00:00:02 -!- ENKI-][ has joined.
00:00:25 <oerjan> however the "best bakes" page has three different listings, the last of which are definitely biased to things that are rated only a few times
00:01:32 <oerjan> the listings and voting system are essential to actually getting the better comics known
00:09:24 <oerjan> oh and also i don't think the Comic Irregulars have revealed anything much about how the mezzacotta comic generation works (barring that i haven't read the forum for today yet)
00:10:30 <oerjan> but there is plenty of speculation on the forum of course
00:16:28 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | sometimes i spend hours hunting some bug.
00:16:44 <oerjan> optbot is a hard working programmer
00:16:45 <optbot> oerjan: I consider it modern art.
00:18:14 -!- puzzlet_ has joined.
00:18:58 -!- puzzlet has quit (Remote closed the connection).
00:21:09 -!- moozilla has joined.
00:26:34 <ehird> http://www.mezzacotta.net/archive.php?date=1995-08-22 <- this actually happened when i was born
00:26:42 <ehird> i am the one on the right
00:31:22 * oerjan was confused there for a moment
00:31:36 <oerjan> i see your intelligence showed early
00:32:29 <ehird> I love my facial expression in the last comic
00:33:29 <oklocod> http://www.mezzacotta.net/archive.php?date=1748-04-04 <<< oh my god, i wonder how long it's been since i laughed this hard at a comic
00:33:54 <ehird> oerjan: http://www.mezzacotta.net/archive.php?year=2001&epoch=ad&month=09&day=11 The sorrows of 9/11 apparently break the algorithm.
00:34:26 <oerjan> i've checked 9/11 before i don't recall it being anything special
00:34:38 <oklocod> did you find it funny, or am i just too tired?
00:34:48 <ehird> oerjan: "This randomly generated comic is just not working today."
00:34:49 <ehird> oklocod: hilarious
00:34:52 <ehird> the facial expressions XD
00:34:52 <oklocod> the expression in the last square was simply priceless
00:35:07 <oklocod> cuz he looks so damn calm first
00:35:11 <oerjan> oklocod: i voted it 100% before you pasted it
00:35:29 <oklocod> okay i have high hopes for #2 too
00:35:47 <oklocod> well i didn't laugh, that just *made sense*.
00:36:19 <oklocod> these things are usually only funny in that they make very little sense
00:36:31 <ehird> http://www.mezzacotta.net/archive.php?date=-5679392-11-17
00:37:00 <oerjan> i think that is from the hall of fame
00:37:47 <oerjan> but i think quality has gone up, obviously because of more people visiting
00:38:04 <oerjan> (on the selected comics)
00:38:21 <ehird> the comics don't change
00:38:32 <oerjan> no but the selections do
00:38:47 <oerjan> the more people visit, the more comics will compete for the selected listings
00:39:10 <oerjan> and so they should become better
00:39:54 <oerjan> i don't think the underlying comics algorithms have changed either
00:40:20 <oerjan> but it's hard to be sure when they are not revealing anything
00:40:25 <ehird> oklocod: http://www.mezzacotta.net/archive.php?date=612-07-27
00:40:33 <oklocod> http://www.mezzacotta.net/archive.php?date=1729-04-03 xD
00:40:43 <ehird> http://www.mezzacotta.net/archive.php?date=612-07-27
00:41:06 <ehird> http://www.mezzacotta.net/archive.php?date=341-07-05
00:41:25 <ehird> http://www.mezzacotta.net/archive.php?date=1995-06-02
00:42:06 <ehird> http://www.mezzacotta.net/archive.php?date=1920-08-20
00:43:10 * oerjan guesses ehird has checked out the forum discussion on silent panels
00:43:22 <ehird> just finished reading the big thread
00:43:28 <oklocod> http://www.mezzacotta.net/archive.php?date=-1286-10-03 <<< xD
00:43:37 <oklocod> okay i laugh at this more than xkcd.
00:44:04 <oklocod> well i guess it's partly because i'm tired and i opened a laugh-gate or something
00:44:48 <oklocod> but really i think it's the faces
00:44:57 <ehird> but yes this is better than xkcd
00:45:48 <oerjan> warning: don't include your vote in pasted links
00:46:34 <ehird> oerjan: they should really use POST.
00:46:51 <oerjan> ehird: there are some bugs like that
00:47:52 <oerjan> the TODO list is growing quickly, and DMM says it won't be handled speedily
00:48:10 <ehird> http://www.mezzacotta.net/singles/collapse_a_wavefunction.php
00:48:20 <oklocod> why is the unbaked cake eaten, and not the fully baked one?
00:48:46 <oerjan> oklocod: hm a conundrum
00:50:51 <oklocod> http://www.mezzacotta.net/archive.php?date=1457-07-02 :D
00:51:12 <ehird> http://www.mezzacotta.net/archive.php?date=1995-08-22
00:51:17 <ehird> the lines are different
00:52:09 <ehird> http://www.mezzacotta.net/archive.php?year=9999999&epoch=bc&month=01&day=01
00:53:29 <ehird> oklocod: my birthday comic, just a bit
00:53:32 <ehird> http://www.mezzacotta.net/archive.php?date=-600000-11-05
00:54:28 <oerjan> well it's cached after it has been visited once, maybe that changes something
00:54:42 <ehird> cause i refreshed after that
00:54:54 <oerjan> or maybe they *shiver* changed the algorithm
00:55:00 <ehird> oerjan: probably :(
00:55:56 <ehird> http://www.mezzacotta.net/archive.php?date=-1875-01-18
00:58:15 <oklocod> http://www.mezzacotta.net/archive.php?date=-2137-11-25 <<< okay the cycle time isn't *that* long
00:58:33 <oklocod> well you prolly know what i mean.
01:00:25 <ehird> http://www.mezzacotta.net/archive.php?year=1992&epoch=ad&month=03&day=18
01:00:56 <oerjan> you mean some of the lines repeat frequently
01:01:11 <ehird> ais523: http://www.mezzacotta.net/archive.php?date=1005-03-16
01:01:14 <oklocod> oerjan: something like that.
01:01:46 <ehird> http://www.mezzacotta.net/archive_new.php?date=888-01-06
01:01:53 <oerjan> there are obviously several corpuses being used
01:02:02 <oerjan> some may be bigger than others
01:02:27 <oklocod> what i don't get is all thet parentheses
01:02:34 <oklocod> why don't they remove those
01:02:34 <oerjan> CHESS IS TRADITIONAL is from the IWC comic
01:03:09 <ehird> http://www.mezzacotta.net/archive.php?date=642-04-16
01:03:37 <oerjan> it was said by one of the Deaths when someone, i think Kyros got a chance to challenge him
01:03:43 <oklocod> hmm, that may be a lie, because i'm not sure i knew whether webcomic is written as one word...
01:03:55 <oklocod> guess i'm just surprised i didn't get it.
01:04:02 <ehird> http://syndicated.livejournal.com/mezzahalloffame/ an almost-daily webcomic version lynched from the halls of fame
01:04:52 <mbishop> that comic is as weird as jerkcity
01:05:28 <ehird> mbishop: i think that says more about jerkcity than mezzacotta
01:17:11 <lament> lguhlughulghulgulgulgulgulghuglghuh
01:19:10 <oerjan> HELP, LAMENT IS DROWNING
01:32:17 -!- Asztal has joined.
01:33:55 <mbishop> a fan in the window and a chair that's not really a chair?
01:33:56 <lament> oh, if he has a vagina then it's all good
01:34:19 <psygnisfive> oh, its not his place. its his friends living room.
01:34:22 <oerjan> if he were rich he'd have fixed that vagina, obviously.
01:34:35 <psygnisfive> lament: so you dont care if its a guy, so long as theres a vagins?
01:35:11 <oerjan> lament is _so_ object-oriented
01:35:28 <lament> this is referential transparency we're talking about
01:35:44 <psygnisfive> interesting in the same way that my homosexuality is interesting even tho i'd fuck a guy who has a vagina.
01:36:22 <lament> i'm guessing he's into women though?
01:36:41 <lament> woman turned homosexual?
01:37:13 <psygnisfive> i know a transguy (that is, female-to-male)
01:37:38 <lament> or otherwise not a lifestyle choice
01:38:13 * oerjan recalls from somewhere about a guy who changed into a lesbian harley biker, or something like that
01:38:38 <oerjan> (probably biked before)
01:39:52 <mbishop> well they could be in the mind
01:39:56 <mbishop> but that's probably illegal
01:39:57 <oerjan> and all three are separable
01:40:21 <psygnisfive> but brain != mind. atleast not technically
01:40:28 <oerjan> psygnisfive: well wait until we develop suitable psi powers
01:40:40 <psygnisfive> but it still wouldnt put a cock inside a mind
01:41:08 <oerjan> impregnation through telekinesis
01:41:10 <psygnisfive> mind is not a physical thing, its an organizational structure and pattern
01:41:50 <oerjan> well that's one theory. science doesn't really know.
01:43:00 <psygnisfive> tht ultimately amount to "i cant imagine it, so its impossible!"
01:43:35 <oerjan> now really. the fact that _he_ doesn't know in no way implies that science does
01:44:05 <oerjan> since science is still incapable of physically reading thoughts directly from brain structure, it doesn't.
01:44:47 <psygnisfive> i dont think theres anything TO know, to be honest
01:45:14 <Dewi> oerjan: except that it can...
01:45:27 <Dewi> oerjan: not to any great degree of precision, but lots of thought can be accessed
01:45:44 <psygnisfive> so what the nature of brain/mind is doesnt matter THAT much since we know it must be necessarily turing-equivalent
01:46:02 <Dewi> oerjan: oh sorry, from *structure*
01:46:12 <psygnisfive> and the functionalist model looks very similar to the brains behavior, to me
01:46:13 <Dewi> oerjan: you're right
01:46:28 <Dewi> oerjan: but I'd argue that that's just a limit of our modelling power
01:46:57 <Dewi> oerjan: our linear computers take a long time to simulate that many neurons, and real brains take years-to-decades to develop
01:47:14 <psygnisfive> so ofcourse we cant read thoughts by looking at structure
01:47:44 <Dewi> psygnisfive: they come from structure plus stimulus though
01:47:54 <Dewi> psygnisfive: and granted, science isn't able to really model that
01:48:02 <psygnisfive> yes but the structure is mechanism + stored data
01:48:11 <Dewi> (because it's too much for any digital computer!)
01:49:05 <psygnisfive> a clear example of why structure is not thought: brain dead people have no structural differences, but they lack the information flow
01:49:30 <oerjan> structure was the wrong word
01:50:57 <psygnisfive> modern science couldn't do similar from looking at a microchip. atleast, not in a reasonable amount of time.
01:51:12 <psygnisfive> and the brain is much larger than a microchip and doesn't operate on the same principles of computation
01:54:52 <psygnisfive> i find it easier to imagine a brain-as-TM-simulator
01:55:43 -!- puzzlet has joined.
02:08:43 -!- puzzlet_ has quit (Read error: 110 (Connection timed out)).
02:26:33 <GregorR> I've been thwarted by shoes again >_<
02:28:03 <oerjan> they're probably jealous of the hats
02:29:20 <oerjan> GregorR: the only solution is that you register choosemyshoes.com
02:30:13 <GregorR> I spend four hours today looking for shoes I can wear.
02:30:20 <GregorR> As it turns out, they don't sell leather-free shoes in Indiana.
02:30:44 <ihope> I suddenly have a feeling GregorR is vegan.
02:34:33 <GregorR> (More accurately, I'm allergic to chromium, which is used to tan most leather, and also process synthetic leather so I can't wear that either yee haw)
02:39:22 <GregorR> But then, I don't think I'd trade my effing weird allergy for the normal array of annoying allergies.
02:39:52 <GregorR> Having to pay $150 for shoes isn't as bad as sneezing a billion times a day for six months per year :P
02:41:56 <ihope> I wouldn't like to sneeze at 11.6 kHz either.
02:49:35 <ihope> Hmm, what would that sound like...
02:50:16 <ihope> Probably just an 11.6 kHz sine, since it would be periodic at 11.6 kHz and you wouldn't really be able to hear the upper harmonics.
03:20:10 <comex> so how exactly is the dialgoue generated
03:20:45 <oerjan> they have not revealed it
03:21:26 <oerjan> what you can say is that it is based on several sources, such as an eliza-like program and their own webcomics
03:22:02 <oerjan> at least some characters depend on what was said previously
03:23:12 <oerjan> see the forum for what has been discussed
03:24:35 <oerjan> s/their own webcomics/their other webcomics/
03:25:27 <oerjan> there are also some public-domain books in there, i think the bible for one
03:34:13 -!- Corun__ has quit ("This computer has gone to sleep").
04:02:04 -!- poiuy_qwert has joined.
04:32:08 -!- rodgort has quit ("Coyote finally caught me").
04:32:21 -!- rodgort has joined.
05:10:35 <Sgeo> xkcd every day this week!
05:12:03 <GregorR> Anybody who considers themselves good at determining whether colors go together, please click wildly at http://home.codu.org/colormatch/
05:13:02 <oerjan> the first two were almost identical. i guess that means they do.
05:13:08 -!- poiuy_qwert has quit (Read error: 110 (Connection timed out)).
05:13:35 <oerjan> black goes with anything, doesn't it >:D
05:13:54 <GregorR> I'm generating some input data to attempt to make a neural net with.
05:14:01 <GregorR> I doubt it'll work well, because I think it's very subjective.
05:14:07 <GregorR> But it's possible that there are some humanish themes.
05:14:08 <oerjan> although to be honest i didn't like the other (purplish) color much
05:23:44 <oerjan> did you just switch the layout or was it part of the original process?
05:24:02 <oerjan> (in any case it seemed to be easier with plaids)
05:25:07 <GregorR> I just switched it on somebody else's recommendation.
05:26:49 * oerjan thinks his eyes are starting to have some illusion effect, so he'll stop
05:35:44 <mbishop> http://www.heyokay.com/wp-content/images/computer programming.jpg
05:36:42 <lament> i want a computer like that
06:14:10 -!- metazilla has joined.
06:14:12 -!- Sgeo has quit (Remote closed the connection).
06:14:18 -!- moozilla has quit (Nick collision from services.).
06:16:28 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | and constants = first letter uppercase (because FOO, etc).
07:13:12 -!- Judofyr has joined.
07:21:46 -!- Judofyr has quit.
07:23:43 -!- Judofyr has joined.
07:38:27 -!- Judofyr has quit.
07:40:39 <GregorR> There's no way that color preference is deterministic.
07:40:52 <GregorR> Clearly I have too little data.
07:43:01 <Asztal> how many votes did you get?
07:44:30 <GregorR> And cancel the last three lines, my previous result seems to have been a bug (although I'm not sure what bug >_> )
07:44:41 <oerjan> heck i'm not even sure _my_ preferences were deterministic
07:45:13 <GregorR> Especially when your name is Gregor.
07:45:18 <GregorR> (Hyper-obscure reference++)
07:47:20 * oerjan squashes a dung beetle
07:48:40 * Jiminy_Cricket hopes oerjan doesn't construe crickets to be dung beetles.
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:14:26 <GregorR> Just hit 70% legitimately 8-D
08:14:37 <GregorR> I doubt you can do much better than 70%.
08:32:48 -!- oerjan has quit ("leaving").
09:20:15 -!- olsner has quit ("Leaving").
10:27:21 -!- M0ny has joined.
10:31:01 <fizzie> 11:32:51 <fizzie> fungot: Plop, goes the BONUS BALL.
10:31:02 <fizzie> 11:32:51 <fungot> fizzie: whoooooooa!!
10:39:20 <ais523> fizzie: what database is fungot using at the moment?
10:39:21 <fungot> ais523: official records state sephiroth is traveling the world like everyone's been saying'...
10:39:45 <fizzie> The script for the Playstation RPG Final Fantasy VII.
10:41:35 <fizzie> Were you one of those logreading people? Tried out 1200 transcribed Penny Arcade comics, and Agora rules, yesterday. Both of those weren't too bad.
10:47:23 <ais523> I don't exactly logread
10:47:32 <ais523> but my bouncer records all the messages while I'm not here
10:47:35 <ais523> then replays them when I join
10:47:44 <fizzie> Oh, right, there was that trick.
10:47:45 <ais523> so it's like I've been here all the time, and I read scrollback rather than logreading
10:48:11 <fizzie> Well, logreading, scrollback-reading; same thing.
10:48:52 <fizzie> fungot: Quote an Agora rule, please?
10:48:52 <fungot> fizzie: any player can flip eir posture to any non-standing value by announcement. if e disqualifies the judge, to all players
10:49:07 <ais523> the first sentence is verbatim from the Agoran rules
10:49:10 <ais523> the second clearly isn't
10:49:53 <fizzie> Shortening the context gives more... "inspired" output. That FF7 thing had one word less context than this.
10:50:13 <fizzie> fungot: Why you have to always quote stuff verbatim? That's not fun at all!
10:50:13 <fungot> fizzie: described elsewhere, and provides a tally of the voters' valid ballots on the various options.
10:50:27 <fizzie> Yes, it sounded suspiciously sensible.
10:50:30 <ais523> oh dear, I seem to have memorised far too much of the Agoran ruleset...
10:50:34 <fungot> ais523: " impossible" is defined here to mean overwhelmingly psychotic behavior, a complete lack of interest in the game, each player's score is set to eir bvlop.
10:50:41 <ais523> ok, that's more interesting
10:51:03 <ais523> btw, setting score to bvlop is not a sensible operation at all, and wouldn't have been even when bvlop was defined
10:51:57 <fizzie> fungot: Can you generate more original stuff if you have a shorter context, like now?
10:51:57 <fungot> fizzie: the stuck switches on an immune player which made the announcement that a judge
10:52:25 <fizzie> It's just that the coherence level decreases rapidly if I shorten the context.
10:52:41 <fizzie> I don't know what a bvlop is.
11:02:41 -!- Corun__ has joined.
11:09:03 -!- moozilla has joined.
11:20:02 -!- moozilla has quit (Nick collision from services.).
11:20:04 -!- megatron has joined.
11:20:25 -!- moozilla has joined.
11:40:10 -!- metazilla has quit (Connection timed out).
11:41:09 -!- Corun__ has quit ("This computer has gone to sleep").
12:16:28 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | Jeeze. . . If you're going tu use C syntax, at least be willing to use the power of pointers..
12:31:34 <ais523> fizzie: it isn't any more, but it used to be base voting limit on ordinary proposals
12:32:02 <ais523> which was 4 if you were a real person, and 0 otherwise, say if you were a computer program or consortium of people
12:32:17 <ais523> fungot: more Agoran inspiration? You should meet Bayes.
12:32:17 <fungot> ais523: what's happening? i think i... what!
12:32:35 <ais523> fungot: that sounded pretty unusal for you...
12:32:36 <fungot> ais523: oh, and once more to turn over a stone or two to see it? will you get involved...' blah, blah...
12:32:54 <ais523> fungot: has your data source changed again?
12:33:20 <fungot> ais523: like you wouldn't be standing here now.
12:33:20 <ais523> anyway, time to go for a while, I'll be back soon, fungot
12:33:20 <fungot> ais523: if i did go to that town across the river...... that's so rude...
12:35:09 <fizzie> Yes, that was fungot impersonating that game again. Maybe I should make it randomly pick from the available sources.
12:35:09 <fungot> fizzie: then it really you, cloud?
12:35:23 <fizzie> fungot: No, I'm not Cloud. You must be confused.
12:35:23 <fungot> fizzie: remember it! why now... no. i don't know what i'm made of! ......urk!
12:35:51 <fizzie> fungot: You're made of some Funge-98 code. Hope this helps.
12:35:51 <fungot> fizzie: nothing. and, where is everybody? if you want! who do they mean.... don't leave me. something's wrong.
12:36:09 <fizzie> fungot: Calm down, everything's all right.
12:36:09 <fungot> fizzie: i thought i would kick their butts.
13:15:26 <fungot> oklocod: wait a minute, cloud! what can he do for ya......... i know this's kinda outta the blue, but...... that name?
13:15:51 <oklocod> also i love this current one
13:16:23 <fizzie> It might be partially because there's also the shorter context thing, the game script was so short.
13:16:29 <fizzie> fungot: Have you been inhaling things again?
13:16:30 <fungot> fizzie: now, i can read sheet music for that! but if you're on a big splash. hold onto your drawers and don't get so caught up in these reports.
13:16:52 <fizzie> I didn't know bots can get high by reading sheet music.
13:16:57 <ais523> random pick from sources would be interesting
13:17:10 <ais523> hmm... what would happen if you supplied fungot with fungot data
13:17:10 <fungot> ais523: did you... jealous? hmm? you... did you say something, and mr. president... i've never seen your faces ' round here before. goin' after me?
13:17:13 <ais523> and used a really short chain?
13:17:20 <ais523> probably it hasn't generated enough to be worth chaining from
13:18:15 -!- oklocod has changed nick to oklopol.
13:18:36 <ais523> hmm... I saw your oko Underload program in scrollback
13:18:43 <ais523> that inspires me to try it for myself, without looking at yours
13:19:09 <ais523> +ul (o)(~:S(ko)*( )S~:^):^
13:19:10 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
13:19:18 <ais523> wow, right first try, not bad
13:19:32 <fizzie> I need to add that output limitation thing to fungot.
13:19:32 <fungot> fizzie: a lot easier. and believe in cloud...... it is not necessary to use that sailor suit. he is!
13:19:44 <ais523> fizzie: don't you have the ...
13:19:44 <fizzie> ^ul (o)(~:S(ko)*( )S~:^):^
13:19:46 -!- fungot has quit (Remote closed the connection).
13:19:53 <ais523> ah, not in the Underload program
13:20:14 <ais523> thutubot never breaks output in the middle of an S command
13:20:20 <fizzie> Hey, that was pretty curious.
13:20:23 <fizzie> RAW >>> :fizzie!i=fis@iris.zem.fi PRIVMSG #esoteric :^ul (o)(~:S(ko)*( )S~:^):^ <<<
13:20:27 <fizzie> *** glibc detected *** ./cfunge: double free or corruption (!prev): 0x080dbee0 ***
13:20:44 <ais523> so in other words, my Underload oko program made fungot crash cfunge?
13:20:49 <ais523> AnMaster needs to know about this
13:20:53 <oklopol> +ul (o)(~:S(ko)*( )S~:^):^
13:20:54 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
13:20:55 <ais523> I wonder how easy it is to reproduce?
13:21:13 <oklopol> okay that's exactly what you had
13:21:21 <ais523> ah, did you come up with it independently?
13:21:24 <oklopol> i was kinda hoping it would have at least some difference.
13:21:30 <ais523> great Underload programmers think alike, obviously
13:21:44 <oklopol> there are many steps you could do in different order
13:21:57 <fizzie> ais523: It seems to be very easy to reproduce, as it crashes whenever I input that program.
13:22:00 <ais523> the ( )S could go anywhere after the first S and before the first ^
13:22:20 -!- fungot has joined.
13:22:29 <fizzie> (It's not fixed, so don't bother testing.)
13:22:32 <ais523> hmm... I suppose it would be neater to do it like this:
13:22:49 <ais523> +ul (o )(~:S(ok)~*~:^):^
13:22:49 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
13:22:55 <ais523> is that shorter, I wonder?
13:23:14 <oklopol> lol, i'm probably making the exact same change as you
13:23:20 <ais523> you have a stray S in your program
13:23:22 <oklopol> keeping (o ) on stack and adding ok
13:23:23 <ais523> but otherwise it's the same
13:23:53 <oklopol> oh, well yeah i just copypasted and changed the beginning
13:24:01 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
13:24:11 <oklopol> that's probably as short as it gets
13:24:42 <ais523> a finite oko would be even shorter, I think
13:25:00 <ais523> +ul (o )(:S(ok)~*):*:*:*^
13:25:00 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko
13:25:03 <ais523> +ul (o )(:S(ok)~*):*:*:*;*^
13:25:07 <ais523> +ul (o )(:S(ok)~*):*:*:*:*^
13:25:07 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
13:25:08 <oklopol> +ul (o)S( o)(~(ko)*:S~:^):^
13:25:09 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
13:25:16 <oklopol> +ul (o)S( o)(~(ko)*:S~:^):^
13:25:17 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
13:25:27 <oklopol> heh, i'm getting pretty fluent at this too
13:25:30 <ais523> +ul (o )(:S(ok)~*):*::**:*^
13:25:30 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko
13:25:46 <ais523> Challenge: given a positive integer, find the shortest way to write it in Underload
13:27:10 <oklopol> and just the number of x's to write, or in base something?
13:27:20 <ais523> Underload constants are probably just as tricky, or more so, to work out than Brainfuck constants
13:27:24 <ais523> :* ::** :*:* ::*:** are the first 5, I suspect
13:27:30 <ais523> hmm... why didn't that work?
13:27:38 <ais523> well, not so much of an individual program
13:27:40 <ais523> but a general task, like [[e:Brainfuck constants]]
13:27:42 <ais523> but a general task, like http://esolangs.org/wiki/Brainfuck_constants
13:27:58 <fizzie> ^bf +++++++++++[>++++++++++>+++<<-]>+.>-.[[<]>[-<+<+<+>>>]<----<<[->>>+<<<]>[.>]<]
13:27:58 <fungot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokoko ...
13:28:01 <ais523> :* ::** :::*** ::::**** works for all integers
13:28:06 <ais523> but it could be shorter for most of them
13:28:16 <ais523> I'm wondering what the general rule is to find the shortest program to produce a given integer
13:28:25 <ais523> probably it's NP-hard or something, tbh
13:28:44 <ais523> well, 4 is :::***, but it's also :*:* which is shorter
13:29:23 <oklopol> is this let-your-nephew-irc-with-your-nick week, first oerjan then you, it's impossible to solve that :D
13:29:28 <ais523> likewise the shortest way to write 5 is ::*:**, and 6 is probably :*::**
13:29:34 <ais523> oklopol: no, it can clearly be brute-forced
13:29:50 <oklopol> when ais comes back, ask him about semidecidability
13:29:55 <ais523> because some of the programs you try might not terminate
13:30:00 <ais523> sorry, not thinking straight there
13:30:14 <ais523> probably there's general rule for programs made of : and *
13:30:23 <ais523> *there's a general rule
13:30:40 <oklopol> probably for any non tc subset
13:30:50 <ais523> hmm... I wonder what the first integer for which the shortest constant is mathematically undecidable is, probably it's pretty high
13:31:13 <oklopol> that can't be solved either, you can only solve a lower bound
13:31:16 <ais523> and any non-TC subset whose halting problem is solvable, clearly it could be brute-forced
13:31:25 <ais523> ah yes, undecidable undecidability
13:31:28 <ais523> I can still wonder, though
13:32:06 <oklopol> +ul (xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx)S
13:32:07 <thutubot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
13:32:24 <ais523> well, clearly that's beatable
13:32:33 <AnMaster> <ais523> AnMaster needs to know about this <-- ?
13:32:35 <oklopol> but that wasn't even the number
13:32:46 <oklopol> i mainly wanted to see how long 79 is :P
13:33:00 <ais523> +ul ::::::*:**::*:**:******(x)~^S
13:33:10 <ais523> +ul (::::::*:**::*:**:******)(x)~^S
13:33:22 <ais523> ugh, must be a typo there somewhere...
13:33:26 <ais523> AnMaster: its' on fizzie's computer, not mine
13:33:49 <ais523> +ul (::::::*:**::*:**::******)(x)~^S
13:33:49 <thutubot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
13:33:55 <fizzie> I can participate in going after it later. But I guess I could take a quick look under gdb right now.
13:34:15 <AnMaster> well I'm leaving for the rest of the day within maybe 10 minutes
13:34:27 <ais523> ah, slightly early bye then
13:34:57 <AnMaster> so 1) how to reproduce 2) any backtrace (with -g or -ggdb3 and hopefully -O0)
13:35:10 <ais523> +ul ((:*):*:*:*:*)(x)~^S
13:35:10 <thutubot> :*:*:*:*:*:*:*:*:*:*:*:*:*:*:*:*
13:35:12 <AnMaster> fizzie, need to go out and buy new clothes and such
13:35:16 <ais523> ugh, hit typo by mistake
13:35:21 <ais523> oklopol: was meant to be, I haven't counted though
13:36:00 <ais523> I think it's probably possible to get it shorter, though
13:36:19 <ais523> +ul (:*:*:*:*:*:*)(x)~^S
13:36:19 <thutubot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
13:36:26 <ais523> is it written as 2*2*2*2*2*2
13:36:32 <ais523> I'm wondering if 2^6 would be shorter
13:36:44 <ais523> +ul (:*)(:*::**)^(x)~^S
13:36:45 <thutubot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
13:37:12 <ais523> one char shorter, I think
13:37:30 <ais523> and obviously the savings go up as you go to larger powers of 2
13:37:38 <ais523> ah, and two chars shorter still:
13:37:44 <thutubot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
13:38:18 <ais523> I wonder if 79 can be written shorter using that sort of trick?
13:39:31 <oklopol> +ul (::*::**:::***~::::::*******)(x)~^S [2, 3, 4]+7
13:39:31 <thutubot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
13:39:34 <oklopol> +ul (::*::**:::***~::::::*******)(x)~^S
13:39:35 <thutubot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
13:40:08 <oklopol> this is what you get for letting the computer do the thinking
13:40:13 <ais523> I think thutubot just stops if it hits an unrecognised character, that makes sense given the way it's programmed
13:40:17 <oklopol> and not really thinking when asking it
13:40:52 <ais523> nothing wrong with that as long as you don't trust the answer to have answered the question you were trying to ask
13:40:59 <oklopol> +ul (::*(::**)(:*)^:*:*~::::::*******)(x)~^S
13:40:59 <thutubot> ::**::**::**::**::**::**::**::**xxxxxxxxxxxxxx
13:41:17 <oklopol> +ul (::*(::**)(:*)^^:*:*~::::::*******)(x)~^S
13:41:18 <thutubot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
13:41:19 <fizzie> Well, it's the free_nogc() in STRN.c:122 that's hitting the glibc double-free thing. Will look at the details later.
13:41:36 <oklopol> mainly because i have to go ->
13:42:00 -!- fungot has quit (Remote closed the connection).
13:42:16 <fizzie> I think I said so, too.
13:42:17 <AnMaster> fizzie, well what is the link the the same version of the source
13:42:21 <AnMaster> so I can test it locally later
13:42:25 <ais523> hmm... where are the links for STRN?
13:42:32 <ais523> I mean, the definition
13:42:49 <fizzie> I'm currently testing whether it works on the stand-alone Underload interp, because that's a lot smaller piece of code.
13:42:53 <ais523> where's the link to those fingerprints?
13:43:04 <AnMaster> fizzie, still I don't think cfunge should crash
13:43:14 <AnMaster> so even if it is a bug in your program I want to debug this
13:43:34 <fizzie> fis@eris:~/src/bef$ echo '(o)(~:S(ko)*( )S~:^):^' | ~/inst/cfunge/cfunge/build/cfunge underload.b98 > /dev/null
13:43:38 <fizzie> *** glibc detected *** /home/fis/inst/cfunge/cfunge/build/cfunge: double free or corruptio
13:43:50 <AnMaster> fizzie, well got a link to that underload.b98 ?
13:43:58 <fizzie> And underload.b98 is at http://zem.fi/~fis/underload.b98
13:44:00 <AnMaster> ais523, 0x5354524e STRN http://www.rcfunge98.com/rcsfingers.html
13:44:13 <fizzie> Easier to run than fungot, no need to use netcat to pretend to be an IRC server or anything.
13:44:45 <AnMaster> fizzie, doesn't crash here, just runs for ages
13:44:58 <fizzie> It does run for quite a long time before crashing, here.
13:44:58 <AnMaster> how long does it take to crash?
13:46:01 <AnMaster> #5 0x000000000041ab94 in finger_STRN_get (ip=0x15d20c8) at /home/arvid/src/cfunge/trunk/src/fingerprints/STRN/STRN.c:122
13:46:02 <ais523> maybe an infinite loop's filling up memory
13:46:05 <fizzie> It's might also be using quite a lot of memory, since it keeps growing that string.
13:46:06 <ais523> until it overwrites something else?
13:46:18 <AnMaster> I think valgrind is the right tool for this...
13:46:21 <ais523> the program is an infiniloop, after all
13:47:00 <ais523> is valgrind Boehm-GC compatible?
13:47:33 <fizzie> "invalid write of size 1" in stringbuffer_finish (stringbuffer.c:99) (from finger_STRN_get (STRN.c:115)) here, but you can run it on your own of course.
13:47:36 <AnMaster> ais523, no but that is optional
13:49:10 * AnMaster debugs, stringbuffer being external code he didn't write himself (was taken from another gpled project I'm a developer on)
13:49:13 -!- fungot has joined.
13:49:24 <fungot> ais523: it's the shinra for all of a sudden and where is everybody? that wasn't even a tough one. if we solve the puzzles, the fourth floor is this...... ha, ha
13:49:38 <fungot> fizzie: huh? there ain't no difference from before!!
13:49:53 <fizzie> fungot: Yes, you certainly seem to be as strange as you've always been.
13:49:53 <fungot> fizzie: wake up. bye, then that's it!! so let's get it.) is where i was...
13:51:21 <ais523> SMEM: "All commands reflect on error with the error code on the stack:" <--- arrgghh
13:51:40 <ais523> I don't think any other Funge commands push something when they reflect...
13:52:21 <AnMaster> seems I got delayed in leaving for other reasons
13:53:32 <AnMaster> was an off by one error in a call to make sure the buffer was large enough
13:56:37 <fizzie> ^raw QUIT :let's get you a brand-new cfunge to run on
13:56:38 -!- fungot has quit ("let's get you a brand-new cfunge to run on").
13:57:02 * ehird installs rc/funge and grabs the fungot code so he isn't contaminated
13:58:39 <fizzie> What can I say -- it certainly works better than RC/Funge.
14:03:06 <AnMaster> fizzie, didn't it crash on too long string?
14:03:07 <ehird> http://www.mezzacotta.net/archive.php?date=1488-06-05
14:04:12 <AnMaster> fizzie, also the bug would only have been triggered on a string that was exactly a multiple of 256 chars long. So that is why I didn't notice it before
14:04:40 <fizzie> Yes, RC/Funge has fixed 1000-byte arrays for about all STRN operations.
14:05:02 <Deewiant> Yes, RC/Funge is like that. :-P
14:05:27 <AnMaster> right, however this bug could have affected other fingerprints, I think the "read line" one in FILE for example.
14:06:18 <AnMaster> since it was in a generic "build string by appending at the end" "library".
14:06:37 -!- fungot has joined.
14:06:43 <fungot> ais523: like you more than sephiroth's shadow?
14:07:13 <fizzie> The program is still an infinite loop with very very long strings, so it might not work very well.
14:07:14 <ais523> AnMaster: I suspect that you're filling up the memory of fizzie's computer
14:07:23 <fizzie> Let's see what it's doing.
14:07:29 <fungot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko okokokokokokokokokokokokokokokoko okokokokokokokokokokokokokokokokoko okokokokokokokokokokokokokokokokokoko okokokokokokokokokokokokokokokokokokoko okokokokoko
14:07:39 <ais523> fizzie: did you just break it by hand?
14:07:49 <fizzie> ais523: No, it terminated by itself.
14:08:01 <fizzie> ais523: There would be a "... out of time" at the end, but that got cut off.
14:08:13 <ais523> +ul (o)(~:S(ko)*( )S~:^):^
14:08:14 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
14:08:18 <fizzie> It runs something like 10000 Underload instructions before time-outing.
14:08:29 <fizzie> Would be better to have a too-much-output thing too, though.
14:08:32 <ais523> probably that isn't enough for serious programs
14:08:37 <ais523> I wonder what Thutubot's limit is?
14:08:43 <fizzie> Serious programs? There are some?
14:08:43 <ais523> it's non-trivial to work out
14:08:52 <AnMaster> ais523, didn't you code thutubot?
14:08:55 <ais523> but that doesn't mean they couldn't be written in theory
14:09:06 <ais523> AnMaster: yes, but the art of counting in Thutu is mostly based on black magic
14:09:28 <ais523> so it would probably take me an hour or so with a calculator to figure out exactly what its timeout was
14:09:54 <AnMaster> fizzie, where does that underload program store it's stack thing?
14:09:57 <ais523> I have written serious programs in a Thutu wimpmode, but the wimpmode does arithmetic
14:10:38 <ehird> http://www.mezzacotta.net/archive.php?date=1398-07-19
14:11:10 <fizzie> AnMaster: It's there in Funge-space, Y=9, negative X values. (Because STRN has that fixed delta, it's easier to have the stack growing that way.)
14:11:54 <AnMaster> fizzie, right. but why then is the funge stack so large according to valgrind at the end of it
14:12:03 <AnMaster> oh wait it reads it into stack every now and then
14:12:27 <fizzie> It shouldn't, but it might easily have some bugs that cause numbers to creep up in the Funge stack.
14:12:31 <fizzie> ^ul ((foobarbazquux)~:^):^
14:12:44 <fizzie> There's a reasonably small stack limit there, too.
14:13:08 <ais523> +ul ((foobarbazquux)~:^):^
14:13:33 <AnMaster> fizzie, well it is just that I found it pointless to shrink the funge stack in cfunge, it is a struct with a pointer to a malloced/realloced array, a size value and a top-of-actual-stack value
14:13:34 <ais523> ah, that took a while...
14:13:40 <ais523> I set the memory limit high deliberately
14:13:45 <AnMaster> so it doesn't shrink it ever, exceptions: t
14:13:47 <fizzie> AnMaster: Actually with the oko program, yes, it does the string concatenation in the Funge stack.
14:14:17 <AnMaster> t will not copy more than needed of the stack to the new ip
14:14:51 <fizzie> I think the implementation of Underload * reads both strings to Funge stack, then uses STRN A to concatenate them, and writes the result back.
14:15:14 <fizzie> Probably would be more efficient just to copy things around a bit in the Funge-space.
14:15:40 <fizzie> Or maybe not; at least the STRN 'G' pop-string loop is likely to be more efficient than a Funge-coded loop.
14:16:23 <AnMaster> considering how I do A... I calloc a new buffer large enough to hold both strings, then strncat them to that buffer. heh, was quite some time ago I wrote that, could rewrite it in a better way I guess
14:17:05 <AnMaster> (reallocing one string and appending to that would be better I bet
14:17:18 <fizzie> (foo)(bar) is stored in the stack as "bar\0foo\0", so * in any case needs to do quite a lot of copying to arrive at "foobar\0". Still, I guess at least TOYS has some funge-space copy operation.
14:24:25 <fizzie> Yes, * is indeed a simple 91g9G N91g1++9G A 91g1+:91p9P (with some other crud to notice stack underflow in there).
14:31:02 <ais523> fizzie: is that using TOYS?
14:31:12 <ais523> if so, you need the fingerprint-switch code too
14:31:28 <AnMaster> since it kind of makes sense in STRN
14:31:44 * AnMaster has rewritten A, now running fuzz tests on it to check that there are no errors
14:34:26 <fizzie> fungot: Do you PLAY with your TOYS?
14:34:26 <fungot> fizzie: you stupid little!? why? can't you just settled down and had a nice girlfriend.
14:34:48 <AnMaster> fizzie, which fingerprints do you use in fungot?
14:34:48 <fungot> AnMaster: all right, it's been a horrific battle. the receptionist. yeah well, good luck, cloud. will guarantee your livelihood once the reactor keeps drainin' up!
14:35:03 <AnMaster> STRN, SOCK, FILE, FING I remember
14:35:03 <ehird> is this still the penny arcade dataset?
14:35:15 <fizzie> ehird: Final Fantasy 7 script.
14:35:42 <fizzie> I keep FING, STRN, SOCK and SCKE loaded all the time; then I use FILE here and there, TOYS for ^reload (the space-clearing part) and SUBR for ^code.
14:38:05 -!- Judofyr has joined.
14:44:19 <ehird> fizzie: you have only ever said 3968 things in #esoteric
14:44:25 <ehird> until optbot was set up
14:44:26 <optbot> ehird: but i would need to try on paper
14:44:36 <ehird> and until optbot was set up, only 58 questions
14:50:56 <fizzie> My own logs list me 4040 comments made with the nick 'fizzie' on this channel before 2008-08-01. That's reasonably close; there's things like splits and such, and I might be counting something wrong too.
14:51:15 <fizzie> And 8802 lines in total, not counting this one.
14:51:25 <fizzie> Apparently I've been quite noisy lately.
14:51:52 <ais523> yes, fungot's been driving traffic
14:51:52 <fungot> ais523: he---y!! here i come, come, and my pay? don't gimme that!? who... who are you saying?
14:54:01 <ehird> I remember when optbot was the traffic-driver
14:54:01 <optbot> ehird: Cannot allocate memory
14:54:04 <ehird> and people conversed with HIM
14:54:19 <fizzie> Well, fungot's decidedly optbot-inspired.
14:54:19 <fungot> fizzie: you can't fool me, liar! maybe we shouldn't stay in here now! ...oops!
14:55:40 <fizzie> And apparently wants out of the channel. Go figure.
15:36:31 <GregorR> (For those who have no idea what I'm talking about, help me train a neural net to recognize whether colors go together by going to http://home.codu.org/colormatch/ )
15:37:21 <ais523> GregorR: what's the copyright status of your hat photos on choosemyhat.com?
15:37:59 <GregorR> ais523: Never thought about it ... ask permission before using, don't use for obscene purposes.
15:38:13 <ais523> well, I'm not planning to for the moment
15:38:32 <ais523> it was just that I was thinking about a programming project which might theoretically some time in the far future need photos of hats
15:38:36 <ais523> and it reminded me of you
15:41:15 -!- Corun__ has joined.
15:43:20 -!- jix has joined.
15:47:51 <GregorR> My neural nets have sex now, btw.
15:48:18 <GregorR> (And being able to say that is the #1 reason to add sexual reproduction to a genetic algorithm)
15:50:07 <GregorR> The sex system is already up to 68% :)
15:50:26 <oklopol> hmm, kinda intriguing, just read a chapter about minimax search for chess plus just a few really simple optimization rules i could easily have come up with myself; then "with alpha-beta search we get to about 10 ply, which results in an expert level of play"
15:50:35 <GregorR> (Due to skew in the input set 61% is free)
15:50:41 <ais523> 10 ply isn't an expert level of play
15:50:44 <oklopol> alpha-beta pruning is this trivial technique for pruning branches minimax will never consider
15:50:59 <ais523> I never got round to implementing it in my own chess program, though
15:51:08 <ais523> but it's pretty trivial
15:51:32 <ais523> it depends on the evaluation function
15:51:37 <ais523> if it's just evaluating material, rubbish
15:51:53 <ais523> you'll survive most tactics, but can easily end up cornered strategically
15:52:01 <oklopol> the evaluation function here is simply counting the amount of pieces, possibly after doing singular extensions
15:52:30 <ais523> I imagine it would fall to a strong aggressive attack, possibly one that throws away material
15:52:32 <oklopol> yeah, just evaluating material
15:52:42 <ais523> or possibly a positional opening trap
15:53:04 <oklopol> well with singular extensions it becomes at least a bit harder to trap it
15:53:14 <oklopol> don't know how hard, i'm not actually that good at chess.
15:54:03 <oklopol> but this is AIAMA, i hear it's considered quite a good book, so i believe what it says.
15:54:18 <ehird> AIAMA is good, yes
15:54:23 <fizzie> Minimax with alpha-beta is what *everyone* (something like 25 out of 30) did for the AI course project-work, which I had to grade.
15:54:26 <ehird> despite me having not read it
15:54:34 <ehird> i have had enough approvals from cool people
15:54:37 <fizzie> That was our course book, too.
15:54:51 <fizzie> Although I think the acronym used was just AIMA.
15:55:02 <ehird> yeah, nobody calls it aiama
15:55:10 <ehird> http://aima.cs.berkeley.edu/
15:55:36 <ehird> cool ,peter norvig is directory of research at google
16:00:19 <fizzie> fungot: Don't you wish you had a real AI brain too? I can loan you that book if you want to write yourself one.
16:00:19 <fungot> fizzie: and with them bringing in the world...
16:16:40 <oklopol> it seems my back can't take sitting.
16:17:08 <mbishop> you're supposed to sit on your butt, not your back
16:20:18 <oklopol> hurts so much i don't even find that funny
16:20:31 <ais523> oklopol: probably you're on the wrong type of chair, then
16:20:43 <oklopol> holy fucking shit... i think i should lie down
16:20:55 <oklopol> perhaps i'll try my armchair.
16:20:58 <ais523> beds aren't really designed for sitting on
16:21:18 <ehird> i have awful posture
16:22:33 <oklopol> i've always had an awful posture, for instance people think i'm quite short, because i'm usually crouching some 10 centimeters down
16:23:10 <oklopol> have had to stand and sit a bit more ergonomically, as my back seems to be starting to... well, die.
16:23:31 <oklopol> and once again i forget a crucial verb
16:23:56 <oklopol> asdads, well at least the armchair helped a bit, thanks for making me realize i have a chair.
16:32:35 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*):*:*:*:*^
16:32:36 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
16:32:39 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*):*:*:*^
16:32:39 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko
16:32:45 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*):*:*:*^^
16:32:45 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko
16:32:53 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*):*:*:*^!^
16:32:53 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokoko okokokokokoko okokokokoko okokokoko okokoko okoko oko o
16:32:59 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*):*:*:*^!!^
16:32:59 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokoko okokokokoko okokokoko okokoko okoko oko o
16:33:04 <ais523> ah, that's what I was aiming for
16:38:51 <ais523> mbishop: have you never seen towers of oko before?
16:38:56 <ais523> of course, it would be better with newlines
16:39:08 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*)::*:*:**^!!^
16:39:09 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokoko okokokokokoko okokokokoko okokokoko okokoko okoko oko o
16:39:19 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*)::*:**:*^!!^
16:39:19 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokoko okokokokokokoko okokokokokoko okokokokoko okokokoko okokoko okoko oko o
16:39:36 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*):*:*:*:*:*^!!^
16:39:38 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
16:39:39 <oklopol> i can almost read that, but what are the !'s all about?
16:39:43 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*)::*:*:*:**^!!^
16:39:43 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
16:39:47 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*):*:*:*:*^!!^
16:39:48 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko ...too much output!
16:39:52 <ais523> +ul ()(o )(:S:a(S^)*~(ok)~*):*:*:*^!!^
16:39:52 <thutubot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokoko okokokokoko okokokoko okokoko okoko oko o
16:40:01 <ais523> oklopol: the first one gets the working string off the stack
16:40:12 <ais523> and the second one gets rid of the repeat of the longest okokoko
16:40:14 <ais523> so it has one peaks not two
16:41:01 <ais523> anyway, the programming technique I used there is one that I don't think I've seen used in any other language
16:41:01 <oklopol> (i actually said an "ah" out loud too, except it was because of my back, and more like "AGHHHHHHH")
16:41:32 <oklopol> basically you store continuations in an explicit data structure
16:42:10 <ais523> yes, it's a bit like that
16:42:24 <oklopol> hmm, i'm seriously considering seeing a doctor. and that is not something i do lightly.
16:43:27 <oklopol> guess i could take those pills that reduce pain, too, but that feels like cheating
16:49:09 <fizzie> "continuations in an explicit data structure" sounds very much what I did in the Prolog-Scheme.
16:49:38 <ais523> incidentally, the Underload divmod-by-constant I wrote used a similar trick
16:49:43 <oklopol> okay i cannot code with this back.
16:49:45 <ais523> to divide by 10, it copied a program n times on the stack
16:49:56 <ais523> each of which popped the 9 elements below it and ran the 10th
17:08:11 <psygnisfive> im going to work on a context free grammar for chinese stroke order
17:11:16 <GregorR> I don't think I can do better than 71% :(
17:11:19 <GregorR> At least not with this data set.
17:11:58 <ehird> i will help you with neural net stuff
17:12:00 <ehird> beacuse i like neural nets
17:12:16 <GregorR> http://home.codu.org/colormatch/ Help me by generating data :P
17:12:52 <ehird> i did but i'll continue
17:12:54 <ehird> GregorR: plz add accesskeys
17:13:04 <ehird> a,s,d for the three buttons respectively
17:13:08 <ehird> GregorR: accesskey="a"
17:13:13 <ehird> on the button elements
17:13:34 <ehird> then alt-key or ctrl-key(on os x) activates it
17:13:38 <ehird> you can do it on input fields too
17:15:12 <ehird> how many trains should I do?
17:15:35 -!- metazilla has joined.
17:15:41 -!- moozilla has quit (Nick collision from services.).
17:16:09 -!- moozilla has joined.
17:16:37 <GregorR> However many you'd like until you get bored :P
17:17:09 <ais523> GregorR: you got some training from me too
17:17:52 <ehird> GregorR: I trained it a bit
17:17:55 <ehird> how does it do now?
17:18:10 <fizzie> I would click them buttons, but I have no clue what colors "match".
17:18:23 <ehird> fizzie: which look aesthetically pleasing together?
17:18:29 <ehird> which colours "go" together?
17:18:40 <fizzie> I know the meaning, I just can't really tell.
17:19:23 <ehird> fizzie: Just click yes if you like how it looks and no if you don't :P
17:19:41 <GregorR> fizzie: That's my problem, that's why I wrote this :P
17:19:48 <fizzie> That's just it, I'd end up doing "can't decide" on just about anything, with maybe a few "no"s in there.
17:19:49 <GregorR> fizzie: I want a computer to tell me if my tie goes with my shirt.
17:20:37 <ehird> fizzie: are you colourblind?
17:20:45 <ehird> If not, I think it's very easy to say "that's pretty" or "that's ugly".
17:20:49 <fizzie> ehird: No, just bad at making decisions.
17:20:50 <ehird> For something that simple
17:21:19 <ehird> GregorR: Does the script only ask for opinions on ones it thinks are good?
17:21:30 <ehird> and make a seperate ones for ones it doesn't like
17:21:39 <ehird> so it's easier - "yep, that's right" or "no, that's wrong"
17:24:16 -!- megatron has quit (Read error: 110 (Connection timed out)).
17:26:52 <GregorR> The script is completely unaware of the neural net.
17:26:56 <GregorR> The script just collects data.
17:31:38 <ehird> GregorR: make it aware
17:31:59 <ehird> GregorR: better idea
17:32:04 <ehird> make it only give ones the neural net isn't sure about
17:32:27 <GregorR> Yeah, it's a good idea, it's just a PITA because I didn't design it that way :P
17:32:52 <ehird> GregorR: well do it :P
17:35:47 <GregorR> Incidentally, do you agree with my assumption that this should be determinable by a computer?
17:38:31 <GregorR> Good, because I want a computer to tell me if my tie matches my shirt, damn it :P
17:39:18 <ehird> GregorR: now modify the script
17:43:05 <GregorR> I don't think you know how difficult that would be.
17:43:09 <GregorR> I would have to implement a neural net in PHP
17:43:20 <GregorR> Which isn't complicated ...
17:43:32 <GregorR> (The 'n' is important :P )
17:43:50 <pikhq> Gregor, that's brilliantly clever.
17:43:55 <GregorR> Besides, to be honest I doubt that would help all that much.
17:43:57 <ehird> GregorR: no you dont
17:44:04 <ehird> just interface with a commandline program
17:44:24 <GregorR> ehird: Presently the commandline program only knows how to evolve things and give their statistics :P
17:44:32 <ehird> GregorR: tweak it a tiny bit.
17:44:52 <GregorR> ehird: Besides, I would have to interface with the /currently running one/, which is IPC.
17:45:58 <mbishop> a computer to determine your fashion? brilliant!
17:46:57 <mbishop> Althought that wouldn't help me, as I don't wear clothes
17:47:13 <GregorR> Yay for text-based communication protocols.
17:50:16 -!- puzzlet has quit (Remote closed the connection).
17:50:30 -!- puzzlet has joined.
17:51:41 <pikhq> GregorR: Neural net to see if colors match.
17:52:46 <GregorR> Ah. When I say I want a computer to tell me if my tie matches my shirt, I'm not making a joke. That is really, truly the reason I wrote this :P
17:53:01 <ais523> GregorR: after all, you get a computer to choose your hat...
17:53:05 <ais523> (via human input, though)
17:53:22 <GregorR> I used to get a computer to do it totally randomly, but that wasn't democratic enough ;)
17:53:23 <pikhq> I'm not saying that it's a clever joke.
17:53:28 <pikhq> I'm saying that it's clever.
17:53:33 <pikhq> Big difference. ;)
17:56:20 <GregorR> I'm just trying to put it into its ridiculous context :P
18:05:45 -!- oerjan has joined.
18:10:25 -!- jix has quit (Read error: 113 (No route to host)).
18:16:28 -!- optbot has set topic: the entire backlog of #esoteric: http://tunes.org/~nef/logs/esoteric | a interpreter/VM/emu could store the compiled code in a cache, meaning that a program would only need to be compiled once.
18:20:34 <ehird> http://divingintointercal.blogspot.com/
18:20:45 <ehird> ais523: http://divingintointercal.blogspot.com/
18:36:51 <oerjan> "The bad news is that the previous sentence is the only good news." :D
18:41:48 <oerjan> "As a programming language, INTERCAL remains every bit as useful as it was over thirty years ago."
18:45:55 <ais523> ehird: I've seen it before
18:45:57 <ais523> has there been a new entry?
18:47:26 <ais523> nope, no new entries since I last saw it
18:47:29 <ais523> I'm worried it's dead...
18:50:13 -!- Corun_ has joined.
18:50:35 <ehird> ais523: last post 07
18:51:08 <ais523> ehird: this is INTERCAL we're talking about, I suspect it requires approx. 12 years before it can truly be considered dead
18:54:28 -!- olsner has joined.
19:05:33 -!- Corun__ has quit (Read error: 110 (Connection timed out)).
19:06:19 -!- Thomsen has joined.
19:06:29 -!- Thomsen has left (?).
19:25:46 <fizzie> Have you bothered to check how many % you'd get if you didn't use your fancy perverted sex-obsessed neural networks, and just used something like a mixture-of-gaussians model estimated from the "pair of colors -> goes-togetherness" data and a fixed threshold to get yes/no out of it?
19:27:00 -!- oerjan has quit ("Antabus").
19:28:42 <fizzie> Maybe it would be significantly less awesome, though.
19:32:09 <olsner> fancy perverted sex-obsessed neural network == brain, right?
19:33:10 <fizzie> No, I think GregorR's having sex with his artificial neural networks, too. Or making them have sex together. Or some other depraved thing, anyway.
19:41:38 <fizzie> Not that I know of, except that I still haven't fixed out an output length limiter in it.
19:44:25 <AnMaster> I just brought a ethernet switch today... opened the box... why the heck is there a cd in it saying "<brand name of switch>\nMy Digital Life" on it
19:44:42 <ais523> because everything comes with random Windows programs nowadays
19:44:52 <ais523> some of them are actually OK, but generally speaking you can throw them away
19:45:07 <AnMaster> well it doesn't even say what's on it really
19:45:19 <AnMaster> manual is my best guess since there is none elsewhere
19:46:57 <AnMaster> ais523, I just can't see how there could be any windows program related to the switch, it is a consumer one, so no webui or settings or such
19:47:07 <ais523> AnMaster: it doesn't have to be /related/
19:47:15 <ais523> my guess is some sort of digital photo album
19:47:43 <AnMaster> well the cd have the name of the product on it too
19:48:06 <AnMaster> the box of the product however says "independent of operating system" hehe
19:48:16 <ais523> maybe it's an audio CD
19:49:23 <AnMaster> well I'll check later, for now I got to move a few computers around, I may lose connection shortly (or it may work without dropping the connection)
19:50:54 <fizzie> Decided to be brave and just check how much I mess up if I try to add that output length limit to the underload interp without any testing.
19:51:17 <ais523> did you limit output length to 1?
19:51:29 <ais523> ^ul (Hello, world!)S( Hello, again!)S
19:51:30 <fizzie> No, not that I know of, and in any case it should add a "... too much output!" after it.
19:51:39 <AnMaster> grr what a cable mess behind the computer
19:51:40 <ais523> it's ending after the first S instruction
19:51:41 <fizzie> Seems like it just stops at the first S now.
19:51:55 <AnMaster> and several unconnected cables
19:51:58 <oklopol> AnMaster: how many computers do you have?
19:52:28 <AnMaster> and the other one is using a temporary 50 meter ethernet cable I happened to have around
19:52:42 <AnMaster> but that just doesn't work well, you can't close the doors
19:52:51 <AnMaster> so I bought a switch to have in this room
19:53:05 <fizzie> I've got my ` test backwards, heh. And the "too much output" just gets lost because it's added too far away.
19:53:27 <fungot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ...too much output!
19:53:37 <thutubot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ...too much output!
19:53:46 <ais523> your output limit is slightly longer than mine, I think
19:54:00 <ais523> I can't remember what mine is offhand
19:54:09 <ais523> but I think that it's at least written in decimal, so I could check
19:54:16 <ais523> my guess is 255 or 256, because I'm like that
19:54:43 <fizzie> It's also longer than my brainfuck limit.
19:54:44 <fizzie> ^bf ++++++++++[>++++++++++++<-]>[.]
19:54:44 <fungot> xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ...
19:54:56 <fizzie> That one's only ff*, so 225.
19:55:57 <fizzie> (To be entirely accurate, the limit is 777** characters for the entire IRC message, including the "PRIVMSG #esoteric :" part.)
19:56:12 <fizzie> The oko program should now be safe to run:
19:56:14 <fizzie> ^ul (o)(~:S(ko)*( )S~:^):^
19:56:14 <fungot> o oko okoko okokoko okokokoko okokokokoko okokokokokoko okokokokokokoko okokokokokokokoko okokokokokokokokoko okokokokokokokokokoko okokokokokokokokokokoko okokokokokokokokokokokoko okokokokokokokokokokokokoko okokokokokokokokokokokokokoko okokokokokokokokokokokokokokoko okokokokokokokokokokokokokokokoko okokokokokokokokok ...too much output!
19:56:32 <ais523> fizzie: what limits do you have?
19:56:58 <ais523> thutubot has number of iterations of the main loop (a time limit), amount of output, and amount of memory used
19:57:09 <ais523> main loop iterations doesn't easily correspond to commands, by the way
19:57:31 <fizzie> Stack length (10k characters), program length (if it tries to extend too far "to the left" -- but I'm not sure that works, I haven't hit it yet), amount of commands executed, and that output limit.
19:57:56 <ais523> fizzie: shall I come up with a massively extending program for you to test program length?
19:58:23 <ais523> oklopol: that extends stack not program
19:58:28 <ais523> and thutubot will run it, but slowly
19:58:40 <thutubot> SSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSS ...too much output!
19:58:58 <fizzie> I think something like (((longsillything)!)~*:^):^ should grow, but that just results in a "out of time" thing.
19:59:29 <ais523> the stack limit must be a lot shorter than the program limit...
19:59:56 <fizzie> I guess it is, in fact.
20:00:21 <ais523> it's much harder to get a program that blows up exponentially if you can't put it on the stack
20:00:37 <fizzie> Well, I think a linearly extending program should work too.
20:00:47 <ais523> I'll try linearly extending
20:01:02 <ais523> ^ul (:^(foo foo foo foo foo foo foo foo foo foo foo foo foo foo foo foo foo)):^
20:02:11 -!- puzzlet has quit (Read error: 60 (Operation timed out)).
20:02:15 -!- puzzlet has joined.
20:03:21 <AnMaster> while rc/funge got lots of odd limits, for cfunge the only limits would/should be how large size_t is
20:03:48 <fizzie> I think it's still possible to create quite a short program, if you first fill the stack with two few-kilobyte strings, and then execute (~:^):^
20:03:55 <AnMaster> (and off_t or whatever it is that you use for files. can't remember)
20:04:07 <fizzie> Er, substitute "short" with "long-running" there.
20:04:13 <ais523> AnMaster: rc/funge's limits are things like 1000, so they're even not odd
20:04:44 * AnMaster don't feel like joking atm, got a fever so probably heading to bed soon
20:05:01 <ais523> sorry, oerjan seems to be inactive and someone has to make the bad puns
20:05:10 <fungot> foofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoofoo ...too much output!
20:05:20 <ehird> Ha, I beat the repressive linear growth stopper.
20:05:45 <ais523> ehird: the first program doesn't grow linearly
20:06:03 <ehird> (:::^^^):^S <- shouldnt this be 3^3
20:06:19 <ais523> ehird: :::^^^ is something silly like 3 nested infinite loops
20:06:29 <ehird> um what are numbers then
20:06:49 <ais523> to output a number in unary use (x)~^S
20:06:56 <ehird> ais523: ah, thanks
20:07:00 <ehird> ^ul (::**):^(x)~^S
20:07:00 <fungot> xxxxxxxxxxxxxxxxxxxxxxxxxxx
20:07:27 <ais523> 1 is the null string, always fun
20:07:29 <oklopol> :::^^^ means roughly f(f)(f)(f)
20:08:00 <ais523> to be precise, it would be \f.f(f)(f)(f)
20:08:06 <ais523> if not for the fact that the stack could change in the meantime
20:08:41 <ehird> ^ul ()(~:(x)~^S( )S(:*)*~:^):^
20:08:41 <fungot> x xx xxxx xxxxxxxx xxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx ...too much output!
20:08:54 <ehird> ais523: um what is that outputting?
20:08:57 <ehird> it should be x xx xxx xxxx
20:09:01 <oklopol> +ul ()(:*)(::**)(:::***):::^^^(x)~^S
20:09:06 <ais523> oklopol's program outputs an insanely large number
20:09:10 <ais523> oklopol: let the first one end first
20:09:13 <ehird> ()(~:(x)~^S( )S(:*)*~:^):^
20:09:22 <ais523> yours is doing powers of 2 I think
20:09:28 <ais523> now i have to work out why
20:09:30 <ehird> I just put :* at the end each time
20:09:38 <ehird> do i have to do ::**
20:10:06 <ais523> and you do addition by adding 1 in a loop
20:10:12 <AnMaster> ais523, if everything comes with a cd these days, how comes the mobile phone (cell phone? Or is that US Eng.?) that *can* connect to a computer didn't came with a cd
20:10:13 <ehird> ais523: great, it went over my head
20:10:19 <ehird> I was just starting to "get" underload.
20:10:41 <oklopol> ehird: basically, if you have ::** and you want to make it :::***, just do exactly what that says, concatenate a : and a *
20:10:54 <ais523> so (:)~* to put a : at the start
20:10:59 <ais523> and (*)* to put a * at the end
20:11:08 <ehird> ^ul ()(~:(*)~^S( )S(:)~*(*)*~:^):^
20:11:09 <fungot> * ** *** **** ***** ****** ******* ******** ********* ********** *********** ************ ************* ************** *************** **************** ***************** ****************** ******************* ******************** ********************* ********************** *********************** ************************ ...too much output!
20:11:14 <oklopol> i guess ais523's was actually less spoiling, because if you can't *program* that, you definitely can't *read* it.
20:11:37 <ais523> esolangs have spoilers nowadays?
20:12:17 <oklopol> ais523: i meant if ehird asks how something is done, it's spoiling if you don't hint, but just write the program
20:12:42 <ehird> what about subtraction?
20:12:54 <oklopol> "doing their homework" on channels with a subject that's actually taught somewhere :P
20:13:02 <oklopol> ehird: that's quite simple
20:13:08 <ais523> subtraction is a pain, oklopol worked it out for emself eventually, and asiekerka gave up
20:13:11 <ais523> it's quite simple if you know how
20:13:17 <ais523> but difficult to come up with in the first place
20:13:24 <oklopol> you know what :::*** actually does?
20:13:30 <ehird> duplicates 3 times, conc...
20:13:50 <ehird> call the number, on (:)
20:13:51 <oklopol> now, after :::, do something, before ***, do something, you can make ::: and *** out of running :::***
20:13:58 <ehird> and concatenate all that together
20:14:13 <ais523> oklopol: incidentally, there's an entirely different way to write +1 in Underload: :(:)~^~(*)~^*
20:14:27 <oklopol> how do you drop one from a string of *'s?
20:14:44 <oklopol> you have to come up with something that forms identity when used with *
20:15:10 <ehird> ^ul (::::::::::**********)(~ (*)~^S( )S :::(:)^!(*)^!()(*)^ ~:^):^
20:15:13 <oklopol> ais523: yeah, that's actually a good way to get substraction
20:15:49 <ais523> oklopol: ah, it would be, I don't think I've ever done it like that though, but probably it would be a computational order more efficient than the way I normally do it
20:16:00 <ais523> whereas for addition it's a computational order less efficient
20:16:05 <ehird> ^ul (::::::::::**********)(~(*)~^S( )S:::(:)^!(*)^!()(*)^~:^):^
20:16:05 <fungot> *********** ** ...out of stack!
20:16:29 <oklopol> ais523: you talking about subtraction or addition?
20:17:29 <oklopol> ehird: :::*** actually makes a (::::) when you call it on (:), so it's not actually drop *one*
20:17:50 <ehird> i don't kbnow what to do then
20:17:51 <ais523> oklopol: I'm talking about that technique for both, it makes subtraction faster but addition slower
20:18:04 <ais523> as for subtraction you can generate lots of !s and ()s
20:18:07 <oklopol> ais523: what do you usually do for subtraction?
20:18:22 <oklopol> first, you use the number on (:), that's a given
20:18:42 <oklopol> but you actually want the effect of two less :'s
20:18:46 <oklopol> so what do you need to put here?
20:19:14 <oklopol> you have(:::***) generate ::::
20:19:50 <oklopol> now, you cannot make "two less *'s than the number"
20:19:55 <ehird> ^ul (::::::::::**********)(~(*)~^S( )S:::(:)^!(*)^(!!)*()(*)^~:^):^
20:19:56 <fungot> *********** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ...too much output!
20:20:01 <oklopol> so you need to make four *'s as well
20:20:24 <oklopol> the problem is, with four *'s on three (x)'s, you concatenate random crap into them
20:20:38 <ehird> so put something on there first
20:20:39 <oklopol> so you need to add something more in the middle of the number
20:20:47 <ehird> ^ul (::::::::::**********)(~(*)~^S( )S:::(:)^!(*)^(!!())*()(*)^~:^):^
20:20:47 <fungot> *********** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ...too much output!
20:20:52 <ehird> ^ul (::::::::::**********)(~(*)~^S( )S:::(:)^!(*)^(!!()())*()(*)^~:^):^
20:20:52 <fungot> *********** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ...too much output!
20:21:04 <oklopol> exactly as many as you have there.
20:21:14 <ehird> so wher be my fuckup
20:21:32 <oklopol> **** run on three (x)'s concatenates exactly two pieces of random crap into it
20:21:32 <ehird> ^ul (::::::::::**********)(~:(*)~^S( )S:::(:)^!(*)^(!!)*()(*)^~:^):^
20:22:46 <oklopol> (:)^ this is equivalent to just :
20:23:00 <oklopol> that would make the :'s you need
20:23:10 <ehird> ^ul (::::::::::**********)(~:(*)~^S( )S:::(:)~^!(*)^(!!)*()(*)~^~:^):^
20:23:47 <ais523> ehird: ()(*)~^ is clearly not what you want
20:23:48 <fizzie> ^ul (:::::*****)(~:(:)~^(!!)*(()())*~(*)~^*:(x)~^S( )S~:^):^
20:23:49 <fungot> xxxxx xxxx xxx xx x ...too much output!
20:24:05 <ehird> ^ul (::::::::::**********)(~:(*)~^S( )S:::(:)~^!(*)^(!!())*(*)~^~:^):^
20:24:06 <ais523> fizzie: that simplifies a bit
20:24:06 <fungot> *********** * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * ...too much output!
20:24:31 <ehird> its running it on the weird !!() thing
20:24:38 <ehird> ^ul (::::::::::**********)(~:(*)~^S( )S:::(:)~^!(*)^(!!())*~(*)^~:^):^
20:24:39 <fungot> *********** ...too much prog!
20:24:55 <oklopol> ehird: what's the first ! about?
20:25:12 <oklopol> you create an concatenated piece of :'s on the stack, and then drop it? :D
20:25:26 <oklopol> you shoud comment your underload
20:26:10 <fizzie> My second thought was to do something like
20:26:11 <fizzie> ^ul (:::::*****)(~:(:)~^(!!!)*(()()())*~(*)~^*:(:)~^~(*)~^*:(x)~^S( )S~:^):^
20:26:11 <fungot> xxxxx xxxx xxx xx x *x ...out of stack!
20:26:19 <ehird> i'll write a more interesting program
20:26:20 <fizzie> That's not pretty either.
20:26:36 <ehird> i wonder how i could do dip
20:26:56 <ehird> dip : 'R 'a ('R -- 'S) -- 'S 'a
20:26:57 <ais523> dip is very neat in Underload
20:27:01 <ehird> ais523: let me work it out
20:27:21 <oklopol> +ul (::::::******)(~:(*)~^S( )S:(:)~^(!!()())*~(*)~^*~:^):^
20:27:23 <thutubot> ******* ****** ***** **** *** ** * ...too much output!
20:27:35 <ais523> oklopol: get it to terminate when it's finished?
20:27:43 <oklopol> yeah was just thinking that
20:28:40 <ehird> ais523: dip = ~a*^
20:29:54 <oklopol> +ul (::::::******)(~:(*)~^S( )S:(:)~^(!!()())*~(*)~^*:()~(~:^)~(!()())~(a)~^^!~!^):^
20:29:56 <thutubot> ******* ****** ***** **** *** ** * ...too much output!
20:30:02 <ehird> (foo) (bar) (koed) -> (foo) (koed) (bar) -> (foo) (koed) ((bar)) -> (foo) (koed(bar)) -> voila
20:30:16 <ehird> i didn't even work ~a*^ out on paper, a first for me
20:30:29 <ehird> i had to close my eyes to concentrate enough though
20:43:19 -!- lilja has joined.
20:43:41 <fizzie> fungot now trained with the intercal manual; but I couldn't figure out how to generate nice non-wrapped text out of groff, so...
20:43:41 <fungot> fizzie: please abstain from gerund list, line label of which was abstained from. however, as these routines might
20:43:54 <fizzie> So it stops dead in the middle of sentences.
20:43:59 <ais523> fizzie: the Revised Manual, or the Revamped Manual?
20:44:09 <ais523> presumably Revised as it was groff
20:44:17 <fizzie> Whatever was in the doc/ directory of c-intercal 0.2.4 or something.
20:44:32 <fizzie> There's a better manual I could've used?
20:44:37 <ais523> yes, there's a new one
20:44:45 <fizzie> Ooh, that sounds better.
20:44:50 <ais523> but it's only in the more recent versions
20:45:18 <ais523> fizzie: http://code.eso-std.org/c-intercal/doc/ick.txi is the source
20:47:40 <fizzie> Let's see what reading that does to fungot.
20:47:41 <fungot> fizzie: maybe don't give your name ( in other programming languages.
20:47:51 <fungot> ais523: intercal-72 c-intercal clc-intercal j-intercal yes all versions
20:48:06 <ais523> it's the "portability" boxes
20:48:13 <ais523> there are lots of them
20:48:17 <fizzie> fungot: Could you ignore that?
20:48:17 <fungot> fizzie: copyright ( c). to install elsewhere ( in this way. note that the program can either decide that it needs to be used; a default value will be available to unpack the installation directory, so otherwise there would be executed.
20:48:32 <ais523> fungot: more Revamped Manual info?
20:48:33 <fungot> ais523: however, clc-intercal uses lowercase to represent multiplication by 1000000 writes a backslash before the current directory by typing `cd ick-0.29') on the user to find the appropriate cross-compiler to translate its programs from the internet, including any `once' or `next from's, `next from's aiming at a command whose line label that the compiler doesn't use them.)
20:49:08 <fizzie> That didn't make much sense.
20:49:16 <ais523> fungot: make more sense
20:49:16 <fungot> ais523: this is tested mostly with that compiler, what actually happens, then the file it was hardcoded as a function ( for instance, `-h' ( currently the only way to pop the top of stack.
20:49:33 <fizzie> So "-h" is the only way to pop. Right.
20:51:31 <fizzie> I don't think the bot is being a very useful intercal reference; I guess I'll switch to some other text.
20:51:49 <fungot> ais523: there is a fatal error ( e000) due to the error that would otherwise not have to simplify the program will end up with a higher character code; the ' marker'; a marker is represented as a thread, one digit at a labeled statement, and c-intercal version 0.29).
20:52:29 <ehird> fizzie: Feed it the first million digits of pi, make it always start on 3.
20:52:42 <ehird> and make it based on 2 chars
20:54:07 <fizzie> Wouldn't that just make it a less-than-stellar RNG?
20:54:41 <ehird> fizzie: Well, maybe.
20:54:46 <ehird> fizzie: Make it order=10 chars.
20:54:52 <ehird> It'd look like pi, unless you knew pi.
20:55:24 <fizzie> I'd think any string of random digits would look like pi, as long as it "starts right".
20:55:36 <ais523> there's that Pi programming language, isn't there
20:55:45 <ais523> which encodes the program as subtle errors in digits of the number pi
20:57:37 <fizzie> In the wiki there's also "Another Pi Language", where the source code is two arbitrary integers; first is the index in pi and second is the amount of digits to read; that is then interpreted as "source file of any language". Unsurprisingly unimplemented.
20:57:59 <ais523> that language should itself be Pi, obviously
20:58:13 <fizzie> Pi seems to be the errors-in-pi one.
20:58:26 <ais523> the "source file of any language" should be Pi source
20:58:38 <ais523> thus you have to find a Pi program embedded in Pi
20:58:49 <fizzie> Yes, I understood that.
20:59:28 <fizzie> At least the Pi article has an implementation that will convert brainfuck into it.
21:18:21 <AnMaster> this new phone only have weird sounds, no classical beeping ones
21:18:28 * AnMaster liked that with his old old phone
21:18:33 <ais523> that's common practice nowadays
21:18:38 <ais523> you have to /pay/ if you want beeps
21:18:43 <ais523> download them from a beep website
21:18:45 <AnMaster> I guess I could make one and transfer it
21:19:14 <AnMaster> ais523, I don't have internet on my phone. Only pay for a cheap connection
21:19:29 <ais523> even then there are numbers you can ring for that sort of thing
21:19:31 <AnMaster> since all I need is to make calls and to send sms
21:19:37 * ais523 doesn't have a mobile at all
21:20:01 <AnMaster> my old phone was an old one with black and white screen. However, you could make your own tunes on it
21:20:20 <ais523> you used to be able to send texts for free
21:20:32 <ais523> but the phone companies realised people would pay for the privilege
21:20:36 <ais523> so they started charging
21:20:40 <ais523> that was a few years ago now
21:20:46 <AnMaster> on my old phone I could make my own beeping sounds
21:20:55 <AnMaster> it didn't even have non-beepy ones
21:21:22 <AnMaster> why can't I just make my own beepy ones on it
21:21:31 <ais523> because people will pay to download them
21:21:43 <ais523> well, some people will
21:21:49 <ais523> and that's all the phone compaines care about...
21:21:58 <AnMaster> I will try to figure out how to export to the *.aac format and then find a laptop with bluetooth to transfer it
21:22:33 <AnMaster> I *can* play the piano and I do have a keyboard + midi cables
21:23:16 <ehird> .aac = apple's format
21:23:34 -!- ab5tract has joined.
21:23:35 <AnMaster> hm wikipedia says "ISO/IEC 13818-7:2003"
21:23:35 <ehird> AnMaster: if you give me a bunch of files i can make them into aacs
21:23:38 <AnMaster> http://en.wikipedia.org/wiki/Advanced_Audio_Coding
21:23:39 <ehird> well, yeah, it's standard
21:23:44 <ehird> but itunes uses it by default
21:23:47 <ehird> as well as the itunes store
21:23:55 <AnMaster> ehird, I would be surprised if there is no tool to convert for linux
21:24:12 <ehird> there seem to be ways to convert FROM it
21:24:15 <ehird> but not to convert TO
21:25:06 <AnMaster> that's what I used on my old phone anyway.
21:26:23 <AnMaster> hm how to render the midi file to beepy sound like a phone.
21:26:24 <ehird> AnMaster: you could just use a ringing sound.
21:26:31 <ehird> also, what phone is it
21:27:37 <ehird> AnMaster: are you sure it doesn't support midi?
21:28:04 <AnMaster> ehird, no, I'm not sure, I just checked what format the existing files were in
21:28:11 <ehird> AnMaster: try and put the midi on.
21:28:41 <AnMaster> ehird, my computer lacks bluetooth, so I'll need to try it later when I get access to a laptop with bluetooth
21:29:06 <AnMaster> ehird, "cable not included with phone" and I didn't think I would need it
21:29:11 <AnMaster> + they didn't have it in stock
21:29:42 <AnMaster> memory card: no I don't have any micro-sd reader or cards, my camera use compact flash
21:29:51 <AnMaster> which of course would be too big for a phone
21:30:47 <AnMaster> and I don't have such a reader either
21:32:10 <AnMaster> why can't they make phones like my old nokia 2100 these days? :(
21:32:19 -!- puzzlet has quit (Read error: 104 (Connection reset by peer)).
21:33:12 -!- puzzlet has joined.
21:44:30 -!- jix has joined.
21:45:35 <ais523> it seems Keymaker just proved Sceql TC
21:52:50 <oklopol> i wish it wasn't just a compilation, i can't exactly reverse-engineer what happens
21:55:04 -!- kar8nga has joined.
22:12:10 -!- ab5tract has quit.
22:17:52 <AnMaster> ais523, that cd with the router: acrobat reader + manual. So it was just badly labeled
22:27:20 <AnMaster> one page manual per language heh
22:27:31 <AnMaster> (wouldn't printing it be easier?)
22:44:30 -!- kar8nga has quit (Read error: 110 (Connection timed out)).
22:45:29 <AnMaster> fizzie, grr I can't find url to fungot
22:45:29 <fungot> AnMaster: guess everyone's here... cloud.
22:45:55 <fungot> ^<lang> <code>; ^def <command> <lang> <code>; ^show [command]; lang=bf/ul, code=text/str:N; ^str 0-9 get/set/add [text]
22:45:59 <fungot> echo reverb rev rot13 rev2 fib wc ul cho choo pow2
22:46:08 <ais523> oh, for a moment I though fungot was a real user, and was trying to figure out what its line meant
22:46:08 <fungot> ais523: no matter how you feel it. then we'll know that's our memory...... calling...... that thing's not human......
22:48:21 <ehird> fizzie: put lovecraft in to it
22:59:39 -!- lilja has quit (Read error: 110 (Connection timed out)).
22:59:57 <comex> fungot: do you still speak Agoran?
23:00:20 <ais523> comex: not any more, fizzie filled it with something else
23:00:27 <ais523> fungot's spouting a computer game script now, I think
23:00:27 <fungot> ais523: i'm... i'm so disgusted with the story about sephiroth.... you really have to kill me!! that young fellow!?
23:11:58 <fizzie> Okay, fungot now has the lovecraft.
23:11:59 <fungot> fizzie: on the other, he knew that it touched on the scenes i had haunted, and i felt myself touched by the contagion of the morbid fnord. to my mind rushed fragments of a more provocative nature. there were cities under the sea.
23:12:59 <fizzie> AnMaster: The raw sources are http://zem.fi/~fis/fungot.b98.txt and if you replace b98.txt with html you'll get the silly only partially done commentation.
23:13:00 <fungot> fizzie: moon over japan, fnord with filth, where a bit of fnord but iranon stayed ever young, and of shadowy companies of heroes that ride fnord white horses along the edges of the aperture itself in my ascent, and managed to avoid any distant glimpse of the broken columns and crumbling sphinxes of that fearful city which lived and died before the years of man.
23:13:33 <AnMaster> ^def source ul (http://zem.fi/~fis/fungot.b98.txt)S
23:13:36 <fungot> http://zem.fi/~fis/fungot.b98.txt
23:13:45 <AnMaster> fizzie, hope you like that idea
23:16:54 <AnMaster> fizzie, also firefox 2 says that a script have stopped responding on http://zem.fi/~fis/fungot.html
23:16:55 -!- megatron has joined.
23:16:56 <fungot> AnMaster: " ia! shub-niggurath! as a foulness shall ye know them,' he asked softly, ' did you see it here... and smell it?' and when mr. ward, it is hard to place the fragmentary dream-glimpses in the fnord lips an all but impassable barrier. from my experience i cannot doubt but that you will want to lead an expedition to the desert and its carven stones.
23:17:03 <AnMaster> wondering if I want to abort it
23:17:03 -!- moozilla has quit (Nick collision from services.).
23:17:20 -!- moozilla has joined.
23:17:27 <AnMaster> fizzie, oh and lovecraft in fungot sucks
23:17:28 <fungot> AnMaster: as the newspapers told, we sailed from boston harbor on september 2nd, 1930, taking a leisurely course down the coast and through the doorway came a faint but peculiarly hateful odor must represent the collected sections of the entity which lake had tried to follow backward the frescoes he had passed the open fnord purple slopes. against these fnord i have proven false to alos, my friend was always far ahead; i could
23:19:32 <AnMaster> fizzie, also does fungot use any other fingerprints apart from FING STRN SOCK SCKE FILE TOYS SUBR
23:19:33 <fungot> AnMaster: he was half crazy himself, and stuff them fnord into an account of his sea voyaging, had a depth and hollowness that could not be applied to analogous qualities in a fnord and promising cat now, and have kept certain doubts and guesses to ourselves with splendid unity and fnord.
23:26:28 -!- metazilla has quit (Read error: 110 (Connection timed out)).
23:28:06 <fizzie> Haven't tried saving underload programs yet.
23:28:10 <fizzie> I don't think it uses any others.
23:28:19 <fungot> (http://zem.fi/~fis/fungot.b98.txt)S
23:28:32 <fizzie> Nice, although it should be in the ^help text optimally.
23:28:48 <fizzie> Of course, given the underload interp I think I could just replace ^help with a program like that.
23:29:12 <fizzie> (The only thing missing would be that currently it gets the '^' command character from whatever you define it in the loader. Oh, well.)
23:29:49 <fizzie> The script does quite a lot of computation; it shouldn't be in an infinite loop, but FF2's JavaScript engine is quite a lot slower than FF3's.
23:30:46 -!- jix has quit ("CommandQ").
23:32:17 <AnMaster> it tooks 3 clicks in that dialog for it to finish
23:32:35 <AnMaster> fizzie, oh you got an fd leak on ^raw quit
23:32:44 <AnMaster> ==16011== Open AF_INET socket 3: 192.168.0.64:38732 <-> 85.188.1.26:6667
23:32:44 <AnMaster> ==16011== at 0x35FAEC1267: socket (in /lib64/libc-2.6.1.so)
23:32:44 <AnMaster> ==16011== by 0x41967E: finger_SOCK_create (SOCK.c:376)
23:34:38 <AnMaster> fizzie, anyway what fingerprints and what functions from which fingerprints are used. Since I consider fungot very important I want to make sure those instructions perform well. For example I rewrote A of STRN to be a bit faster today (could probably be made even faster)
23:34:39 <fungot> AnMaster: the reason why arthur jermyns charred fragments were not collected and buried lies in what was certainly not english. i boarded it and looked vainly about for the light fnord as i was, i could at least bar others from the outside, so the party was somewhat abated. vast walls shot up, and in
23:35:35 <AnMaster> also STRN is kind of weird. I mean it copies between stack and funge space, yet converts them to unsigned char* in between. I mean... what a waste. But I guess using strings of funge_cell could break stuff
23:46:51 <fizzie> Yeah, I don't bother closing that single socket I have.
23:47:05 <fizzie> I do properly close the language model files I open every time someone speaks to the bot.
23:47:25 <fizzie> fungot: Go back to IRC chat for now.
23:47:26 <fungot> fizzie: or simply sub by 47
23:50:52 <fizzie> Let's see... from FILE I use G/P for the state file (not performance-critical at all) and then R/S a whole lot for the babbling; FING Y/Z but not much; from STRN it's mostly G, P, N, L, C, A, S, V approximately in that order of frequent use (so G/P most, S/V pretty much in the state file saving/loading only); ...
23:51:45 -!- Corun__ has joined.
23:52:10 <fizzie> From SOCK it's just R/W most of the time, from TOYS only S to clear the old code when ^reloading (so it might be good if S'ing to value 32 would actually clear those cells), and from SUBR only a C/R pair for ^code.
23:59:09 -!- Corun_ has quit (Read error: 110 (Connection timed out)).