←2007-08-04 2007-08-05 2007-08-06→ ↑2007 ↑all
00:00:18 <ehird`> xor, are you a new generation of irp?
00:00:23 <ehird`> xor: 1 2
00:22:28 -!- sebbu has quit ("@+").
00:39:11 -!- ehird` has quit (Read error: 104 (Connection reset by peer)).
01:03:35 -!- poiuy_qwert has joined.
01:04:13 <pikhq> Hmm. . .
01:04:19 <pikhq> A new person for our insanity to rub off on?
01:06:04 -!- peek_you has joined.
01:06:18 <pikhq> . . . Or someone with more than one IRC client.
01:22:39 -!- poiuy_qwert has quit (Read error: 110 (Connection timed out)).
01:26:25 <pikhq> ,[.,]
01:29:26 -!- peek_you has changed nick to poiuy_qwert.
01:29:50 -!- ihope has quit (Read error: 110 (Connection timed out)).
01:44:02 -!- peek_you has joined.
01:56:24 -!- poiuy_qwert has quit (Read error: 110 (Connection timed out)).
02:29:30 -!- peek_you has quit (Read error: 110 (Connection timed out)).
02:45:37 <pikhq> [set ::pikhq::ACT_score 33]
03:17:36 <xor> O.o
03:17:45 <xor> is the act out of 1600 points?
03:18:24 <pikhq> 36.
03:19:06 <xor> oh ok
03:33:41 -!- GreaseMonkey has joined.
03:35:46 -!- Shel2476 has joined.
03:38:53 <pikhq> 'Lo.
03:51:00 -!- GreaseMonkey has quit ("rebooting to remove defunct modules from system, brb").
03:54:09 -!- GreaseMonkey has joined.
03:58:18 -!- sekhmet has quit (Read error: 110 (Connection timed out)).
06:46:02 -!- RodgerTheGreat has quit.
06:52:43 -!- cherez has quit ("Leaving.").
06:54:00 -!- cherez has joined.
06:54:45 -!- cherez has quit (Client Quit).
06:59:02 -!- cherez has joined.
07:18:19 -!- Shel2476 has quit (Read error: 104 (Connection reset by peer)).
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:12:33 -!- sebbu has joined.
08:15:42 -!- jix has joined.
10:08:18 -!- oerjan has joined.
10:15:58 -!- GreaseMonkey has quit ("gnight").
10:50:57 -!- sebbu2 has joined.
10:55:06 -!- sebbu has quit (Read error: 110 (Connection timed out)).
10:55:06 -!- sebbu2 has changed nick to sebbu.
11:55:45 -!- tokigun has joined.
13:28:59 -!- oerjan has quit ("Lunch").
14:50:15 -!- RodgerTheGreat has joined.
14:50:35 <RodgerTheGreat> hi everyone
15:23:27 <SimonRC> hi
15:25:06 <RodgerTheGreat> hello, SimonRC
15:27:26 -!- sekhmet has joined.
15:30:14 <SimonRC> RodgerTheGreat: did you read the newest xkcd?
15:30:31 <RodgerTheGreat> haha- yes, a nightmare indeed
15:30:41 -!- sebbu2 has joined.
15:31:23 -!- ehird` has joined.
15:31:53 <SimonRC> RodgerTheGreat: The forums contain two sorts of people: those who think it would be great and those who think it would be terrible.
15:32:04 <SimonRC> I am definitely the latter type
15:32:20 <RodgerTheGreat> yeah- I flipped through the individual comic thread
15:32:31 <RodgerTheGreat> I am in agreement with you sir
15:32:53 <SimonRC> I alarmed myself by realising I was thinking like the Catholic chruch confronting geocentrism.
15:33:42 <RodgerTheGreat> I'm sure i would be comparable to how religious people would feel if there was a clear, instantly understandable and recognizable proof for the non-existence of god.
15:33:44 <ehird`> what would?
15:33:49 <ehird`> what would be terrible
15:33:57 <RodgerTheGreat> (not featuring a puff of logic)
15:34:03 <ehird`> where what how
15:34:07 <RodgerTheGreat> http://www.xkcd.com/
15:34:23 <ehird`> oh, the thread discussing science suddenly being discovered not to work
15:34:25 <ehird`> and what it would be like?
15:34:39 <ehird`> the discussion thread for that comic
15:34:42 <RodgerTheGreat> yes
15:34:57 <ehird`> i think it would be awesome :p
15:35:03 <RodgerTheGreat> why?
15:35:14 <RodgerTheGreat> explain your reasoning
15:36:37 <ehird`> i have none
15:36:42 <SimonRC> Slartibartfast: "I would rather be happy that right, any day." || Arthur: "And are you?" || S.: "No, that's where it all falls down I'm afraid."
15:36:50 <SimonRC> *than
15:36:52 <ehird`> i'd just like to see science turn out to be utterly and completely wrong
15:37:06 <ehird`> :p
15:37:31 <SimonRC> how do English people distinguish between the two meanings of "science"?
15:37:52 <ehird`> what are the two meanings
15:37:52 <RodgerTheGreat> but consider for a moment that science is responsible for the level of comfort and safety you enjoy in our modern age
15:38:02 <ehird`> RodgerTheGreat, oh yes, i like science and all that
15:38:05 <RodgerTheGreat> I think it's Science versus science
15:38:07 <SimonRC> ehird`: the knowledge and the method
15:38:08 <ehird`> i'd just like it to be completely wrong
15:38:11 <ehird`> SimonRC, ah
15:38:19 <SimonRC> ehird`: which one do you want to be wrong
15:38:27 <ehird`> SimonRC, hmm - both, for the hell of it
15:38:28 <ehird`> :-)
15:38:32 <RodgerTheGreat> Science would be the method, science would be the community (in my thinking)
15:38:47 <ehird`> i would say scientific method
15:38:49 <SimonRC> If the Method is wrong, then the world will be very strange indeed.
15:38:50 <ehird`> vs science
15:38:59 <ehird`> SimonRC, EXACTLY!
15:39:37 <RodgerTheGreat> if the scientific method is flawed, and the universe *isn't* comprehensible, it'd be the most depressing thing I could think of
15:39:39 <SimonRC> For a start, you would have little guarantee that much would be the same from one day to the next
15:39:45 <SimonRC> RodgerTheGreat: indeed
15:40:01 <RodgerTheGreat> to know that our reach is ultimately limited would be crushing
15:40:04 <ehird`> i'd think it was hilarious
15:40:04 <SimonRC> Most stuff you like wouldn;t work
15:40:05 <ehird`> :D
15:40:24 <SimonRC> e.g. food would not work consistantly
15:41:05 <ehird`> well it >does<
15:41:25 <ehird`> so even if it turned out to be wrong, we could live in the knowledge that for several billions of years nothing has actually changed much
15:41:29 <SimonRC> If the knowledge is wrong but not the method, then we can cope rather better.
15:41:31 <RodgerTheGreat> so >does< the scientific method
15:41:36 <SimonRC> It's called "progress"
15:41:40 <oklopol> </irl> <irc>
15:41:44 <oklopol> mörning all
15:41:48 <RodgerTheGreat> hi, oklopol
15:41:57 <RodgerTheGreat> nice o-umlaut you have there
15:42:08 <SimonRC> oklopol: yay! You can produce the ö character correctly, unlike everyone else.
15:42:18 <SimonRC> UTF-8 FTW
15:42:23 <ehird`> áéíóú
15:42:36 <RodgerTheGreat> Ö RLY?
15:42:39 <oklopol> SimonRC: i kinda have it in my keyboard
15:42:42 <SimonRC> ¿Qué?
15:42:44 <oklopol> :P
15:42:44 <SimonRC> :-)
15:42:46 <ehird`> Hmm
15:42:50 <RodgerTheGreat> I has an option key
15:42:50 <ehird`> ú
15:42:52 <ehird`> Hmm
15:42:54 <ehird`> ú
15:42:57 <ehird`> LOWER COMMA!
15:43:00 <SimonRC> oklopol: often people use a non-good character set
15:43:09 <oklopol> those bastards
15:43:18 <RodgerTheGreat> haha
15:43:24 <SimonRC> e.g. 8-bit character-sets are non-good
15:43:25 <oklopol> uh, typing is so much easier than speaking
15:43:55 <SimonRC> indeed. For example, pronounce: -6~8e5-y;e57)-p'0'-cD_c7#=-'^C_@6
15:44:54 <RodgerTheGreat> which is made even worse because some people don't understand words like "dash" or "hash" or "carat" properly
15:45:05 <RodgerTheGreat> not to mention "tilde"
15:46:22 <SimonRC> ITYM "caret"
15:46:42 * SimonRC recalls calling that "tidal".
15:46:43 <SimonRC> :-)
15:47:00 <RodgerTheGreat> ah, fuck- you're right
15:47:04 <RodgerTheGreat> thank you, SimonRC
15:47:20 * RodgerTheGreat begins cutting an incorrect piece of data out of his brain
15:47:59 -!- sebbu has quit (Nick collision from services.).
15:48:10 <SimonRC> ouch
15:48:41 <RodgerTheGreat> making fewer mistakes is worth the pain. It doesn't even hurt once you get through the skull, anyway
15:51:44 <oklopol> "Archos: screw religion, I'm trying to make a silencer!" <<< this is cool
15:51:51 <RodgerTheGreat> lol
15:52:27 <RodgerTheGreat> Archos was a lot of fun once he'd had a few days to soak up conversations
15:53:04 <SimonRC> "was"?
15:53:18 <RodgerTheGreat> well, I mean, he was eventually reset
15:53:33 <RodgerTheGreat> I never made a system to save state, so that version of Archos "died"
15:53:48 <ehird`> is archos source code open?
15:54:49 <RodgerTheGreat> I guess I could dredge it up, but I built it around the JMegaHal system: http://www.jibble.org/jmegahal/
15:55:08 <RodgerTheGreat> you'd probably have more fun coding something similar on your own
15:55:33 <ehird`> ah megahal
15:55:35 <ehird`> that markov chain thing
15:55:44 <ehird`> but markov chains are really shitty, aren't they?
15:55:48 <ehird`> i mean, they never produce coherent stuff, barely
15:55:51 <ehird`> mostly just nonsense?
15:56:18 <RodgerTheGreat> well, the main thing I was doing with Archos was working on filtering the input and output for the chain
15:56:55 <RodgerTheGreat> I achieved a fair amount of success making it more coherent
15:57:07 <RodgerTheGreat> but it's ultimately still kinda random
16:00:23 <oklopol> english is too hard... why not make your own esoteric natural language for it to speak
16:00:55 <ehird`> maybe lojban#
16:01:00 <ehird`> lojban should be easy to generate
16:01:03 <SimonRC> better idea
16:01:10 <SimonRC> better idea: ask on #conlang
16:01:13 <ehird`> :p
16:01:13 <oklopol> yes, but less fun than making your own, naturally
16:01:23 <ehird`> but anyway
16:01:27 <ehird`> mvldo is meant to fool normal people
16:01:30 <ehird`> most people on irc skeep english
16:01:36 -!- ihope has joined.
16:04:28 <ehird`> skeep
16:04:30 <ehird`> i like that word
16:04:31 <ehird`> skeep!
16:05:02 <oklopol> me too
16:05:11 <oklopol> i assumed that was on purpose
16:06:10 <ehird`> i skeep english
16:07:11 <oklopol> skeeping english with a great pseed
16:07:19 <oklopol> i aslo like pseed
16:07:28 <oklopol> some words are better that way
16:14:42 <ihope> I disargee.
16:17:11 <SimonRC> "pseed" sounds like a unix program
16:17:19 <SimonRC> maybe a system call
16:17:22 <ehird`> this has the good side of elinimating tpyos
16:17:49 <ehird`> i skeep english with a great pseed and with elinimated typos
16:32:13 <ehird`> RodgerTheGreat, do you think markov chains are part of the future of AI?
16:32:17 <ehird`> or are they not worth it?
16:32:23 <RodgerTheGreat> dunno
16:32:37 <RodgerTheGreat> they're cheap and easy to make, and have a good return on invested coding time
16:34:33 <ehird`> but are they worth it for the downsides?
16:35:47 -!- sebbu2 has changed nick to sebbu.
16:38:30 <ihope> I think Markov chains as they are are not very useful.
16:38:45 <ihope> Not very useful for "the future of AI", that is.
16:39:18 <ihope> By the way, text-to-speech is silly. How, exactly, is "I think that green goo should not be eaten" pronounced?
16:39:25 <ihope> And, for that matter, what, exactly, does it mean?
16:39:36 <ehird`> what good alternatives are there for sentence construction?
16:39:37 <ehird`> to markov chains
16:40:04 <ihope> Markov chains assume that language is... flat.
16:40:23 <ihope> That each construct is merely a certain type of word followed by some other construct.
16:41:01 <ehird`> there must be a way to generalize it to a parsetree
16:41:07 <ihope> Indeed.
16:41:12 <ihope> Can't be too hard.
16:41:15 <ehird`> the first step would be lojban
16:41:19 <ehird`> as lojban can be unambigiously parsed
16:41:22 <ihope> First, though, come up with a good way to describe English syntax.
16:41:29 <ehird`> there are already many ways
16:41:47 <ihope> I should look at Lojban. Does it have things like emphasis?
16:41:55 <ehird`> in what way?
16:42:04 <ehird`> it's a full language that can express anything english or anything else can
16:42:07 <ehird`> and it's unambigious
16:42:14 <ehird`> and without exceptions to rules
16:42:25 <ehird`> lojban.org
16:44:10 <ihope> So there's a Lojban sentence meaning "What in the world IS 'pink' ice cream"?
16:46:42 <ehird`> Um, yes, I'd imagine so
16:47:03 <ehird`> Or at least an approximation
16:47:18 <ehird`> ask in #lojban
16:50:54 <ehird`> :)
19:07:42 -!- oerjan has joined.
19:42:36 <pikhq> Lojban is easy to parse, for that matter.
19:42:52 <pikhq> Step one of writing the parser: Download the Lojban Reference Grammar.
19:42:56 <pikhq> Step two: compile it.
19:43:47 <ehird`> yes
19:43:55 <ehird`> which is why it would be a perfect target for generalized markov chains
19:44:04 <pikhq> Yeah.
19:44:14 <ehird`> i think a refined version of parse-tree-generalized markov chains is the future of sentence construction
19:44:15 <ehird`> imo
19:44:35 <pikhq> They're perfect for any Turing-test-relevant AI, for that matter. . .
19:44:52 <ehird`> what is?
19:44:58 <ehird`> parse-tree-generalized markov chains?
19:45:05 <pikhq> Err. No. . . .
19:45:08 <ehird`> oh
19:45:12 <pikhq> s/they're/it's/
19:45:26 <ehird`> ah
19:45:27 <ehird`> lojban
19:45:31 <ehird`> well yeah
19:45:33 <pikhq> Lojban, or any other language with a regular grammar, is perfect for that. . .
19:45:35 <ehird`> but englihs generation isn't that hard, you know?
19:45:38 <ehird`> *english
19:45:44 <ehird`> i mean, lesser than other problems in AI
19:45:45 <ehird`> certainly
19:54:39 <SimonRC> <amazon>People who investigsated `Lojban' also investigated `Ithkuil'.</amazon>
20:03:23 -!- jix has quit (Nick collision from services.).
20:03:37 -!- jix__ has joined.
20:08:52 -!- ehird` has quit (Read error: 104 (Connection reset by peer)).
20:14:59 <ihope> Lojban, Ithkuil and Ilaksh.
20:15:37 <pikhq> Is Ithkuil the language where you say "Cthulhu rygel! Cthulhu rygel!"?
20:15:44 <pikhq> :p
20:16:03 <oerjan> i don't think so.
20:16:26 <oerjan> that might be Old R'lyehan, or something.
20:16:38 <ihope> It's the one where "Oumpeá äx’ääuktëx" means "On the contrary, I think it may turn out that this rugged mountain range trails off at some point".
20:17:42 <ihope> "Äx’ääuktëx" is really quite the word.
20:18:44 <ihope> The whole sentence, spoken: http://home.inreach.com/sl2120/Sound_Files/Intro-1.mp3
20:19:36 <ihope> Hmm, I left out a word.
20:20:46 <ihope> Rather, a letter.
20:21:07 <ihope> That should be "äx’ääluktëx", except with a cedilla below the l.
20:21:31 <SimonRC> Ilaksh is a newer, mor pronouncible version
20:21:35 <SimonRC> with extra tones!
20:22:20 <ihope> More pronounceable indeed.
20:22:30 <SimonRC> whatever
20:22:47 -!- ehird` has joined.
20:23:37 <ihope> Though it's rather incomplete, no?
20:36:53 <oklopol> hmm... is there a nice ithkuil tutorial? if it has a thousand words, that's like insanely trivial to memorize
20:37:05 <oklopol> hmm... i think someone here already linked me one
20:37:08 <oklopol> some time ago
20:38:04 <SimonRC> depends on what you mean by "word"
20:39:27 <oklopol> a token consisting of letters the meaning of which i have to learn
20:39:38 <SimonRC> hmm
20:39:48 <oklopol> where 'which' refers to 'token'
20:39:57 <SimonRC> In that case it has thousands and thousands of words
20:40:14 <oklopol> yes, but only 1000 or so primitives
20:40:32 <SimonRC> but some of the derivations from those primitives is not obvious
20:40:53 <SimonRC> lietk which way round "man" and "woman" are
20:41:22 <oklopol> anyway, i assume it doesn't have as many redundant synonyms as most languages
20:41:42 <oklopol> i'm a pretty assumptive person
20:41:50 <oklopol> i don't know squat about the language
20:44:45 <SimonRC> It has *LOTS* of inflections.
20:45:18 <SimonRC> maybe 10000 - 100000 forms of each word
20:47:42 <oerjan> well, once you add several types of inflections you generally multiply their number together, so 10000 - 100000 is not quite as impressive as it sounds. There might be natural languages with the same.
20:49:34 <ihope> Anti-anti-anti-anti-anti-anti-anti-anti-anti-anti-anti-anti-anti-antidote!
20:49:48 <ihope> Kingdom of Loathing taken... farther/further.
20:53:35 -!- RedDak has joined.
21:02:41 <ehird`> No person is hitherto known to be able to speak Ithkuil; its creator, for one, does not: “I don't speak Ithkuil, never have, never will, never claimed to.”
21:02:48 <ehird`> Ithkuil (Iţkuîl) is an outstandingly complicated human language constructed by the American linguist John Quijada between 1978 and 2004.
21:02:52 <ehird`> i don't think a tutorial is going to be very likely
21:03:05 <SimonRC> yup
21:03:11 <lament> sounds boring
21:03:31 <ehird`> as far as i can tell
21:03:34 <lament> it's easy to make complicated things
21:03:35 <ehird`> nobody has even said "hello world" in it
21:03:43 <lament> making simple things is hard :)
21:03:52 <ehird`> RodgerTheGreat, would a neural network be useful for an ai bot?
21:03:56 <ehird`> combined with markov chains, maybe ?
21:04:04 <RodgerTheGreat> maybe
21:04:07 <xor> ha
21:04:08 <RodgerTheGreat> that could be interesting
21:04:12 <oklopol> err... so no one can speak it at all? i thought no one speaks it fluently
21:04:15 <xor> i think my modem was overheating
21:04:25 <ehird`> oklopol, nobody at all
21:04:27 <ehird`> by the looks of it
21:04:33 <oklopol> xor = bsmntbombdood!
21:04:42 <xor> good job
21:04:46 <ehird`> xor: 1 2
21:04:50 <oklopol> you and your poor xor encryption
21:04:51 <ehird`> your irp component still doesn't work :p
21:04:56 <xor> it was dropping connections every 2 or 3 seconds
21:05:06 <ehird`> RodgerTheGreat, I think it could help in sentence construction
21:05:18 <oklopol> ehird`: there's an example of it spoken in... ihope's link
21:05:28 <ehird`> which link
21:05:29 <oklopol> @ wikipedia that is
21:05:40 <RodgerTheGreat> or potentially in allowing the bot to "learn" more in terms of high-level sentence construction
21:05:46 <oklopol> http://home.inreach.com/sl2120/Sound_Files/Intro-1.mp3
21:05:50 <xor> markov chains!
21:07:13 <ehird`> markov chains have disadvantages
21:07:21 <ehird`> now, if you combined markov chains and neural networks
21:07:27 <ehird`> plus some evolutionary technique
21:07:30 <ehird`> that should produce good stuff
21:07:52 <RodgerTheGreat> as always, in theory
21:08:06 <ehird`> maybe with a teaching backdoor into a bot
21:08:07 <ehird`> so like
21:08:10 <ehird`> <user> hello, bot!
21:08:16 <ehird`> <bot> oojamaflip gazunks
21:08:28 <ehird`> <user> bot: !no hi user
21:08:29 <RodgerTheGreat> I imagine this might be a bit like my memory optimization algorithm for PEBBLE. Brilliant in concept, miserable failure in execution.
21:08:45 <ehird`> !no X meaning "your last response to me was wrong, it should have been X"
21:08:48 <ehird`> and indeed
21:08:52 <ehird`> but the only way to figure out is to try
21:08:55 <ihope> Teaching backdoor... I like that.
21:09:16 <ihope> What's this PEBBLE thing all about?
21:09:18 <pikhq> Of course, I think your bit about *sequential access* could be much more useful.
21:09:29 <ehird`> pebble is a high level language compiling to brainfuck
21:09:30 <ehird`> pikhq's.
21:09:46 <oerjan> well, if your evolutionary technique includes simulating a small planet, you might get somewhere.
21:10:03 <oerjan> in just a few billion years
21:10:37 <ehird`> no, no
21:10:41 <ehird`> response evolution
21:10:45 <ehird`> not organism evolution
21:10:46 <ehird`> sheesh
21:10:59 <ehird`> combined with survival of the fittest
21:11:20 <oerjan> i'd say organism evolution has a better track record - at least one success
21:11:25 <oklopol> yeah, oerjan, that was very stupid of your, trying to make a joke
21:11:29 <oklopol> *you
21:11:37 <ehird`> "ok, this response X does badly, but this other one Y does well, and a part of the response X is also in another response which does badly. So, replace that part of X with the corresponding one in Y"
21:11:39 <ehird`> something like that
21:11:40 <ihope> What counts as a success when it comes to response evolution?
21:11:46 <ehird`> oklopol, :p
21:12:11 <ihope> Creation of intelligence?
21:12:32 <oklopol> ihope: i guess success is the lack of !no
21:12:34 <oklopol> or something
21:12:46 <ihope> !no isn't necessary, surely.
21:13:03 <ihope> I want non-Terran intelligence.
21:13:14 <oklopol> well, there has to be some primitive that indicates success.
21:13:20 <ihope> Too bad it's impossible to create.
21:13:35 <ehird`> honestly, the only way to test all of this is to implement it
21:13:48 <ehird`> who codes java and wants push rights to the mvldo hg repo :)
21:14:38 <ihope> What's an hg repo?
21:14:51 <ehird`> a code repository of the mercurial format
21:14:55 <ihope> Or a hg repo, be it so.
21:14:56 <oklopol> mvldo hg repo huldv grps mdfmd ofol
21:15:00 * pikhq contemplates adding sequential access to PEBBLE. . .
21:15:01 <ehird`> hg/mercurial = like svn, but distrobuted
21:15:02 <oklopol> i can make up words too
21:15:10 <ehird`> distrobuted = there's not one central repository.
21:15:16 <ehird`> every single copy contains the entire repo history
21:15:18 <ihope> oklopol: y's cn ym!
21:15:21 <ehird`> all the commiting is local
21:15:28 <ehird`> and you share changes by "pushing" your updates somewhere
21:15:30 <ehird`> it's great
21:15:39 <ehird`> http://www.selenic.com/mercurial/
21:15:42 <ehird`> it's really easy to use
21:18:14 <ehird`> but anyway, yeah
21:18:17 <ehird`> push rights to mvldo's
21:18:35 <oerjan> mercurial is non-terran, isn't it? might get somewhere there.
21:19:40 <ihope> If it's not Terran, then what is it?
21:21:42 <ehird`> but anyway, if someone wants push rights and knows java etc. just shout
21:21:59 <oklopol> shout
21:22:00 <ehird`> then, i guess, some rough design documents and little experiments
21:22:03 <oklopol> let it all out
21:22:03 <ehird`> then, the full stuff
21:22:13 <oklopol> these are the things we can do without
21:22:14 <oklopol> ...
21:22:15 <oklopol> err
21:22:19 * oerjan is shocked that ihope doesn't know about the planet Mercury.
21:22:41 <oklopol> come on
21:22:46 <oklopol> i'm a'talking to your
21:22:47 <ihope> Mercurial is Mercury?
21:22:48 <oklopol> come on
21:23:00 <oklopol> *you
21:23:16 * oerjan is shocked that ihope doesn't know how nouns may be turned into adjectives.
21:24:00 <ihope> Mercury has intelligence?
21:24:23 <oerjan> well, it might, if mercurial gets somewhere.
21:25:00 * oerjan realizes his joke must be _really_ awful if it requires this much explanation.
21:25:17 <oklopol> the best jokes are the ones no one ge4ts
21:25:21 <oklopol> *-4
21:27:12 <ihope> Indeed.
21:27:14 <ihope> I hate irony.
21:27:16 <ihope> (joke above)
21:28:09 <xor> ihope xor ihope
21:28:46 <ihope> Not in scope: `ihope'
21:29:24 <oklopol> xor: i think /me would allow a nice prefix way to to that
21:29:38 * ihope xor ihope
21:29:48 <ihope> Or infix.
21:29:49 <oklopol> well, that too
21:30:05 <ihope> Too bad prefix is sort of superior to infix.
21:30:26 <oklopol> prefix & postfix are, yeah
21:30:43 <xor> sexp!
21:31:08 <oklopol> really, xor, you are just using that term for the shock value
21:31:49 <xor> shock value?
21:32:22 <ihope> It's like sex with a p on the end.
21:32:36 <ehird`> no, the real way is irp-style
21:32:38 <ehird`> xor: 1 2
21:32:47 <ehird`> interpret, you, damnit
21:32:50 <ehird`> fine
21:32:55 <ehird`> i'll be an irp execution station
21:32:58 -!- ehird` has changed nick to and.
21:33:08 <xor> ((lambda (ehird) (ehird ehird)) (lambda (ehird) (ehird ehird)))
21:33:12 <and> someone perform an operation on me damnit"
21:33:14 <and> and: x y
21:33:26 <oklopol> :P
21:33:30 <oklopol> and: ihope oklopol
21:33:32 <and> or and: and: x y and: z a or something
21:33:36 <and> oklopol
21:33:58 <ihope> and: and: x z and: y z
21:34:00 <oklopol> i thought you'd give like a percentage
21:34:09 <and> z
21:34:12 -!- and has changed nick to ifte.
21:34:16 <ihope> ifte/
21:34:19 <ifte> ifte: cond true false
21:34:20 <ihope> Er, ?
21:34:22 <ifte> if then else
21:34:30 <ifte> someone change nicks to and
21:34:39 <ifte> quick =p
21:34:42 <ihope> Um...
21:34:43 -!- ihope has changed nick to and.
21:34:56 <and> No's mine but... wait.
21:35:01 <ifte> ok, whenever you see "and: a b" if a is false return false otherwise return b.
21:35:07 <ifte> and to return
21:35:17 <ifte> say "whoever sent the message containing the and: retval"
21:35:18 <ifte> ok
21:35:19 <and> Er, wait, shouldn't we be using CPS?
21:35:24 <and> Um, wait.
21:35:27 <and> Lemme read that.
21:35:37 <xor> rpn is was better than unparenthized prefix
21:35:44 <ifte> now, someone say "ifte: and: false true true false"
21:35:51 * oerjan wonders what is going on
21:36:07 <oerjan> ifte: and: false true true false
21:36:39 <ifte> CALCULATION IN PROGRESS...
21:36:48 <and> How slow.
21:36:54 <ifte> hey, it's your turn
21:36:57 <and> Oh?
21:36:58 <ifte> you've recieved and: false true
21:37:02 <ifte> and you have to return to ifte
21:37:07 <and> ifte: false
21:37:08 <ifte> ok, say that again oerjan :P
21:37:27 <ifte> WAITING FOR INPUT...
21:37:33 <oerjan> ifte: i thought you would have to pass it on
21:37:38 <oerjan> ifte: and: false true true false
21:37:39 <ifte> ah, yes, true
21:37:43 <ifte> and: false true
21:37:47 <and> ifte: false
21:37:53 <ifte> oerjan: false
21:37:55 -!- ifte has changed nick to ehird`.
21:37:59 <and> That was fun.
21:37:59 <xor> lol
21:38:01 -!- and has changed nick to ihope.
21:38:02 <ihope> Maybe.
21:38:04 <ehird`> hooray! irp generation 2 has its foundings!
21:38:09 <ehird`> an infix-based message-passing language!
21:38:18 <xor> delagated computation
21:38:23 <ehird`> yes, whatever
21:38:24 <ehird`> :p
21:38:37 <ehird`> of course, lots of calculations would include random stuff
21:38:42 <ehird`> like if ifte finds a word it doesn't know
21:38:43 <ehird`> like and:
21:38:46 <ehird`> it'll have to say
21:38:50 <ehird`> and: paramcount
21:38:57 <ehird`> or similar, so it knows how many places to look ahead
21:38:59 <ehird`> but still :p
21:39:21 <xor> sexp is better
21:39:22 <ehird`> notably, this irp generation can have computer components and human components working together
21:39:32 <ehird`> xor: yes but sexp doesn't look like irc message
21:39:36 <ehird`> and: 1 2 looks like an irc message.
21:39:39 <ehird`> it is directed to and
21:40:20 <oerjan> there is a problem in that each operator must know the argument count of the others
21:40:27 <ehird`> read above
21:40:34 <ehird`> "operator: paramcount" or similar would be used
21:40:40 <oerjan> ah
21:40:41 <ehird`> then operator is expected to return the parameter count
21:40:49 <ehird`> then it could be stored in RAM, or something
21:41:07 <ehird`> when an operator changes paramcount
21:41:21 <ehird`> it must deny all further operations with an error, stating that you must ask paramcount again before it will accept
21:41:26 <ehird`> therefore, changes in paramcount propagate
21:43:46 <ehird`> http://nonlogic.org/dump/text/1186346246.html two example irp sessions
21:43:59 * SimonRC reads up
21:44:03 <ehird`> note that both could have the non-usr operators as human, program, or mixed
21:44:06 <ehird`> it'd work fine
21:44:22 <ehird`> err
21:44:25 <ehird`> get should have get replying
21:44:25 <ehird`> not set
21:44:29 <ehird`> :)
21:44:37 <ehird`> obviously get/set would have to be the same person/program
21:45:36 <ehird`> and with at least one human operator involved somewhere in an expression, if e.g. 99 bottles of beer was being executed, that human user could give an error and stop it
21:47:14 <SimonRC> (Depending on which review you read, darcs does distributed version control better than Mercury. Darcs uses branches for many purposes.)
21:47:25 <xor> the problem is that each operator has to be able to parse everything
21:48:28 <ehird`> SimonRC, everything is a branch in mercurial
21:48:36 <ehird`> xor, so what?
21:48:49 <xor> so, code is duplicated everywhere
21:49:14 <ehird`> and..
21:49:30 <xor> each operator has large amounts of the same code
21:49:32 <xor> which is stupid
21:49:34 <SimonRC> ehird`: Have a look at darcs then
21:50:08 <SimonRC> in darcs, a source tree is just a set of patches applied to the empty tree
21:51:28 <ehird`> same with hg
21:51:59 <SimonRC> Well, darcs was written by a quantum mechanic
21:53:29 <ehird`> i am aware
21:53:32 <ehird`> i used to use darcs.
21:53:37 <ehird`> but it is glacier-style slow
21:53:40 <SimonRC> wht is the difference?
21:53:40 <pikhq> ehird`: That'd be IRP version 3, BTW.
21:53:49 <ehird`> pikhq, ?
21:53:53 <SimonRC> darcs is only slow in a few cases
21:54:05 <pikhq> IRP version 2 was developed in #irp, with a bot assigning programs to programmers.
21:54:07 <ehird`> no - many
21:54:08 <SimonRC> Some of the algorithms have SUPRISE exponential running times
21:54:19 <ehird`> pikhq, sounds bad :)
21:54:21 <pikhq> It also had a syntax to seperate programs & comments.
21:54:31 <ehird`> #irp3, gogogo
21:54:34 <ehird`> :D
21:54:42 <pikhq> [Please do foo, bar, and baz]
21:54:54 <pikhq> <irpbot> Job #1 requested by pikhq.
21:54:56 <pikhq> !accept 1
21:55:05 <pikhq> [:Foo, bar, and baz]
21:55:07 <ehird`> irp3 is betar!
21:55:12 <ehird`> therefore #irp3 is betar
21:55:14 <pikhq> <irpbot> Job #1 completed.
21:56:08 <ehird`> ok nobody likes #irp3
21:56:08 <ehird`> :p
21:57:22 <ehird`> :(
22:27:22 -!- cherez has quit ("Leaving.").
22:33:15 -!- RedDak has quit (Remote closed the connection).
22:45:59 <GregorR> "IRP" is a trademark of GregorCorp and cannot be used without explicit written permission from Gregor Richards.
22:47:03 <xor> GregorCorp is fraudulent
22:47:08 <ehird`> i agree
22:47:25 <ehird`> get your butt into #irp3 and realise how awesome it is :p
22:48:13 <ehird`> it is a-maze-ing
22:49:26 <ehird`> :D
22:50:00 <xor> lynx is a-maze-ing
22:50:47 -!- ehird` has changed nick to ifte.
22:52:27 -!- ihope has changed nick to and.
22:52:43 -!- and has changed nick to ihope.
22:53:24 <ifte> oi! someone go into #irp3 and run an expression
22:53:30 <ifte> you can do ifte: x y z (if x then y else z)
22:53:37 -!- ihope has changed nick to and.
22:53:40 <ifte> and and: x y (if x then y else x)
22:53:48 <and> "and and"?
22:53:53 <and> Oh.
22:53:53 <ifte> and "and: x y"
22:53:55 <ifte> you can enclose sub-expressions with ()
22:53:58 * and nods
22:54:07 <ifte> so ifte: (and: true false) true false is "if true and false then true else false"
22:54:12 <ifte> now! get in there and run something!
22:55:14 -!- and has changed nick to ihope.
22:55:50 <ifte> ... nobody wants to run an expression :(
22:55:52 <xor> () aren't needed
22:56:06 <ifte> xor: to avoid !count they are
22:56:10 <xor> that's the whole point of prefix/postfix
22:56:14 <ifte> ummm ok
22:56:16 <ifte> parse this:
22:56:21 <ifte> ifte: and: true false true false
22:56:25 <ifte> you do not know how many arguments and takes
22:56:37 <xor> you have to
22:56:39 <ihope> : is essentially an infix operator there.
22:56:44 <ifte> exactly, xor
22:56:45 <xor> variadic functions aren't allowed...
22:56:50 <ifte> err, its not just variadic
22:56:53 <ihope> You could use a prefix operator instead.
22:56:56 <ifte> its that there could be 100 operators you don't know about
22:57:06 <ifte> and you cna't remember all their arity
22:57:07 <ifte> so, ()s.
22:57:15 <xor> ...which is why it's a problem to have every operator do its own parser
22:57:28 <ifte> oh, we'll just become a big centralized human called Man
22:57:32 <ifte> and we'll ask Universal AC all our questions
22:57:36 <ifte> simple, really
22:57:44 <pikhq> proc ifte {expr-arg} {set val [uplevel 1 expr $expr-arg];uplevel 1 expr {($val == 0 && $val == 1) ? 1 : 0}}
22:57:47 <ihope> Hmm...
22:58:10 <ihope> Actually, lemme sort of do something else for a while.
22:59:44 -!- oklofok has joined.
23:07:03 <ihope> It's an oklokok!
23:07:56 <ifte> hmm
23:08:18 <ifte> oklofok: oklofok: oklopol: oklopol: oklopol: oklopol: oklopol: oklopol: oklofok: oklofok: oklopol
23:08:20 <ifte> it's binary!
23:09:51 <oklofok> what am i now?
23:09:56 <oklofok> oklofok, okay.
23:11:04 <oklofok> ihope said something about koks and i got all confused
23:11:26 <ihope> fok + pol = kok.
23:11:54 <oklofok> i did catch myself a girl during the excursion by the way, thanks for asking
23:12:06 <oklofok> not that anyone remembers what i said 2 weeks ago :P
23:12:21 <oerjan> 1542
23:12:43 <oklofok> 1542? isn't that the amount of different languages spoken in india?
23:12:49 <oklofok> oh, that's 1652
23:13:22 <oerjan> no it's the binary number ifte mentioned above
23:13:43 <ifte> how do you know what was 0 and what was 1
23:13:55 <oklofok> you can assume it starts with 1...
23:14:01 <oerjan> exactly
23:14:10 <oklofok> unless it has 1 digits
23:14:49 <oklofok> i don't know what happens if there's a negative number of digits
23:14:56 <ifte> it implodes
23:15:04 <oklofok> ah, obviously
23:15:15 <oerjan> antibinary
23:15:24 <ifte> how do you know i wasn't using two's complement, also? :)
23:15:45 <oerjan> with 11 bits?
23:16:07 <oklofok> 2's complement doesn't work with any number of digits?
23:16:48 <oerjan> well, in theory.
23:16:58 <oklofok> ah
23:17:00 <oerjan> i was more worried about getting endianness wrong
23:17:00 <oklofok> i see what you mean
23:17:48 <oklofok> yeah, you can prolly assume it's little endian and not two's complement if it doesn't have 2^n digits
23:17:50 <oklofok> err
23:18:05 <oklofok> i actually just guessed little endian, i never remember which is which
23:18:20 <oerjan> me neither
23:18:28 <oklofok> i have some issues with which-is-which-ishness
23:18:37 <oklofok> as i hear most people do
23:18:53 <oerjan> i vaguely recall that the end in case is actually the beginning
23:19:24 <oerjan> i.e. little endian means the _first_ bit is the least significant
23:19:52 <oklofok> yeah, so you just think little+end / big+end, and then reverse the meaning
23:20:22 <ifte> little endian = ends with the littlest bit
23:20:26 <oklofok> oh
23:20:29 <ifte> big endian = ends with the biggest bit
23:20:33 <ifte> thus "endian"
23:20:42 <ifte> it's a big ended number, so it's a big endian number
23:21:23 <oerjan> yep, but with the end being the first end
23:21:36 <ifte> oh, wait i'm wrong
23:21:42 <ifte> "Again, big-endian does not mean "ending big", but "big end first"." -- wp
23:22:06 <oerjan> yeah, it is exactly the opposite of what intuition tells me, which is how i managed to remember it :D
23:23:21 <oklofok> then i'd've remembered correct
23:30:08 -!- ifte has quit (Read error: 104 (Connection reset by peer)).
23:33:18 <oklofok> by the way, i checked the number of languages spoken in india, and oerjan was wrong
23:33:38 <oklofok> i was right
23:33:44 <oklofok> i'm on a roll here
23:33:52 <GregorR> Every language that has ever been spoken is spoken in India.
23:34:38 <oerjan> how can i be wrong when i haven't commented on the issue? :/
23:34:43 <oklofok> well, be that as it may, i have proof it's exactly 1652
23:35:06 <oklofok> oerjan: rarely can you take anything i say seriously.
23:36:15 <oklofok> especially at this time of night, when i'm keeping silent channels alive.
23:36:23 <oerjan> even more rarely than what i say? that's horrible
23:37:21 <oerjan> incidentally that Wikipedia article mentioned Sanskrit as a rare little-endian language
23:39:19 -!- sebbu has quit ("Leaving").
23:39:35 <oerjan> that 1652 is probably decreasing as we speak.
23:40:07 <oklofok> "kuinka montaa erikielt puhutaan Intiassa ::= 1 652 eri kielt"
23:40:28 <oklofok> i guess you should've played #tietovisa for some 100 hours to get my joke.
23:40:47 -!- jix__ has quit ("CommandQ").
23:41:00 <oklofok> (finnish for "how many differentlanguages are spoken in india ::= 1 652 different languages")
23:41:55 <oklofok> anyway, wikipedia says > 1500 languages
23:42:23 <oklofok> i guess 1652 is over 1500, thoughj
23:42:25 <oklofok> *jhjhtg
23:47:32 <oerjan> what does tietovisa mean?
23:47:50 <oklofok> "visa" is kinda like "trivia"
23:48:01 <oklofok> "tieto" is "knowledge"
23:48:25 <oklofok> just "visa" would work fine for the channel name
23:51:00 <oklofok> also, turns out also #visa is a finnish trivia channel
23:51:04 <oklofok> (qnet)
23:51:19 <oklofok> *is also a
←2007-08-04 2007-08-05 2007-08-06→ ↑2007 ↑all