binEng> _hype_rsystem, yes
04:00pm
* jdl/#tunes grins
<binEng> ofc, marking it as comments and filtering it out in the script is dead easy...
<jdl> bin: XML's advantages in this case: Designed for text; supports structured documents; lots of XML-manipulation tools and libraries available (I think -- ask fufie).
<binEng> looks simply ugly to me
<Fare> XML's disadvantages are: not designed; made clumsy by text; does not really support structured documents; XML-manipulation tools are very heavy and suck
<jdl> bin: XML?  I haven't used it, but it does look good for structured, text-heavy data that needs to be hand-edited.
<Fare> Try SEX
<jdl> s/text-heavy/text-intensive/
<Fare> XML doesn't look good, but for dead simple stuff where just anything would look good.
<binEng> this isn't very advanced stuff really
<jdl> Fare: I'd be more inclined to believe you if it didn't seem like you were just saying the opposite of everything I said.
<Fufie> umm.. xml is not ideal but it's the compromise people agreed to
<Fufie> and it serves it's role as a common way to express dataformats well
<binEng> Fare: Any work done to improve the tunes glossary in any way? How would XML suit for that?
<Fufie> using s-expressions might have been better but we all know how far the allergy against it has come..
* eihrul/#tunes watches as CMUCL calculates the factorial of 50,000.
* Fufie/#tunes hopes eihrul compiled the function
<Fare> xml is just yet another non-sensical "standard" we suffer because of proprietary software -- the proprietary software makers want to hoard code, so they agree on data.
* eihrul/#tunes nods.
<eihrul> Fufie: 50 pages of pure big number....
<Fare> bE: no and yes XML might be better than nothing.
<jdl> Fare: Your criticism smacks more of ideology than technology.
<Fare> as far as technology is concerned, XML is far far behind SEXP
<Fare> and it doesn't allow for inclusion of arbitrary text (as the horridly complex SGML did)
<Fufie> bE: if you make a simple xml-format for the glossary, and use e.g CSS to format it, I can help you when the hypermedia-system is released for making the (adaptive) behaviour you want of the various "chunks" :)
04:10pm
<binEng> hmm
<binEng> jdl: You said you'd go through and comment every entry.. you'd like to do (some of) that now?
<jdl> Fare: I haven't looked at SEXP (and I've barely looked at XML) but in this case I think the criteria are the ability to get up and running quickly while expressing document structure and still allowing hand-editing.  Is SEXP still better
<jdl> (particularly for first criteria)?
<Fufie> stand-alone sexp doesn't do the job, but a markup-system could easily be based on it.. like laml
<Fufie> laml?
<abi> laml is Lisp Abstracted Markup Language at http://www.cs.auc.dk/~normark/laml/
<Fare> jdl: there are many many ready to use SEXP processors. Any LISP or Scheme system, really.
<Fare> jdl: laml, latte, the Rice PLT thing, and many more
<jdl> bin: I can't do it now, I'm finishing up a "C-Like" code to demonstrate some Prism concepts.
<binEng> k
<Fufie> jim: are you making progress on your C example?
<jdl> fufie: Should be done tonight if I don't spend too much time on IRC.  :)
<Fufie> it's a subset of C I guess
<jdl> fufie: I'm pretty much done, just cleaning it up now.
<Fufie> because defining a meta-forma for C seems to be a big task
<Fare> oops, latte is not SEXP
<Fufie> meta-format
<jdl> fufie: Yes, supports everything important except structs and externals.
<Fare> externals?
<eihrul> globally accessable data
<Fufie> send a cc to sds-dev when you're done
<jdl> fufie: will do... I hope to be done soon.  Which reminds me, I should get back to that.  :)
* jdl/#tunes is away (msg me to get my attention)
<Fufie> hmm.. I am two weeks too late in preparations for this semester
<Fufie> I'll never get up in time for lecture at 10am tomorrow
04:20pm
<binEng> goodnight
<Fufie> the undergrads started last week :)
-:- SignOff binEng: #TUNES ( <k!14>)
04:30pm
-:- lar1 [larman@sdn-ar-014casfrMP132.dialsprint.net] has joined #tunes
<lar1> Hey
04:40pm
-:- SignOff jdl: #TUNES (Leaving)
-:- SignOff lar1: #TUNES (Leaving)
-:- SignOff pyro: #TUNES (pyro has no reason)
-:- malc [malc@masq.redline.ru] has joined #tunes
<malc> Fufie: alive?
05:30pm
-:- SignOff malc: #TUNES (tata)
-:- lar1 [larman@sdn-ar-014casfrMP132.dialsprint.net] has joined #tunes
-:- SignOff eihrul: #TUNES (Ping timeout for eihrul[usr5-ppp95.lvdi.net])
-:- eihrul [lee@usr5-ppp95.lvdi.net] has joined #tunes
<lar1> Free BeOS... this is A Good Thing
06:20pm
<air> eh?
<eihrul> lar1: free beer, man
<lar1> eihrul:  Yup
<lar1> air:  The next BeOS is free for personal use... see /.
<air> damn
<air> those bastards
<air> i paid something like $85 for my copy
<lar1> What?  You don't like BeOS?
<lar1> Heh
<eihrul> lar1: free beer is worthless
<lar1> eihrul:  No... you can drink it can't you?
<eihrul> shrug... but you can't get the beer from any other supplier
<eihrul> you always have to go to them if you want the beer
<lar1> Uhh, no
<lar1> Why can't you get another brand?
<eihrul> you can
<eihrul> and that's highly recommended
<eihrul> one with an open recipe :)
<eihrul> but if you choose to stick with that beer
<eihrul> then you're ruled by the supplier of it
-:- hcf [nef@me-portland-us833.javanet.com] has joined #tunes
<eihrul> beer is just nasty anyway
<lar1> Mmm
<lar1> May I call you a neo gpl hippie?
<lar1> air:  Are you testing BRiX on 486es?
<eihrul> hey, fare proved the gpl sucks already :)
<air> hmm, tv looks SOOOOO MUCH better when u use 24bit color instead of 15bit
<eihrul> and that it hinges upon US corporate law and such
<lar1> Fine a neo free-as-in-freedom hippie
<eihrul> lar1: public domain, hippy... maybe, bugroff hippy, maybe...
<air> lar1: 486, k6 and p2
<air> lar1: and pentium
<air> lar1: why?
-:- fester [fester@adsl-pool-1-C7B2B028.chicago.il.ameritech.net] has joined #tunes
<lar1> air:  My code isn't running on anything other then pentiums
<lar1> air: When you fixed BRiX to run on real boxen, was there a bug that prevented it from running on 486es?
<air> no
<air> do u use any pentium instructions?
<lar1> No, I don't think so... its all pretty basic stuff I am doing
<air> u sure?
<lar1> Would invalid opcode triple fault?
06:30pm
-:- fester [fester@adsl-pool-1-C7B2B028.chicago.il.ameritech.net] has left #tunes []
<air> if u dont have the exception setup correctly
<lar1> Hmm, I don't
<lar1> Perhaps I should install my execptions :)
<lar1> What ever genious decided x86 was little endian is on my hit list now...
<lar1> s/was/was going to be
<eihrul> gah, little endian is nice
<lar1> Little endian is rubbish!  Why would anyone want their data backwards?
-:- ult [noone@user-37kbama.dialup.mindspring.com] has joined #Tunes
-:- lar1 is now known as lar1-eating
<eihrul> uh, little endian isn't backwards
<eihrul> its our numbering system that's backwards
06:40pm
<lar1-eating> Whatever :)
<eihrul> it would be so much easier to read in number from text if they were little endian!
<lar1-eating> brb
<ult> eihrul: no, our numbering system isn't backwards.
<eihrul> yes, it is... damn it!
<ult> no, it's not.
<eihrul> troll
<ult> bah. Your the troll.
<eihrul> you're, you mean
<ult> ;)
<air> wtf, beos will be free for newbies but i have to pay to upgrade my copy. thats craps
06:50pm
<eihrul> you worried they'll beat out brix? :)
<air> eh?
<air> beos doesnt even compare to brix
<air> and how does my question relate to brix?
<eihrul> it isn't a question of quality
07:00pm
-:- lar1-eating is now known as lar-homework
-:- jdl [jiml@ultra1.inconnect.com] has joined #tunes
<jdl> Anybody awake?
<abi> i am
-:- ult_ [noone@user-37kbaro.dialup.mindspring.com] has joined #Tunes
-:- SignOff ult: #TUNES (Read error to ult[user-37kbama.dialup.mindspring.com]: Connection reset by peer)
-:- jdl is now known as jdl_away
07:20pm
<hcf> ho hum
<eihrul> ?
<hcf> yawning
07:30pm
-:- lar-homework is now known as lar1
<jdl_away> This is *way* off-topic... but can anyone tell me how to set my computer's clock from Linux?
<lar1> There is no time command?
-:- jdl_away is now known as jdl
<jdl> There is, but I can't figure out how to use it.  (Man page brings up C docs)
<eihrul> date?
<abi> i heard date was Tue Jan 18 2000
<eihrul> abi: not that date
<jdl> Ah ha!  Yep, it's date.  Okay, I think I've filled my newbie question quota for the day... thanks.
-:- rares [rares@wtrb-sh3-port28.snet.net] has joined #tunes
-:- jdl is now known as jdl_away
-:- tim [user9783@blastbc01-69.bbsi.net] has joined #tunes
-:- tim [user9783@blastbc01-69.bbsi.net] has left #tunes []
<lar1> :)
-:- SignOff rares: #TUNES ([x]chat)
08:20pm
-:- lar2 [larman@sdn-ar-005casfrMP130.dialsprint.net] has joined #tunes
-:- rares [rares@wtrb-sh4-port210.snet.net] has joined #tunes
-:- SignOff lar1: #TUNES (Ping timeout for lar1[sdn-ar-014casfrMP132.dialsprint.net])
-:- lar2 is now known as lar1
<lar1> air: mov cr0, eax is ok on a 486, right?
08:30pm
-:- water [water@tnt-10-171.tscnet.net] has joined #tunes
<water> 'lo
<hcf> hel`
<lar1> Hey water!
<water> hcf: could you get me any papers that are specifically about bisimulation or bisimlarity?
<hcf> water: possibly ;)
<hcf> additional keywords?
<water> maybe adding in coinduction as a keyword would help
<water> darn it, i had some great ideas for arrow papers and implementation today
<water> also some good slate stuff
<water> oh good, tril's helping me with web stuff
<hcf> abi: tell water about laml
<water> thx
* eihrul/#tunes bows.
<water> hey eih
<water> we got a mlist now
<eihrul> hcf told me
<water> k
* eihrul/#tunes has been taking a detour through "Advanced Compiler Design and Implementation."
<hcf> i knew cuz i'm on the tunes-cvs mlist ;)
08:40pm
<water> ah
<water> eih: i have that book
<water> heh. it seems the tunes front page is handled entirely manually :)
<water> damn, tonight's going to be a long night
<eihrul> why so?
<water> heh
<eihrul> water: some of the writing is little crude....
<eihrul> s/is/is a/g
<water> a lot to add the slate web content and a lot of formatting to add, as well as mlist introduction
<water> eih: yeah, oh well
<eihrul> but, its still more than i'd find searching around otherwise....
08:50pm
<water> hmm with luck i should have beos working online in a few days
<water> alright, down to the slate discussion
<water> the best reduction of the beta semantics for applying an object as a function seems to be an implicit method in every object's protocol
<water> which i believe actually subsumes the accessor methods
<water> in that, i see little conflict
<water> we simply have the system's meta-object handle variable access and cloning
<water> in the meta-object would be stored the implicit cloning of the object on state-access
<water> which roughly yields immutable state
<water> an interesting question is how "versioning" of state should be handled
<water> some references would be "shallow", referring to an object and not its value, while others would have varying types of "deepness", taking values of variables selectively, so that updates to other parts of the object's state would not update the value of the reference
<water> this could be represented by a bitfield
<water> this make sense so far?
<eihrul> sorry, was having a discussion on tries in another window, let me read everything :)
09:00pm
<eihrul> hmmm
<eihrul> if objects are cloned when state is updated
<water> perhaps lazily
<eihrul> then why must you need to worry about the references themselves needing to be modified?
<water> because cloning will actually push the previous object value *away* from the reference
<water> the object would essentially be the last value it had
<water> that is, the current value
<eihrul> but then references are mutable :)
<eihrul> which means the enclosing object is mutable
<water> not if i have the above system in place
<water> you could even make partially-immutable references
* eihrul/#tunes is almost losing faith in immutable objects.
<water> hm
<eihrul> actually, moreso the stream representation needed to simulate mutability with immutable objects
<water> well, references being mutable shouldn't be a problem if they're implicit most of the time
<eihrul> even in things like a method
<water> i.e. if you know the semantics of the reference, back-tracking can be done on demand
<eihrul> okay, how would exploit the timeline of an object for practical benefit?
<eihrul> s/would/would one/g
<water> well, there's the obvious debugging use
<water> but mostly you get just different semantics available
<water> hm
<eihrul> well, what if an object is really the head of the stream?
09:10pm
<water> "really" as opposed to implementationally?
<eihrul> -really
<eihrul> just a superfluous word inserted
<eihrul> in that when it is modified, it clones a copy of itself and pushes it behind it in the stream
<eihrul> this way the stream is always at a definite location
<water> actually, a timeline would help for any interactive analysis
<water> sure
<water> for instance, you could model mouse-input with a stream
<eihrul> the challenge would then become when to optimize away the tail of the stream
<water> yes that would be a challenge
<eihrul> its probably not impossible
<water> you would have to refine the type of stream to handle the semantics of its use
<eihrul> just noone's made a language that considered objects as streams yet :)
<water> yeah, i know
<ult_> objects solely as streams?
<ult_> sounds like it would have too darned much overhead...
<water> well, like i said... if you knew ahead of time when an object's stream was discardable, then you could potentially allocate finite resources to it
<water> i.e. if you had static-style declarations
<water> and immutability handles that :)
<ult_> true
<ult_> but then you'd want lots of typing
<water> that's one reason for immutable objects... you can define what variables in context can change and what remain constant
<water> er.. partial-immutability
<water> it essentially gives you a mechanism for static typing and/or constants
<water> but this mechanism can still be dynamically modified when necessary
-:- SignOff lar1: #TUNES (Ping timeout for lar1[sdn-ar-005casfrMP130.dialsprint.net])
<eihrul> hmm, sedgewick has disappeared
-:- lar1 [larman@sdn-ar-013casfrMP055.dialsprint.net] has joined #tunes
<hcf> sedgewick?
<eihrul> robert "algorithms in c++" sedgewick
09:20pm
<eihrul> ah, it lives yet
<eihrul> water: problem it seems is that you almost cannot determine in any way but prophesising the future that an object's history won't be accessed
<eihrul> in that any history not created will never exist
<water> hm
<eihrul> look-ahead-in-time-buffers, the next great compiler optimization trick! :)
<water> then we'd have to provide mutable state by default
<eihrul> so long as the default can be easily changed, it is irrelevent no?
<eihrul> so more important than the default itself is how its changed
<water> hm
<water> seems like a meta-property
<eihrul> and there could be many *defaults* :)
<water> maybe we could clone the meta-object for each kind of case
-:- washort [washort@d115.narrowgate.net] has joined #tunes
<water> with objects of various sizes, that could get complex
<water> otoh
<water> maybe we could provide all of the desired behavior by pretending that all the meta-objects are one meta-object
<eihrul> you mean localing meta-properties to certain sets of objects rather than all objects?
<water> yeah
<eihrul> s/localing/localizing/g
<eihrul> then that brings up the question of how two objects with different semantics would interact :)
<water> yep
09:30pm
<water> there is also the possibility of treating integers as types of references to an "iterator" object
<washort> water: hey, saw your post to the squeak list.... and the responses. =)
<water> wash: hey cool
<washort> water: Self and BETA? sounds like a rather interesting combination. =)
<water> yeah, it's pretty scary in the details, as we're finding
<water> particularly because i'm trying to work out immutable objects
<washort> *nod*
<washort> it's a stretch though, Self seems to be the most flexible OO language i've ever heard of; BETA seems extremely, um, fascist
<washort> (in a good way of course :)
<eihrul> water: so what of mobius is there yet?
<water> well, we're working that out from the self perspective
<washort> water: *nod*
<water> eihrul: mostly the abstract ideas, since i have no working slate system yet
<eihrul> well, i mean what ideas....
<hcf> water: for bisim poop, http://citeseer.nj.nec.com/cs?dd=2 w/ query: coinduction and bisimulation and bisimilarity, for more results remove bisimulation or bisimilarity
<water> oh geez
<water> well, i'm looking at providing purely functional versions of slate using meta-objects alone
<eihrul> water: my mind is gravitating more how would mobius allow one to specify immutable objects more than how to provide them :)
<water> i.e. ones where an object's protocol isn't defined locally, it's defined in a shared type-space
<water> well, sure
<water> that part of mobius would be built into the language, though
<water> like i said, the internal representation of mutability could be a bit-field
* rares/#tunes thinks somethings overdone dunno why
<water> huh?
<rares> i don't know maybe I'm just used to emergent systems
<water> define emergent
<rares> emergence-ology: the study of convenient side-effects
<water> oh
<water> yeah, i think it's been overdone :)
<rares> you're doing it from the inside out I'm used to the other way around
<water> what's the other way like?
<rares> the other means killing a lot of birds with a few well aimed stones
09:40pm
<rares> sort of like letting your model grow on its own
<water> hm
<washort> "build everything from the top down, except the first time"?
<rares> only you know it won't go outside certain constraints even though they're not explicitly drawn
<water> well, look, i gotta keep working
<water> eih: where were we?
<eihrul> mobius, bit fields, (im)mutability
-:- SignOff jdl_away: #TUNES (Leaving)
<water> hm
<water> well, we could treat the meta-object as completely immutable and tell it to turn immutability on or off for a given variable
<water> that way the default meta-object settings stay the same for new objects
<water> and the implementation would fiddle with the bit-fields which get read on accessor messages
<eihrul> processors don't like bits :)
<water> huh?
<water> all we gotta do is mask
<rares> performance drops like a stone when you start reading bitfields
<water> hm
<rares> but a good mask can avoid some of that
<water> how so?
<rares> if you want to read a yes or no from one byte
<rares> the silly way to do it
-:- Downix [down@d-gnaps-22.ici.net] has joined #tunes
<water> i think there's a better way
<rares> course
<rares> i'll get there
<rares> you want to test bit 6 for 1
<Downix> hmm
<rares> silly way:
* eihrul/#tunes ponders the Tao of KISS.
<rares> shift 1 6 places masak shift back down 6
09:50pm
<rares> that's the stupid way
* water/#tunes agrees
<Downix> Anyone know where I can find an introduction to 68k asm?
<water> Dx: motorola.com?
<rares> especially when you want to check all the bits even tually
<Downix> water:  Doh,  hehe
<rares> 24 instructions is silly
* ult_/#Tunes kisses Tao
<Downix> What is Tao?
<abi> Tao is an expression of the Water or tao, the (so-called) ultimate os at http://www8.pair.com/mnajtiv/tao.html or at http://www.tao-group.com
<water> k
<rares> what you could have instead is a table of results from masks
<rares> and instead of shifting back
<rares> just use the results directly
<water> describe this table
<eihrul> tis easier to just uniformly clone already :)
<rares> gimme a sec
<rares> to formulate
<eihrul> if slots themselves are objects
<eihrul> they can have their own meta-properties, perhaps
<water> oh yeah
<water> damn it, i forgot my own assumption
<water> that slots are independent objects
<water> no, we still have the problem of *references*
<water> rares: continue
<eihrul> could you kindly repeat the problem? :)
<rares> say you have 8 bits
<water> eih: even if the objects referred to by an object were mutable or immutable, the references would have a semantics independent of them
<rares> each one means something
<rares> some are related
<rares> some not
<rares> for the realted ones you can test them all then
<rares> at the same time
<rares> for the not so releated ones
10:00pm
<eihrul> water: perhaps references and immutability just don't mix :)
<rares> you can still skip the silly back shift
<water> maybe i could tag the reference
<water> with just a bit
<rares> by hard coding if blahtest = 16 instead of 1
* ult_/#Tunes ponders deeply
<water> ok
<ult_> Shmarty Pants
<eihrul> all this bit testing is just to decide which data swims down the stream?
<water> for now, yes
-:- witten [witten@adsl-gte-la-216-86-196-179.mminternet.com] has joined #tunes
* ult_/#Tunes bonks eihrul
* eihrul/#tunes thinks references don't serve much of a purpose with immutable objects.
<water> hm
<water> yeah they would
<water> you could use them to give you constant values for some iteration function
<witten> hm. slate sounds good
<water> hi witten
<witten> i'm making something similar
<witten> hello water
<water> cool
* rares/#tunes thinks he'll stick the shell surrounding all this good stuff and let you figure it out
<water> for instance, i could enumerate certain kinds of methods that satisfied some property
<water> np thx
<ult_> cast void bolt on eihrul
<ult_> oops ;)
<water> and then i could pick any one of those methods generated by the iterator and because it belonged to that stream, i would know that it possessed certain properties
<eihrul> i'm already there, bro
<ult_> Where?
<ult_> da void?
<water> i could even use that idea to describe other languages
<water> (potentially)
<hcf> water: dont forget to add mentions of irc&mlist to slate-home
<water> bitten: you a tunes member?
<water> hcf: just uploaded the changes
<witten> water: who, me?
<water> but i still must make a news page
<water> witten: yeah, typo :)
<witten> water: no, but i've followed tunes for quite a while
<water> ok
<water> slate's supposed to be a tunes hll proposal
<witten> only a proposal?
10:10pm
<water> well, if they don't want it, then it's not tunes :)
<witten> "they"?
<washort> does TUNES necessarily need a low level language? =)
<water> the majority of tunes members
<washort> i know previous systems in the same vein did not. :)
<water> washort: not sure
* washort/#tunes hopes the Crusoe is something really cool
<witten> how do you implement an OS in a hll?
<washort> witten: Lisp Machine. =)
<washort> make it a *good* high-level language, mainly. :)
<witten> oh, custom hardware?
<washort> witten: it helps
<witten> hm
<water> well, you can describe an instruction set and architecture operationally with lisp
<washort> stack-based anything would be enough, in my amateur opinion
<water> and then if you re-write the lisp system in lisp and cross-compile
<washort> *nod*
<water> that's the basic idea
<washort> water: amazing how many ppl don't seem to see that. =)
<water> yeah
<washort> water: philosophical question - do you think that there's any technological reason to keep using C, considering the state of the art?
<washort> i know there is much social and historical reason to
<eihrul> washort: that's more of a moral question
<washort> plus a lot of C code lying around 
<rares> well I'm going to relieve X of some of the swapo it's wasting
<washort> eihrul: Heh :)
<eihrul> washort: C is constraining, repressive, and down-right evil
* water/#tunes focuses on writing emails to slate respondents to inform them of the mlist
<washort> eihrul: constraining? why, you can do *anything* in it. ;-)
<eihrul> more *how* than *what*
<washort> eihrul: *nod* exactly. 
<washort> eihrul: you can do it with zeros and ones. are you going to? not likely. =)
<witten> it's scary how i've come to some of the exact same conclusions as the slate people
<washort> eihrul: i paid a visit to #C++ on efnet the other day, asking what made C++ worth using :)
<water> witten: i *am* the slate people :)
<washort> eihrul: they couldn't tell me, except for "it's OO" and "It's better than C"
<witten> haha
<eihrul> washort: just need the first dynamic languages to out-optimize C, then let all the nay-sayers wither away :)
<witten> water: well then i guess we're both Self loser :)
<witten> s
<rares> i'm not going to trust a full GUIed system until I no longer need text to express structure some thing no visual tool provides
<washort> eihrul: I think Dylan may do it. ever looked at it?
<water> heh
<water> i'm on the self mlist
<witten> eihrul: how do you have a dynamic language out-optimize C?
<eihrul> washort: dylan is not sexp
<washort> eihrul: True.
<eihrul> witten: how does C optimize?
<washort> eihrul: but it's "good enough" for a lot of stuff.
<water> witten: by having it dynamically optimize for the current case faster than c can
<washort> eihrul: I favor incrementalism. :)
<witten> eihrul: static typing, for starters :)
<witten> water: hm. like sun's hotspot?
<eihrul> witten: not static typing but type information
<washort> hotspot? heh.
<water> witten: sure
<eihrul> witten: static is just a means to an end
<witten> eihrul: a fast end :)
* eihrul/#tunes shrugs.
<rares> just don't get lost this is too cool a project
<eihrul> fast on what?
<eihrul> witten: its just dislocation... you do more, compiler does less
<eihrul> in the end, how fast is it?
* washort/#tunes likes his OO languages to be very dynamically typed and his functional languages to be very statically typed
<eihrul> i can do more in 2 lines of lisp code than i can in 200 of C :)
10:20pm
<witten> eihrul: faster than if you don't tell it the types and it has to be resolved as runtime
<washort> except Lisp, which is special. :)
<water> washort: i've found a good way to formalize a reason for that
<washort> water: nifty
<eihrul> witten: nah, it's just caching... 
<witten> eihrul: caching?
<water> it'll wind up in slate docs eventually
<eihrul> yes, caching, the answer to everything
<washort> water: i think it has a lot to do with the fact that my experience with dynamically typed OO langs ha been very good while my experience with statically typed "OO" langs has been ratehr bad
<witten> eihrul: so you're suggesting that runtime type resolution be cached, and that'll be as fast as static typing?
<washort> (i.e. C++, Java)
<eihrul> witten: just remembering runtime determined type information should have the same effect, theoretically
<washort> then there's stuff like Haskell which gives me a headache easily but is really cool :)
-:- SignOff hcf: #TUNES (Leaving)
-:- hcf [nef@me-portland-us713.javanet.com] has joined #tunes
<ult_> Haskell is -evil-.
<washort> ult_: really?
<washort> why?
<ult_> monads burn yer gonads.
<washort> haha
<washort> seriously, why?
<ult_> hmm. It's a nice enough language in theory...but it looks to be a lot like Scheme...cool for teaching functional programming and the like...but...wouldn't be great in real world applications.
<witten> eihrul: i suppose i'd have to see it to believe it
<ult_> Frankly, the syntax looks like something some visual Haskell program generated =)
<washort> mmm, i've got a couple friends who are rather fond of it --
<washort> ult_: eh, it's no worse than Python. :)
<ult_> Python is evil ;)
<ult_> Python == Unix basic!
-:- water_ [water@tnt-10-171.tscnet.net] has joined #tunes
<washort> ult_: the XML library for Haskell is really cool. wish i could wrap my head around it
<Downix> lol
<witten> eihrul: in the mean time, i'll stick to my old fashioned static typing :)
<rares> abi lang-list
<washort> ult_: i don't love python very much either, but it's paying the bills at the moment
<water_> langlist?
<abi> langlist is at http://cuiwww.unige.ch/langlist
<washort> ult_: by way of Zope...
<washort> ult_: and it could be mcuh, much worse
-:- SignOff water: #TUNES (Read error to water[tnt-10-171.tscnet.net]: Connection reset by peer)
<washort> say, Java
-:- water_ is now known as water
<ult_> True.
<ult_> Haskell's not a horribly bad language, but i just don't see it being used in the 'real world'
<rares> thx water
<washort> ult_: i'd at least like to try. :)
<rares> it depends on what you need
<rares> i define a circle as radius and center
<washort> Interestingly enough, the most sucessful "real world" FP language at the moment seems to be Erlang.
10:30pm
<washort> (which is dynamically typed, and a bastard child of Prolog...)
<rares> you could define it by naming every point and including it in a set
<eihrul> ult_: well, the haskell compiler is in haskell....
<washort> eihrul: most FP languages are very good for writing their own compilers. ;)
<eihrul> washort: gah, lisp or bust! :)
<washort> eihrul: I like Lisp... pity there's nothing to Lisp as Squeak is to Smalltalk.
* eihrul/#tunes cackles.
<rares> Leak
<eihrul> rares: shhh
<washort> hm? :)
<rares> I know nossing
* washort/#tunes waves his recently-checked-out library copy of "The Art Of The Metaobject Protocol"
<washort> the more I learn, the more I like. :)
<eihrul> washort: squeakish lisp => leak
<water> cool
<water> amop, that is
<washort> water: *nod*
* eihrul/#tunes has that... still need to read that
<washort> eihrul: URL?
<eihrul> but need to finish acd&i first
<eihrul> washort: its still in research phase :)
<washort> so is Squeak. :)
<eihrul> well, unapplied research phase
<washort> hehe
<washort> most of Tunes seems to be there....
<rares> umm you know you need a new processor when bladenc -br 192 takes 35 minutes
<eihrul> washort: well, whether or not tunes or slate pans out, i'm making leak :)
<washort> eihrul: *nod*
<washort> eihrul: CL, CLOS, CLIM and all? :)
<eihrul> i was thinking of using an unCommon Lisp
<washort> any particular reason?
<eihrul> common lisp is *huuuge*
<washort> well, sure :)
<water> k, i let all the slate respondents know about the mlist and gave them some general news
<eihrul> tunes would get finished first
<eihrul> washort: something ala eulisp, perhaps with a hint of prototyping i'm thinking :)
* washort/#tunes doesn't know eulisp
<eihrul> well, i'm not too thrilled with object system as is
<eihrul> but otherwise its somewhere to start from
<rares> abi langlist
<abi> rumour has it langlist is at http://cuiwww.unige.ch/langlist
<washort> i've never heard of eulisp....
<eihrul> abi: eulisp?
<abi> hmmm... eulisp is http://www.bath.ac.uk/~masjap/EuLisp/eulisp.html and http://www.aiai.ed.ac.uk/~jeff/lisp/eulisp.html and ftp://ftp.maths.bath.ac.uk/pub/eulisp/
<eihrul> washort: i just heard about it a week ago myself
<eihrul> but, its old apparently
<water> damn, it's one of the first lisps i ever used
<washort> newer than CLOS apparently.
<witten> who needs langlist when you've got http://www.ionet.net/~timtroyr/funhouse/beer.html
<washort> witten: heh
<washort> witten: that doesn't cover everything though
<eihrul> washort: more coincidental i believe
<eihrul> though, perhaps newer
<washort> y'know, it probably wouldn't be too hard to put the sexp syntax back into dylan
10:40pm
<washort> the gwydion ppl might be amenable if code was contributed...
<water> i'd bet dylan'd be happy to be getting more SEXP ;)
<washort> sexp?
<abi> rumour has it sexp is see s-exp
<washort> t!
<eihrul> nah... i'd rather start from scratch
<washort> eihrul: see, i *hate* reinventing wheels :)
<witten> NIH syndrome is fun for the whole family!
<eihrul> washort: sure, but i've yet to learn how to make certain wheels
<eihrul> crude, unornate wheels are not worth it
<eihrul> but elegantly crafted wheels have a certain appeal in their study :)
<washort> it's frustrating watching all the unix weenies desparately trying to do same when all the really good wheels are lying out back (lisp, smalltalk, garbage collection, dynamic typing, you name it)
* water/#tunes nods
<eihrul> well, they're just trying to turn their squares into a wheel
<washort> eihrul: heh, true
<witten> you know what.. the unix weenies have got something done. :)
<washort> the square clunks along pretty good, relatively though :)
<witten> even if it's ugly and square
<washort> witten: worse is better.
<eihrul> witten: sure, and 20 years from now, they'll recreate lisp
<witten> washort: there you go.
<washort> witten: to quote a famous paper about Lisp. :)
<witten> washort: yah, read it
<washort> eihrul: well, aren't you trying to do that now? :)
<eihrul> washort: no, i'm just reviving it
<eihrul> but moreso that i wasn't born around the 70s to see the evolution of lisp :)
<eihrul> so unfortunately, i'm doomed to reinvent it!
<washort> eihrul: poor you. 
<washort> poor us. =)
* washort/#tunes is content to read about the history of lisp. ;)
<washort> but i wouldn't mind finding out that the Crusoe chip is ideal for building Lisp Machines around. ;)
<rares> my ehole take on it is: I should be able to control similar objects in similar ways without all the possiblities having to be spelled out in command buttons
<washort> or Smalltalk Machines for that matter. =)
<eihrul> washort: but the utopian lisp has yet to be created....
<witten> eihrul: what's wrong with the current lisps?
<washort> eihrul: of course.  the MOP obviates that need to a large extent. :)
<eihrul> witten: they cost more than a used car
<water> lol
<witten> eihrul: you silly longhair, you
<washort> eihrul: Implement a small lisp in itself, then we'll see where we can go =)
<washort> witten: free is important. =)
<eihrul> washort: that's the idea
<witten> washort: i know, i know :)
<washort> eihrul: ok, i;m cathing the idea now :)
<rares> unchained software versus software in a bottle in the middle of the sea
<washort> witten: the Unix pushers never would have gotten anybody hooked if they hadn't given away free samples. :)
<washort> rares: heh.
* water/#tunes starts adjusting the slate-semantics page to current policy
<Downix> indeed washort
<witten> washort: the early commercial unices weren't really free, were they? had to buy the hardware
* washort/#tunes found a copy of the Unix-Haters Handbook at the library. now i'm an equal-opportunity flamer. :)
<rares> at least the samples have longer durability than most of the stuff made these days
<Downix> lol
<washort> witten: you'd have to do that anyway. :)
* ult_/#Tunes ponders
10:50pm
<ult_> goodnight and goodluck eihrul.
<eihrul> live long, and prosper
-:- SignOff ult_: #TUNES (Leaving)
<water> damn it's 11pm already here
<eihrul> yep
<Downix> Abi, MorphOS?
<abi> downix: wish i knew
<washort> it's 1a here...
<Downix> Ok, anyone here ever heard of "MorphOS"?
<rares> what's it do?
<Downix> Apparently it runs on the "common" microkernel, Quark, and runs amiga software
<water> running amiga software is no small feat
<rares> 2am here
<Downix> water:  No, which makes me wonder....
<water> that's be cool to see an os project with scores of items in its feature list, and at the bottom you see "runs all amiga software"
<water> :)
* Downix/#tunes laughs, I know
<lar1> heh
<Downix> MorphOS:  http://www.morphos.de
11:00pm
<water> darn it, i'm somehow torn between having objects' slots be a literal part of the object and having slots be independently referenced by the object
<witten> water: make 'em separate
<water> it seems to make thinggs more complicated though, considering i'm experimenting with immutability controlled by meta-objects
<hcf> water: on slate-members, either remove the mention of an mlist or link to slate-collab
<water> hcf: oh yeah
<witten> water: which objects are you talking about? the object containing the slots, or the slot containing an object?
<water> not sure, perhaps that's why i'm confused
<witten> water: well i tried combining the slot with the object it contains, and that simply doesn't not work
<water> i'm considering allowing an object to specify its state's mutability slot-by-slot
<water> but if slots are just references
<witten> water: make a slot contain other information (like a "mutable" boolean value or something)
<water> that's not the problem, though
<witten> what is it?
<water> the problem is when you consider that slots from different objects might reference the same object with different semantics
<witten> correct..
<witten> oh, so you're asking about the case if an object is referenced from one slot as mutable and another slot as immutable?
<water> it could be dealt with, but not simply afaik
<water> sure
<witten> hm
<water> the real issue is the relationship between object and meta-object systems
<witten> meta-object systems being the slots?
<water> because when you change a slot's style of reference, it's really making a clone of the existing meta-object for the object to a new one that provides the desired accessor semantics
<water> s/semantics/behavior
<water> since all accessors are messages
<water> hcf: k, i linked to collaboration and uploaded
<witten> "style of reference"?
<water> yeah, mutable or immutable access to objects
<witten> okay
11:10pm
<water> btw, there are very interesting uses of this
<witten> such as?
<rares> read/write or read-only for us old hackers
<water> well, you could treat the integers as immutable references to a "primitive iterator" object's past state
<rares> anyway I have work to do folks so I shall leave thee now
<water> s/past//
<witten> water: why would you want to do that?
<water> it just seems more unifying
<witten> i'm not sure how
<water> it would make the core set of programming objects smaller
<water> hm
-:- rares [rares@wtrb-sh4-port210.snet.net] has left #tunes []
<witten> okay, i think i see what you're saying
<water> well, at least partial-immutability would allow you to take things as constants dynamically
<water> which could help with optimization
<water> sorry, these are all very new ideas to me
<witten> no problem
<witten> they're new to me too :)
<water> at least, from the implementation standpoint
<water> heh
<water> btw, the slate idea arose from a project i did in college
<water> i had a lisp-based object system with a direct-manipulation gui
<water> but the code is lost now
<witten> that sucks
<water> oh well
<water> i did it after learning of the tunes project
<water> i thought "that's easy" and went out and did it :)
<witten> and was it? :)
<water> not quite
<water> i learned a lot, though
<witten> heh
<witten> yeah, that's the best part of it, i would think
<water> at any rate, i'm trying to fit all the language-unification lessons i've learned in the last few years into slate
<witten> it seems to me that the mutability toggle would have to be in the object itself, not the slot
<water> hm
<witten> cool
<water> i'm not so sure about that
<witten> if it's not, then you can end up with an object that is simultaneously mutable and immutable
<water> well, it's not really either
<water> you just have (branching) streams of state
<witten> then what good is the boolean value if it's not either?
<water> and objects tap into the stream in various ways by name
<water> (damn, that's an esoteric programming concept)
<witten> heh
11:20pm
<hcf> water: slate-members s/recieved/received/  ;)
<water> doh!
<water> i used to have great spelling habits a few years ago
* eihrul/#tunes welcomes water to the land of IRC.
<water> fixed
<lar1> heh
<water> :P
<eihrul> off to read more acd&i
-:- SignOff eihrul: #TUNES ([x]chat)
<water> well, the issue does arise if you change an object's meta-variable, "name"
<water> since now references won't follow the new branch in the stream
<water> unless... i don't know, maybe there's an exception to that
<water> that might raise some thorny questions about meta-circularity
<water> hm... i recall this paper that talked about avoiding such by the use of a "top" meta-object that simply described itself
<water> otoh, maybe references wouldn't (or shouldn't) be text-symbol based
<water> darn it, it sounds like i'm veering too far from the slate intent
<Downix> uh oh
<water> yes, that's too far
<witten> water: i tried to think of a way not to do it with text symbols
<witten> i couldn't come up with anything halfway useful
<water> if people want nameless (fully-intensional) references, they could smply make do with a new mop
<water> witten: that's what arrow's for :)
<witten> arrow?
<abi> hmmm... arrow is a homo-iconic information manipulation system intended to support programming and human languages in the way that Tunes should. at http://www.tunes.org/papers/Arrow/, http://www.tunes.org/~water/
<water> hehe
<water> thanks, abi
<abi> water: no worries
<water> now, i've looked far and wide, and there's no idea more esoteric than arrow afaik :)
11:30pm
<water> but i'm sure that it's workable
<witten> how?
<witten> or must i read this entire paper to get it? :)
<water> well, not all the answers are in the (now obsolescent) paper
<water> but the concept is there
<witten> hm
<water> well back to the subject
<Downix> ok
<water> well, although it seems pretty hard to implement, i'll stick with the immutability by reference for now
<water> until i find a conflicting issue with meta-objects or something
<witten> i'll be interested to see how you work it out
<water> subject change: the meta-object should provide state-change of multiple variables all at once if requested
<witten> how?
<water> well, it would be like accessors of multiple variables in smalltalk
<water> myObject x:1 y:2 z:3
<witten> oh
<witten> that's a syntax issue, ain't it?
<water> well, syntactically, it'd still be the same
<witten> same as what?
<water> but i'd like the mo to encapsulate the state-change all at once
<water> well, it would look exactly like a sequential state-update
<water> syntactically
<witten> hm
<water> but the system wouldn't update the stream with 3 new states, but one
<water> (this is just a proposal)
11:40pm
<witten> why does the number of state updates matter?
<water> well, if you take it literally, it'd wind up with more space taken up
<water> uh oh
<water> i just realized that the mo intercepts all state-updates and can change the default behavior
<water> hm
<water> i'd better just leave it as an implicit optimization for the default case only, to avoid weird bugs
<witten> i have no idea what you're getting at now, but that's okay :)
<water> this language will be quite slow, if i use the default oo ideas
-:- SignOff lar1: #TUNES (RavenOS -- NOW!)
<Downix> hmm
<water> i just meant that state-update might have weird conflicting effects
<Downix> water:  You need good hardware for it then
<water> lol
<witten> water: what aspects of OO will slow it down?
<water> just the usual architecture of oo environments
<Downix> water:  *pitches Eddas some more*
<water> but then, i doubt a slate environment *should* look traditional at all
<water> hm... traditional oo would probably be smalltalk w/ mvc
11:50pm
-:- SignOff Downix: #TUNES ([BX] Occifer, take me drunk, I'm home)
<water> the audience has worn thin, it seems
<water> oh yeah, it's getting late
<witten> apparently so
<water> it's to be expected
<water> i'm so dedicated to my ideas... perhaps i should be *committed* :)
<witten> haha
<witten> i know how that is
<water> the tunes people know that i have often gotten very obsessed
<hcf> water: why rnt the tunes ppl more like u?
<water> ok, uploaded some changes to semantics
<water> hcf: because tunes seems to attract the wrong crowd, i guess
<hcf> what attracts the right crowd?
<water> heh. official academic or commercial support, i guess
<water> we're too "unprofessional" a band of coders
[msg(TUNES)] newlog 2000.0119
IRC log ended Wed Jan 19 00:00:02 2000