IRC log started Sun Apr 25 00:00:01 1999
[msg(TUNES)] permlog 1999.0425
-:- smkl [sami@MCCV.rdyn.saunalahti.fi] has joined #tunes
-:- hcf [nef@me-portland-us319.javanet.com] has joined #tunes
!koontz.openprojects.net!! Remote CONNECT vinge.openprojects.net 8005 from lilo
!mccaffrey.openprojects.net!! Remote CONNECT drexelhill.pa.us.opirc.nu 8005 from lilo
-:- tcn [tcn@cci-209150250088.clarityconnect.net] has joined #tunes
<tcn> who's here?
10:50am
* hcf/#tunes is
<tcn> what's up?
<tcn> ok
<tcn> umm.. how do I type in chat?
<hcf> it differs depending on ur client
<hcf> just join #tcn
11:00am
-:- SignOff smkl: #TUNES (((())))
-:- _QZ [brand@p0wer.qzx.com] has joined #tunes
<tcn> hey qz
<_QZ> hey tcn
<_QZ> anything new on retro?
<tcn> nah, too busy
12:00pm
-:- SignOff tcn: #TUNES (tcn has no reason)
-:- tcn [tcn@cci-209150250086.clarityconnect.net] has joined #tunes
<tcn> anything new w/ brix?
12:10pm
>>> tcn [tcn@cci-209150250086.clarityconnect.net] requested PING 925066600 925236 from #tunes
<hcf> tcn: here?
<hcf> 
<tcn> yeah
<hcf> RETRO #8 bootloader
<hcf> _
<hcf> the _ is the blinking cursor
<hcf> and thats it
<hcf> and btw, 'rawrite3 image a:' doesnt work cuz rawrite3 doesnt take cmdline params
<tcn> oh yeah :)
<tcn> What OS are you using?
<hcf> dos6.22
<tcn> we could figure out where it's hanging
<hcf> using image.1999-0417
<tcn> you get GNU make and NASM, and compile it yourself..
<tcn> not today, though
<tcn> I was gonna compile a DOS version of NASM for someone else, when I have more time.
<tcn> i'll add you to the list
<tcn> You can add instructions to tell you how far it got before it crashed, and figure out where the problem is
<hcf> it does the same thing on my 486/25
<tcn> is that a PS/2?
<hcf> no
<tcn> how much RAM?
<hcf> 24
<tcn> hmm, weird
<tcn> what kinda video card?
<hcf> herc ;)
<tcn> no wonder!
<hcf> herc-mono
12:30pm
<hcf> i know
<tcn> the video buffer isn't at 0xB8000 on that machine.
<tcn> ok, it's time to improve that :)
<hcf> but why does the RETRO #8 ... party get printed?
<tcn> that's done through the BIOS
<hcf> thotso
<tcn> that's the first thing it does
<tcn> what kinda video card does the PS/2 have?
<hcf> vga-mono gas plasma display
<tcn> oh yeah
<tcn> forgot
<hcf> the text mode mem should work
<tcn> it might have the buffer at 0xB0000
<tcn> hey, try booting it, then press some keys.. if the PAUSE key reboots it, it's working.
<hcf> the ps/2's text mode mem is at 0xb8000
<hcf> thats no the prob
<tcn> ok, it's probably either the 386 or the MCA bus
<hcf> probly the mca
<tcn> I might have used a 486 instruction somewhere
<tcn> Sometime I'll have to read about MCA
12:40pm
<hcf> got something
<hcf> tried it on a at&t 386/25 w/ vga
<hcf> R...
<hcf> 0287
<hcf> AX:0011 BX:0000 CX:0000 DX:0000 Disk Error
<hcf> bad floppy?
<hcf> btw, should having or not having a match co matter?
<tcn> umm
<tcn> shouldn't matter
<tcn> that's a floppy error.. try it a few times
<tcn> hey, it does have a 1.44M drive, right? :)
<hcf> yeah, all 1.44m drives
<tcn> if it boots a DOS floppy, the problem is retro :)
12:50pm
>>> tcn [tcn@cci-209150250086.clarityconnect.net] requested PING 925069230 220638 from #tunes
<tcn> abi, seen fare?
<abi> tcn: bugger all, i dunno
<tcn> abi, seen fare
<tcn> abi: seen tcn
<abi> tcn was last seen saying something on IRC Sun Apr 25 12:57:44 1999
<tcn> abi: seen fare
<tcn> abi: seen hcf
<abi> hcf was last seen saying something on IRC Sun Apr 25 12:45:59 1999
<tcn> abi: seen Tril
<abi> Tril was last seen saying something on IRC Sat Apr 24 21:50:20 1999
<tcn> abi: seen tril
<tcn> abi: seen abi
<tcn> haha
-:- hcf_ [nef@me-portland-us113.javanet.com] has joined #tunes
-:- SignOff hcf: #TUNES (Read error to hcf[me-portland-us319.javanet.com]: Connection reset by peer)
-:- hcf_ is now known as hcf
-:- GMOL [gmol@24.66.11.51] has joined #tunes
01:00pm
<tcn> maneesh yadav?
<GMOL> Tis
<GMOL> Have some time to talk?
<tcn> yeah, sure.. did you read my email?
<GMOL> yeah..
<tcn> oops, I should send that to tunes too :)
<GMOL> Yeah..that's ok....here are some of my thoughts..
<GMOL> Tunes is about meta programming, and coming with a bootstrapped environment is a part of it...
<GMOL> But consider all the things that limit you (I know these becuase I too finished my pseudo 386 ukerenel)
<GMOL> As far as I can tell there is no memory management...you going to first write the page initilization code...then you'l want vmem...then GC...believe me you'll never get to all of it
<GMOL> <How bout you respond to that one...I gotta go...can we meet back in ~30min)
<GMOL> >
<GMOL> ?
<tcn> you can try :)
<GMOL> k...c u in 30
<tcn> ok
<tcn> yes there will be vmem and GC.. real soon (for scheme)
-:- binEng [bineng@dialup82-4-14.swipnet.se] has joined #tunes
<tcn> howdy
<abi> hi, tcn
<binEng> hi
<binEng> This book, GEB, is very interesting.
<tcn> yeah
<binEng> D'ya know where more on the same subject can be found?
<tcn> I just started reading Metamagical Themas (mostly a collection of Hofstadter's articles from Scientific American in the early 1980's)
<binEng> oh
<tcn> there's one place :)
<binEng> :)
<tcn> he wrote some other books too, and lots of articles & papers
01:10pm
<tcn> You might try www.ai.mit.edu
<binEng> Anything on the thought of what concious is and how it can be constructed, anywhere?
* Tril/#TUNES is back from the dead. Gone 15 hrs 20 min 26 secs
<binEng> greets, Tril
<Tril> hi, all
<tcn> hey tril
<tcn> just reply to your mail
<tcn> replyING
<Tril> ok
<binEng> tcn: Think I've been there already without finding anything
<Tril> bineng: consciousness is reflection :)
<Tril> (btw, I don't actually believe that...)
<binEng> Tril: yes, it has something to do with that
<tcn> you could say that
<binEng> Ok, so you tell me what it _really_ is about
<Tril> i'm being sarcastic. I hope tunes will show once and for all that sentient machines are impossible.
<binEng> You *hope*? You don't wish sentient machines?
<tcn> we already have sentient people, who needs sentient machines? :)
<Tril> actually, I think being under complete user control is the closest thing to sentience a machine will ever reach.
<tcn> hehe
<tcn> yeah, a computer should be more like an ox than a human :)
<binEng> Tril: Don't you want sentient machines, or don't you think it's possible (or both)?
<Tril> both. It's part of my religious faith.
<Tril> Tcn: yes, but how sentient are they?
<binEng> oh. religion.
<Tril> tcn: tell me when you hit send so I can recheck my mail box.
<tcn> ok
<binEng> Have anyone of you read Brian's paper?
<tcn> maybe we could create a machine that *evolves* into something sentient or semi-sentient
<Tril> I read an older draft
<tcn> but it would be a chaotic process
<Tril> not the new one yet
<tcn> I'm planning to read it, next week.
<binEng> tcn: the idea is not that bad actually... after all, that's how we did.
<binEng> I've had a quick look at the paper. It seems pretty dense.
<binEng> ...not dense in the meaning of basic or something if that's what you thought.
<tcn> binEng: But we can't control it.. if we could, it wouldn't be sentient :)
<binEng> tcn: are you sure about that?
<tcn> If we could make a sentient machine that's smaller/faster/stronger/smarter than us, that could be dangerous
01:20pm
<binEng> I guess so
<tcn> hmm.. people can be brainwashed, and we call them sentient..
<tcn> maybe we're really just semi-sentient :)
<binEng> but can that be a reason not to try to understand how 
<binEng>  work?
<tcn> lemme put it this way:
<binEng> intelligence/anything work?
<binEng> Without doubt, there's degrees of sentience.
<tcn> How did the universe start? Who/what created it? Or did it -- reality, everything around us -- just appear from nowhere?  Is there another universe containing this one?  What contains it? etc...
<tcn> I can't answer that
<tcn> So I chalk it up to God
<tcn> Consciousness may be one of those unsolvable puzzles.
<tcn> Or we might solve it, but not in our lifetimes.
<binEng> Isn't that a romantic view of it?
<binEng> We can't credit some god whenever there's something we can't explain.
<tcn> we can't explain everything though
* Tril/#TUNES becomes aware of being unconscious.
<tcn> haha
<Tril> bin: no we have to figure it out :)
<binEng> guesso
<tcn> Even if we can make a computer conscious, I'm afraid it'll be so complex that we won't be able to figure out the mechanism of consciousness.. and the computer's mechanism will be different from ours.
<binEng> I don't think consciousness is so overly complex. Sure, higher levels of intelligence is probably really nasty, but basic consciousness shouldn't be impossible.
<binEng> What's the line between consciousness and intelligence anyway?
<binEng> What's the difference?
<tcn> i can't answer
01:30pm
<tcn> anyway, I'm gonna finish this email to tril
<binEng> feel free.
<Tril> if I told you, would you believe it?
<binEng> If you told me what? The difference?
<Tril> or whatever
<binEng> I would consider anything you told me, but I can't promise belief.
<tcn> Tril: check you mail
01:40pm
>>> tcn [tcn@cci-209150250086.clarityconnect.net] requested PING 925072916 162790 from #tunes
<Tril> tcn I replied
<tcn> oh ok
<tcn> I think you're saying each object can be a member of only one type, but that type can be a member of several types
<Tril> no, an object can be member of many types. And I didn't say anything about types of types yet. That's not until later.
<binEng> Where would a concious system get its motivation from? How could it feel rewarded?
<binEng> I can't seem to figure that out...
<tcn> yeah, but if a type can be a member of several types, objects can be members of many types by proxy.
<tcn> (I'm concerned about efficiency here :)
<Tril> bin: program something, and try it out! :)
<Tril> tcn: don't get confused between member of a type, and subtype of a type.
<tcn> binEng: food, pleasure, mental illness ;)
<binEng> What language would be best for that?
<tcn> whatever you're best at
<Tril> bin: LISP :)
<Tril> I say that because the AI class I took last quarter is from this big book, and all the programs are in LISP.
<tcn> I wrote an AI in basic once
<binEng> tcn: artificial beeings don't need food. Pleasure is another word for motivation, I guess. Mental illness can't be a motivation.
<tcn> ;)
<binEng> tcn: allow me to doubt...
<binEng> Tril: I don't know LISP. What would make it suitable?
<tcn> binEng: it was a tic-tac-toe AI!
<binEng> tcn: I don't think I'd call that A_I_.
<tcn> hehe
<tcn> it was more like a Genetic algorithm
<Tril> bin: what?
<Tril> what do you mean "make it suitable"
<tcn> binEng: what languages DO you know?
<binEng> The lang needs to be easy/fast to use and easy to monitor.
<Tril> tcn: so I guess you don't want to program this simple exercise...
01:50pm
<binEng> tcn; C(++), Basic, asm... mainly.
* binEng/#tunes is thinking... aren't there other languages he knows?
<Tril> batch
<tcn> Tril: I'm just trying to figure a good way to program it.. more efficiently
<binEng> That doesn't count
<Tril> if it's 4dos it does. sort of.
<Tril> tcn: I don't think you understand it, first of all :)
<tcn> LISP is easy, once you learn it :)  That could take awhile
<Tril> more efficiently than what, I might add? I didn't specify an implementation.
<Tril> bin: you gotta learn a real language.  Try Haskell or scheme
<binEng> some other time
<binEng> Java, maybe
<Tril> my ai class forced me to learn lisp, and I was glad I learned it
<binEng> What did you learn on Ai, then?
<tcn> Tril: I need to convince myself about some things in your spec...
<tcn> Tril: A type *is* an object, right?
<Tril> search, game playing, neural nets, logic programming/inference, and one more thing.. inductive learning
<Tril> tcn: Not yet!  The types and objects are separate creatures, for now.  Later, we'll add bindings so that all the types are accessible AS OBJECTS. But for now they can be just types, not objects.
<binEng> What's inductive learning?
<tcn> Tril: But i've already got objects; it would be easier if types were objects..
<Tril> inductive learning is where you have a bunch of examples (data), and you feed it to a program one at a time, and it "learns" a pattern in the data. That is, it tries to find some logical rules that describe the data, which is shorter than just storing the data itself (it tries to find the simplest rules possible that fit all the examples)
<tcn> Tril: i don't suppose it matters to you if types are objects, anyway
<binEng> ic
<Tril> tcn: you have objects in YOUR system, or are you talking about the objects in my system?
<binEng> No hint on where conciousness comes from?
<tcn> Tril: What will we do with your system? Shouldn't I write some sort of interpreter for it?
<Tril> bin: we just learned the main ideas from the field of AI today. They haven't got that far, yet.  In fact, nobody has any clue.  You will probably want to look for some books where people have written about consciousness.  There's a whole group of people studying that, not necessarily with AI
02:00pm
<Tril> that is, there are some people speculating on the nature of consciousness, and hanging out at conferences together, writing papers, etc. But IMHO they probably have no clue, either.
<tcn> binEng: "philosophy of mind"
<binEng> Tril: Then there _is_ a separation between conciousness and intelligence (AI), at least
<binEng> tcn: is that a book?
<Tril> tcn: some sort of UI, anyway. Maybe a command line inspector for objects/types, or a graphical one if you know how.
<tcn> binEng: No, it's a subject
<tcn> Tril: I'm thinking, we should work it out better before coding it
<tcn> Tril: Or just code what makes sense so far :)
<Tril> tcn: I'd be glad to explain it to yo better before you code it :)
<Tril> OK: An object is a unique ID. The data related to the object is stored entirely by what type(s) the object is "in". Got it?
<Tril> Maybe it will make more sense if I send you the first half of the document..
<binEng> Tril: Is that your OS-type model?
<tcn> Tril: the spec?
<Tril> bin: I'll send you one, too :)
<tcn> Is it http://tunes.org/~dem/tunes/Specifications.html?
<Tril> This is a file which is describing the storage of state in my reflective arch. I'm hoping it is something simple enough someone can begin coding it now. It's the base that goes under what you see in the Specifications.html
<tcn> ok, let's see it :)  This should help..
<tcn> I could always implement it independently of Retro's object system
<Tril> I mailed it to both of you
<Tril> yes, you SHOULD implement it independently of Retro's object system, because it's significantly different.
<Tril> you can use retro objects to implement it, if you want, but don't call them the same.
<tcn> much better
<binEng> Tril; have you mailed me anything?
<Tril> bin: I sent a letter to bineng@tunes.org, yeah
<Tril> hmm, that won't work , will it.
<Tril> (bad address)
<Tril> sorry
<binEng> k
<Tril> there it goes.
02:10pm
<binEng> ...and here it comes
<tcn> that doc helps alot, tril
<Tril> good.
<tcn> Maybe I'll code it in forth :)
<tcn> probably in asm
<binEng> I'd be inclined to call that mail a fragment...
<Tril> bin: you have to fill in the missing parts:)  or ask?
<binEng> With or without emphasizing on 'you'?
<Tril> I think without.  But, I will fill in the missing parts.
<Tril> bin: Have you read the Specifications.html lately?
<tcn> Tril: What's with the cross-products? Is that just lambda-calc notation for passing multiple arguments?
<binEng> k. I'm quite busy with GEB and AI right now
<Tril> oh, that's why you're asking all these questions :)
<binEng> that would make a resonable explanation
<Tril> tcn: Search for "Relations" on a line by itself in Specifications.html.  Read the paragraph that follows.
<Tril> abi, shut up
<binEng> I didn't hear abi?
<tcn> ok
<Tril> she interpreted "Search for" and sent me some msgs.. she said no matches were found :)
<binEng> heh
<tcn> haha
<Tril> i don't think she actually looked in the right file..
<binEng> search for your brain
<binEng> hehe
<binEng> <abi> Looking for your brain:
<binEng> <abi> no matches found for your brain
<tcn> search for the meaning of life
<tcn> she says life is meaningless :)
<Tril> oh i see. "saerch for x" searches for the x in all of her factoids..
<tcn> for her I suppose it is
<tcn> saerch :)
<abi> *sigh*
<Tril> hi hcf
02:20pm
<Tril> I see your mouth moving over there
<Tril> ventriloquist
<abi> pay no attention to the geek behind the curtain
<binEng> abi: how's going?
<abi> no idea, bineng
<binEng> she's totally lost
<tcn> Allright I gotta go do some other things
<binEng> cya
<Tril> ok...
<tcn> I'll work on that code later, tril.. prolly not till next weekend
<tcn> take care
-:- SignOff tcn: #TUNES (tcn has no reason)
<Tril> Making HTTP connection to www.Specifications.com
<Tril> stupid lynx
<binEng> :)
* hcf/#tunes is away 25/04/99 17:23:54 EST : ...
<Tril> abi stats
<abi> Since Thu Apr 22 10:14:19 1999, there have been 50 modifications and 12 questions.  I have been awake for 3 days, 4 hours, 10 minutes, 3 seconds this session, and currently reference 423 factoids.
<binEng> what's a factoid?
<Tril> that's one of her bits of knowledge
<binEng> k
<Tril> each one is a keyword and a definition
<Tril> separted by "is" or "are"
<Tril> abi, each one?
<abi> each one is a keyword and a definition
<Tril> see...
<Tril> that's a factoid
<binEng> I thought... IRC could be a good way of communicating with an AI, since human-type interfacing isn't that easy to implement.
<binEng> Of course, the language used wouldn't be anything like plain English.
<binEng> The simplest way would be to just implant thoughts into its mind. That could actually work, I think.
<Tril> bah. the language should be English.
02:30pm
<binEng> well - how would you teach that to a sentient machine? You can't like point at something and tell the word for it.
<binEng> The means of communication are very limited.
<Tril> I'm going to program English into tunes. not a sentient machine
<binEng> :)
<Tril> speaking of tunes .. these ff3 tunes are great
<Tril> the music is usually the best part of any game.
-:- SignOff GMOL: #TUNES (Leaving)
<binEng> Besides, humans have a natural feeling for language... something that isn't given to machines that easily.
<binEng> Not Quake ;)
<binEng> err, Quake 2
<Tril> haven't heard the Quake NIN track yet.
<Tril> definitely not q2's
<Tril> (haven't heard that either)
<Tril> DOOM, on the other hand..
<binEng> Yeah, DOOM!!
<Tril> I'm gonna port doom to tunes some day..
<binEng> Do that. After you've written Tunes, that is...
02:40pm
<Tril> maybe first thing :)
<Tril> or second. we need a text adventure first. then doom
<Tril> i didnt' post on that games thread, but I do want to make a lot of games with tunes
<binEng> yes... but the system comes first.
02:50pm
-:- Downix [down@survivoronetwothree.ne.mediaone.net] has joined #tunes
<Downix> Hey-lo all
<binEng> hi
<abi> what's up, binEng
<Tril> hi
<Downix> How is Tunes coming?  Anything done yet?
* Tril/#TUNES points Downix at the topic of the channel
<binEng> what topic
<Tril> there's some stuff you can read. The Arrow paper, and some specs I'm working on
* binEng/#tunes sees no topic
<Tril> bineng: look at the title bar of the window for the channel
<binEng> Are you sure there's a topic?
<Tril> what happens if you type /topic
<binEng> #Tunes There isn't a topic.
-:- Pitu [pali@ih-239.arrakis.es] has joined #Tunes
-:- Pitu [pali@ih-239.arrakis.es] has left #Tunes []
<Tril> I'm not saying he's asking something that is off topic, and to shut his mouth, I'm saying there are some URLs in the channel topic that he can read.
<Tril> bin, well, I guess you are desynched?
<binEng> must be
<binEng> All I know is that I don't have any URLs in _my_ topic line
-:- Tril has changed the topic on channel #Tunes to: http://tunes.org (Reflective Computing) http://tunes.org/papers/Arrow/ (Arrow paper available!) http://tunes.org/~dem/tunes/Specifications.html (Tril's draft specs)
<Tril> is that better?
<binEng> there we go
<binEng> there you've got a little more info on those specs
03:10pm
<binEng> Why never more than one argument to a function?
<Tril> well, use a structure to contain more than one argument, and use the structure as the one argument. 
<Tril> I'm not sure why, it just seems simpler.
<binEng> maybe... maybe not
<Tril> You have to describe what the function does for every possible combination of arguments. It is easier if each combination is represented by a separate object.
<Tril> You have to have a set of objects for the function to take as input, otherwise it's not a function.
<binEng> hmm
03:20pm
<binEng> I don't think grouping the arguments would solve or even ease the problem of acceptable inputs... unless you define some group which can be used for many functions.
<binEng> Do you have any thoughts on when you'll round this system up and consider it final?
<Tril> a group which can be used for many functions- that's what a type is. A set of objects that can be used as the input for any number of functions.
<Tril> and the output of functions.
<Tril> what system?  The specs?
<binEng> yah
03:30pm
<binEng> A thought - is {true, false} really == {yes, no}, {zero, one}, etc?
<Tril> not equal, but isomorphic.
<binEng> better
<Tril> two domains are isomorphic if they have the same structure.
<Tril> or you know what it is
<binEng> kinda, yes
<binEng> I learned that from GEB
<Tril> I dunno when the spec will be done. I'm adding more headings in it now.
<binEng> Just wondering if you wanted to finish this soon, or if you rather would keep thinking of it.
<binEng> Your specs reaches into the internal works of programs, don't they?
<Tril> I want to finish it soon so I can start working on tunes
<binEng> ok
<Tril> in what way
<binEng> I should be going now... TTYL!
<Tril> ok, ask me later
-:- binEng [bineng@dialup82-4-14.swipnet.se] has left #tunes []
03:40pm
-:- SignOff hcf: #TUNES (Leaving)
-:- SignOff Tril: #TUNES (Ping timeout for Tril[sloth.wcug.wwu.edu])
-:- Tril [dem@sloth.wcug.wwu.edu] has joined #TUNES
-:- mode/#Tunes [+o Tril] by ChanServ
* Tril/#TUNES  is away: (afk) [BX-MsgLog Off]
-:- Fare [rideau@quatramaran.ens.fr] has joined #Tunes
<Fare> gakuk
<Fare> not here for long
<Fare> dem: here?
<Fare> Well, have a nice day!
-:- SignOff Fare: #TUNES (Connection reset by pear)
06:10pm
-:- SignOff Downix: #TUNES (Ping timeout for Downix[survivoronetwothree.ne.mediaone.net])
-:- AlonzoTG [Alonzo@client-151-200-120-23.bellatlantic.net] has joined #tunes
-:- Downix [down@survivoronetwothree.ne.mediaone.net] has joined #tunes
<Downix> Hey
<AlonzoTG> y0
-:- SignOff Downix: #TUNES (Ping timeout for Downix[survivoronetwothree.ne.mediaone.net])
08:00pm
-:- Downix [down@survivoronetwothree.ne.mediaone.net] has joined #tunes
-:- SignOff AlonzoTG: #TUNES (Have Nice Day :))
-:- SignOff Downix: #TUNES ( AtlantiS (v1 . 3b): lost in time ... [ 6 . 02 . 97 ])
-:- vladmeir [morbidm@216.88.72.130] has joined #Tunes
-:- vladmeir [morbidm@216.88.72.130] has left #Tunes []
-:- SignOff _QZ: #TUNES (BRiX [http://www.qzx.com/brix] :: sleep)
[msg(TUNES)] newlog 1999.0426
IRC log ended Mon Apr 26 00:00:00 1999