00:32:30 -!- oklopol has quit (Connection timed out).
00:45:36 -!- ehird` has quit (Read error: 104 (Connection reset by peer)).
01:00:54 -!- staplegun has joined.
01:06:15 <staplegun> ... ... ... ... ... ... ... ... ... ... ... ... ... ...|... ... ... ... ... ... ... ... .._,,-~~~-,-,,_
01:06:16 <staplegun> ... ... ... ... ... ... ... ... ... ... ... ... ... ...|... ... ... ,,---,,_... ...,-~",-":__-,: : :"-,:::"'~,,
01:06:19 <staplegun> .,,-~~--,, ... ... ... ... ... ... ... ... ... ... ...|... ... ,~": : --,,:"~,,-"::::/:,-". . ."'~-,,: \,::::_,,"~-,,
01:06:20 <staplegun> /: : : : : : :"'~-,,... ... ... ... ... ... ... ... ... |... ... /: : : : :~"'\,: \::__/:|o--,,. . . . .\,: ¯¯: : : : : :"-,
01:06:22 <staplegun> \: : : : : : : : : : :""~,,... ... ... ... ... ... ... |... ... .\: : : : : :,,-"~": : : "'~~-,:"'~-,,_|: : :,-"¯¯¯¯"-,:|
01:06:24 <staplegun> .'~,: : : : : : : : : : : : :"'~-,,_... ... ... ... ...|... ... .."-,_: ,-" : : : : : : : : : : : "~,--~": : (o~--,,_. . |:|
01:06:25 <staplegun> ... ."'~-,,: : : : : : : : : : : : :"-,~--,... ... ... |... ... ... ...,/: : : : : : ,,--,-,~-,,: : :"~-,,: : :"~,___:"-/,/__
01:06:27 <staplegun> ... ... ...,"~"~--,,_,,-~"`"`": : "-,::"'~~--,,_..| ... ... ... ..|: : : ,: :,-". ,-"./. . . ."-,,_: : :"~-~": : : :"'-,,~,:"-,,
01:06:29 <staplegun> ... ... .,/: : : : --,,:|: : : : : : : : :"-,,::::::::::::"~--,,_... ... \: : :|: /. .,/. . |. . . ,-".|. "~,,_____,,,,,__:\: : : : |
01:06:32 <staplegun> ... ... .|: : : : : : : "|: : : : : : :,: : : :\:::::::::::::::::::::"'~~-,,"-,:|: |. /. . . .|. .,-". . |. ,-"'. \. . .,/|. . . .",:\: : ,/
01:06:35 <staplegun> ... ... ..\, : : : : : ,/ : : : : : : :|: : : ,/::::::::::::::::::::::::::::::::::"\: "-,,___,\,/___,,\/___. |,-". |. . . ,/.|: |,~"
01:06:38 <staplegun> ... ... ... ",-,,,__/: : : : : : :,/,_~"-:::::::::::::::::::::::::::::::::::::"-,,____,, : : : : : : : : : "'~,,,/.,,~". ,/: /
01:06:41 <staplegun> ... ... ... /: : : : :"-,,___,,-": : ,"-,-:::::::::::::::::::::::::::::::::::::::::::'\: : : : : : : : : : : : : : : "-"__,-":,-"
01:06:44 <staplegun> ... ... ... '\,_: : : : : : /-,,___,,"~-----~~~~,~---,,__::::::::::::::::::::::|: : : : : : : : __,,--~"~,,___,,-"
01:06:47 <staplegun> ... ... ... ... ."'~---~"... ... ... ... ... ... ...|---~"::::"'~-::::::::::::::::,/: : : : : : : : : :\"~,,
01:06:49 <staplegun> ... ... ... ... ... ... ... ... ... ... ... ... ... .|::::::::::::::::::::::::::::,,": : : : : : : : : : : |:::::"-,,
01:06:52 <staplegun> ... ... ... ... ... ... ... ... ... ... ... ... ... .|:::::::::::::::::::::::,,-": : : : : : : : : : : : :,|:::::::::"-,
01:06:55 <staplegun> ... ... ... ... ... ... ... ... ... ... ... ... ... .|::::::::::_,,,--~~": : : : : : : : : : : : : : :,/::::::::::::::\,
01:06:58 <staplegun> ... ... ... ... ... ... ... ... ... ... ... ... ... .|----~~" : : : : : : : : : : : : : : : : : : : :,/::::::::::::::::::\,
01:07:01 <staplegun> ... ... ... ... ... ... ... ... ... ... ... ... ... .|: : : : : : : : : : : : : : : : : : : : : : :,-,"::::::::::::::::::::::\
01:07:04 <staplegun> ... ... ... ... ... ... ... ... ... ... ... ... ... |: : : : : : : : : : : : : : : : : : : _,,~"SL'\,::::::::::::::::::::::\
01:07:07 <staplegun> ... ... ... ... ... ... ... ... ... ... ... ... ... |: : : : : : : : : : : : : : : : :,,-"... ... ... .\:::::::::::::::::::::::\
01:07:10 <staplegun> ... ... ... ... ... ... ... ... ... ... ... ... ... |: : : : : : : : : : : :_,,--~"... ... ... ... ...\:::::::::::::::::::::::\
01:09:29 -!- Sgeo_ has quit (Read error: 110 (Connection timed out)).
01:11:59 <staplegun> yes .. god forbit i disrupt your indepth conversation with less than a few kilobytes of data
01:17:10 -!- staplegun has quit ("ChatZilla 0.9.78.1 [Firefox 2.0.0.8/2007100816]").
02:25:36 -!- staplegun has joined.
03:37:06 -!- Sgeo has joined.
05:09:54 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
05:23:04 -!- calamari has joined.
05:24:19 <calamari> probably old news, but saw this on slashdot: http://www.wolframscience.com/prizes/tm23/solution_news.html
05:25:59 <calamari> "Alex Smith is an undergraduate studying Electronic and Computer Engineering at the University of Birmingham, UK. He has a background in mathematics and esoteric programming languages."
05:26:45 <calamari> dunno.. didn't find anything on the wiki with his name, but that doesn't mean much
05:30:16 -!- oklopol has joined.
05:30:23 <pikhq> Would the real Alex Smith please stand up?
05:31:48 <oklopol> would('ve) be(en) so cool if he were from here
05:33:33 <oklopol> also, i wish i'd joined 15 seconds later
05:34:08 <oklopol> even cooler than these two put together would be if i actually were alex, unfortunately i'm not
06:16:03 -!- calamari has quit ("Leaving").
06:39:54 -!- oklopol has quit (Read error: 110 (Connection timed out)).
07:02:54 -!- staplegun has quit ("ChatZilla 0.9.78.1 [Firefox 2.0.0.8/2007100816]").
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
09:12:31 -!- oerjan has joined.
11:05:39 -!- staplegun has joined.
11:11:32 -!- oerjan has quit ("leaving").
11:16:40 -!- bsmntbom1dood has joined.
11:27:49 -!- bsmntbombdood has quit (Read error: 110 (Connection timed out)).
11:29:01 -!- staplegun has quit ("ChatZilla 0.9.78.1 [Firefox 2.0.0.8/2007100816]").
12:21:41 -!- zuzu_ has quit (Remote closed the connection).
12:21:44 -!- zuzu has joined.
12:47:19 -!- oklopol has joined.
14:00:21 -!- RedDak has joined.
14:34:39 -!- INTERCAL has joined.
14:35:00 <INTERCAL> foreach (user in channel) do print("Hi!");
14:49:32 <INTERCAL> This program seems bugged. a = 3 + 5; print(a);
14:53:37 -!- INTERCAL has quit.
15:03:20 -!- Sgeo has joined.
15:42:35 -!- ehird` has joined.
15:47:08 -!- ais523 has joined.
15:50:12 <ais523> pikhq: you were right, the person who solved the 2,3 Turing machine problem is on #esoteric
15:50:30 <ais523> (I'd like to thank whoever it was who put the original prize in the topic and let me know about it in the first place)
15:51:11 * ais523 was logreading and saw people discussing their discovery
15:51:26 <ehird`> wait, was it solved? that wolfram turing machine thing?
15:51:36 * ais523 also realises their gender is now public knowledge, but continues stubbornly using gender-neutral pronouns anyway
15:51:46 <ais523> ehird': see http://wolframprize.org
15:51:50 <ehird`> the person was in #esoteric???????
15:52:23 <ais523> pity I'm in the wrong timezone to have been in either of the conversations about it
15:52:29 <ehird`> I don't know what to say it's amazing
15:52:56 <ais523> if you want proof, just deduce my email address from my IRC username and the facts given in the article, and I'll send you a reply (which the University servers auto-stamp with my real name)
15:54:33 <ehird`> did it involve writing a mapping to brainfuck? ;)
15:55:14 <ais523> no, cyclic tag systems
15:55:40 * ais523 has a client running in an X terminal window on Windows which tends to hang for no apparent reason every now and then
15:56:12 <ehird`> hehe, perhaps a bit simpler ;)
15:57:33 <ehird`> (you're crazy too, btw ;))
15:58:03 <ais523> I'm in #esoteric, so that was so obvious you didn't need to point it out
16:00:36 <ais523> it's a strange feeling, sitting at a computer trying to stop an INTERCAL compiler while you know at any second a journalist might phone and ask questions ranging from informed to inane
16:01:19 <ehird`> so, who's going to write a brainfuck interpreter in it?
16:01:41 <ais523> it runs at a speed worse than two to the two to the number of steps, so I suspect nobody any time soon
16:01:54 <ais523> (at least with the initial condition I used)
16:02:03 <ehird`> YEAH WELL THEORETICAL-LAND HATES YOU TOO :<
16:02:53 <ais523> you'll just have to find a more efficient way to do it. (I suspect there probably is one.)
16:03:23 <ehird`> is there any not-trivial program in cyclic tag systems written?
16:03:30 <ehird`> well, not-trivial = dividing two numbers
16:04:53 <ais523> ...and it seems that it's lament that I have to thank for unwittingly letting me know about the problem in the first place by putting it in the topic
16:05:11 <ehird`> so i guess he gets half the prize money? ;) hahaha
16:08:35 -!- oklopol has quit (Read error: 110 (Connection timed out)).
16:16:26 -!- RedDak has quit (Remote closed the connection).
16:22:35 <ehird`> ais523: you are all over google
16:23:09 <ais523> you have to be careful with the search terms; it's easy to phrase it in such a way that you get people with the same name as me who are not me
16:23:18 <ais523> but there are still quite a lot of results even then
16:23:24 <ehird`> i searched for: wolfram 2 3
16:23:30 <ais523> s/then/if you exclude false positives/
16:23:32 <ehird`> tons on tons of results, near-all have your name
16:23:40 <ais523> ehird`: that's what I would have recommended
16:27:17 <ais523> based on my own Googling attempts
16:28:23 <ehird`> "I just won $25k from Wolfram! Time to see how I'm doing on Google."
16:30:25 -!- sebbu has joined.
16:33:57 <ais523> every time I've looked back at my IRC terminal I see more comments I can't think of a good response to
16:34:31 <ehird`> ais523: the trials in the life of a computer scientist
16:35:34 <ais523> bsmntbom1dood: my current working version of the C-INTERCAL compiler contains a compiler for a language designed specifically for writing INTERCAL optimizer idioms
16:35:52 <ais523> the compiler is written in yacc, which means in this case yacc is a compiler compiler compiler
16:36:08 <ehird`> ais523: i hope you tell this to the journalists, they need to know what an exciting life you have
16:36:46 <ais523> surprisingly most of them weren't very interested
16:37:06 <ais523> but there are a few esolang references dropped into some of the articles if you look closely enough
16:38:18 <ais523> Wolfram's related blog entry, for instance
16:45:39 <ehird`> "Hey guys good news! I found a polynomial-time algorithm running on this 2,3 Turing machine for an NP-complete problem!" -- person I know
16:46:09 <ehird`> (It's probably funnier first time due to the fact it's on a rather idle post tagged on after "Edit: ")
16:47:10 <ais523> I'd be quite impressed if someone came across an NP-complete problem that could manage that
16:49:28 -!- sebbu2 has joined.
16:50:50 <ehird`> "Only God and Chuck Norris can solve the halting problem.
16:50:50 <ehird`> Edit: Also Rudy Giuliani. It turns out that for any input, the answer is 9/11. :/"
16:52:06 <ais523> the neat thing about the Turing machine in question is unlike with most languages, a program that lasts an infinite amount of time has to be infinitely long
16:52:25 <ehird`> wait... doesn't that make it NOT universal?
16:52:32 <ais523> it would be nice if there was a proof which filled space with a simple repeating pattern rather than the complicated nested pattern I used
16:52:33 <ehird`> don't you have to be able to do finite infinite loops?
16:53:01 <ais523> the point is that the initial condition is always derived from a finite original program using a non-Turing-complete-in-itself process
16:54:20 <ais523> if you think about it, all Turing machine initial conditions are infinitely long because the tape itself is infinitely long and you have to specify every element on it
16:56:33 -!- sebbu has quit (Connection timed out).
16:56:33 -!- sebbu2 has changed nick to sebbu.
17:02:27 -!- ais523 has quit.
17:29:14 <lament> setting up the initial conditions themselves can be represented as a program
17:29:19 <lament> and the program will be finite
17:29:42 <lament> the only infinite thing is that the tape starts off as infinite zeros
17:30:00 <ehird`> but hey, you directly caused him to find out about it and then win it
17:36:12 -!- oklopol has joined.
17:46:08 -!- Nucleo has joined.
17:52:48 <ehird`> haha, i haven't read it yet
17:53:04 <lament> i mean your conversation with ais523
17:53:37 <lament> if you want your hard esoteric problem solved - show it to #esoteric people!
17:53:57 <ehird`> LMAO, from earlier today: <pikhq>Oooh. <pikhq>Anyone we know?
17:54:14 <lament> of course, #esoteric avoided all publicity
17:54:27 <lament> maybe we should ask ais523 to mention it somewhere :)
17:54:36 <ehird`> <oklopol>would('ve) be(en) so cool if he were from here
17:55:00 <ehird`> lament: imagine how bad reddit was -- and they're more intelligent than most people, on average -- and multiply it times a few million
17:55:04 <ehird`> lament: do you still want to do that?
17:55:28 <lament> it's not _that_ important a result
17:55:34 <lament> it's probably just the people who read wolfram's blog
17:55:42 <ehird`> he's had newspaper articles, iirc
17:55:43 <lament> which are hopefully much smarter than redditers
17:55:47 <ehird`> (not published yet ofc)
17:55:59 <lament> newspaper articles aren't gonna mention an irc channel, because they don't know what that is
17:56:16 <lament> maybe we could post to lambda the ultimate if he's fine with that
17:56:21 <ehird`> it's really funny, this line:
17:56:24 <ehird`> <ais523>pikhq: you were right, the person who solved the 2,3 Turing machine problem is on #esoteric
17:56:31 <ehird`> i didn't catch on that he meant him until a bit later
17:57:21 <lament> also the redditers thing wasn't that bad in the long run
17:57:32 <lament> the idiots left within a day
17:57:34 <ehird`> it's redditors, just fyi
17:57:54 <ehird`> bsmntbom1dood: ais523 solved it
17:58:28 -!- lament has set topic: Esoteric programming language discussion | FORUM AND WIKI: esolangs.org | CHANNEL LOGS: http://ircbrowse.com/cdates.html?channel=esoteric | IRP in #irp | Don't spam the channel with EgoBot commands, /query EgoBot | Don't spam the channel with toBogE commands, /join #toboge | Don't spam the channel with bsmnt_bot commands, take him to your own channel. | Congratulations ais523 for winning the Wolfram research prize!.
17:58:33 <ehird`> no, it's evidently possible
17:58:52 <ehird`> anyway: <ais523>if you want proof, just deduce my email address from my IRC username and the facts given in the article, and I'll send you a reply (which the University servers auto-stamp with my real name)
18:00:39 <ehird`> we all know you are ais523 ;)
18:03:10 -!- jix has joined.
18:06:14 <bsmntbom1dood> someone should post something about #esoteric and the prize
18:11:59 <pikhq> That's pretty spiffy.
18:45:37 <lament> i'm sitting on a piano bench!
18:47:54 <oklopol> how many pianists do we have here?
18:48:14 <lament> i'm not a pianist, i'm a musician.
18:48:26 <oklopol> well yeah, i'm more of a general musician myself
18:49:33 <pikhq> Well, there's Gregor. . .
18:50:43 <oklopol> that would be a small percentage
19:00:07 <pikhq> ... Is everyone in here a musician? XD
19:01:04 <lament> you have a 3-day grace period to learn to play some instrument
19:01:16 <ehird`> I can play the kazoo...
19:04:13 <pikhq> It doesn't even have to be a complex instrument. . .
19:04:27 <ehird`> i can make random whines on the theremin?
19:04:34 <pikhq> Stallman, for example, sings (nasally) and plays the recorder (can't comment on quality)
19:04:48 <ehird`> stallman is a crazy hobo though.
19:04:50 <pikhq> Learn the theremin. . . We could use a theremin for the Esoteric Ensemble. :p
19:04:53 <ehird`> I'm not. I don't think.
19:05:04 <ehird`> i actually have a theremin
19:05:12 <ehird`> i thought it would be awesome until i realised i suck at anything musical
19:05:32 <lament> i think theremins kinda suck, but that's really cool if you actually have one
19:07:19 <ehird`> http://uk.youtube.com/watch?v=mW0B1sipLBI the formal proof that theremins are awesome
19:08:39 * SimonRC just thought of a great math trick
19:08:58 <SimonRC> I just realised how to measure the area of an arbitrary polygon
19:09:13 <SimonRC> you pick some point way outside the polygon
19:09:33 <SimonRC> then, procede round the polygon's edges
19:09:50 <SimonRC> outside the convex hull, or whatever it's called
19:10:45 <SimonRC> consider the area defined by your point and each edge, which is a triangle...
19:11:32 <SimonRC> if you calculate the ares of all the triangles, then add all the ones from edges that went "left" and subtract all the ones from edges that went "right", you have the area of the polygon!
19:12:16 <SimonRC> the point you are measuring from can be a corner of the bounding box, for example
19:12:36 <SimonRC> and the are of the triangles is given by the cosine rule, I think
19:12:55 <SimonRC> in fact, you may be able to use any point outside the polygon, but I can't be sure
19:14:13 <SimonRC> maybe even points inside the polygon work...
19:16:12 <SimonRC> oh, wait, the area of the triangles can be found without the cosine rule
19:18:52 <ehird`> who wants a random .signature-C challenge?
19:19:27 <lament> is the prize a theremin?
19:19:41 <ehird`> probably not, i like my theremin :P
19:20:43 <ehird`> but the challenge is: generate the sierpinski triangle (in 4 lines or less of C, no standard compliance required - gcc hacks allowed), and print it out in ascii (any size, really... the bigger the better, because you can get more detail. I'd say something like ~200x200)
19:20:49 <ehird`> you must do it with this method:
19:21:01 <ehird`> 1. Take 3 points in a plane, and form a triangle 2. Randomly select any point inside the triangle and move half the distance from that point to any of the 3 vertex points. 3. Plot the current position. Repeat from step 2.
19:21:12 <ehird`> you'll need a lot of iterations of course
19:23:22 <SimonRC> I am getting a funny result here...
19:23:50 <SimonRC> If you ahve a triangle inside a bouunding rectangle...
19:24:05 <SimonRC> suppose that it has one corner at 0,0
19:24:38 <SimonRC> the other corners are at (w,x) and (h,y), so the bounding box has size w wide by h high...
19:24:58 <SimonRC> the are of the triangle seems to be (wh-xy)/2
19:26:24 <SimonRC> (I got this by assuming that w=h=1 initially, then rescaling
19:27:06 <lament> calling a y-coordinate "x" is possibly the most confusing thing ever.
19:27:25 <lament> i suppose you actually mean the coordinates are (w,y) and (x,h)?
19:27:51 <SimonRC> yeah, I meant the corners are (w, y) and (x, h), I think....
19:28:32 -!- oerjan has joined.
19:36:28 <SimonRC> oops, I now realise that there are some triangles that won't work for
19:36:49 <jix> anyone knows alex smith?
19:37:08 <jix> alex smith
19:37:15 <jix> http://blog.wolfram.com/2007/10/the_prize_is_won_the_simplest.html
19:37:34 <jix> But at 20:53:59 GMT on Saturday, June 30--just 47 days after we announced the prize--we received a submission, with the description of the submitter given as "Alex Smith is an undergraduate studying Electronic and Computer Engineering at the University of Birmingham, UK. He has a background in mathematics and esoteric programming languages."
19:39:16 <oerjan> jix: read the logs for today :D
19:39:44 <jix> ah cool :)
19:44:49 -!- Arrogant has joined.
19:45:07 -!- RedDak has joined.
19:48:00 <jix> today is a great day :)
19:48:19 * GregorR stabs jix and steals his happy.
19:49:12 <jix> my doom cart arrived in UK (and is on the way to me), we got 1000eur for our robot team, and i think it's really cool that someone from here solved that problem
19:50:03 <Arrogant> I wish I had a robot that could play Doom
19:50:27 <jix> nah the doom cart and the robot stuff is completely unrelated
19:51:36 <SimonRC> duh, I could just applythe method I just rediscovered in the difficult case
19:59:05 * pikhq always feels smart here
19:59:09 <oklopol> well, anywhere else too, except for programming..
20:01:35 <oerjan> this is extremely annoying. now i'll have to solve the P vs. NP problem just to stay in the pecking order here...
20:02:14 <oklopol> what if i do that tonight just to annoy ya?
20:02:16 <jix> is there a prize for the collatz problem?
20:04:22 <oerjan> hm, nothing listed on the wikipedia page
20:04:38 <oklopol> http://mathworld.wolfram.com/CollatzProblem.html
20:06:59 -!- jix has quit (Nick collision from services.).
20:07:09 -!- jix has joined.
20:08:34 * oerjan wonders if David Morgan-Mar has ever been on the channel. But i guess his esoteric language days were in the past.
20:09:24 <SimonRC> I know Edwin Brady, and he has been here.
20:09:28 <oerjan> the author of Irregular Webcomic, and also Chef
20:09:47 <SimonRC> or should that be: Edwin " \t\n" Brady
20:12:33 * ehird` is parsing the whole of the brown corpus into his markov chain
20:12:37 <ehird`> and now it's gone again
20:13:20 <ehird`> something must be awfully odd at
20:13:49 <ehird`> lines 32000 to lines 33000
20:13:53 <ehird`> because it is lagging on them
20:14:32 <oerjan> ehird`: did my ping automatically move you into a message window? o_O
20:14:49 <ehird`> (i would not use it if i was not on windows beyond my control right now)
20:14:56 <SimonRC> if the point on the triangle are (0,0), (x1,y2), and (x2, y2), the area is (x1*y2 - x2*y1)/2
20:15:05 <SimonRC> for any arrangement of points
20:16:33 <jix> i know someone who knows someone who knows mandelbrot
20:17:32 <oerjan> but is that knowledge self-similar?
20:17:41 <ehird`> SimonRC: not '', but ..
20:17:45 <ehird`> (windows retarded keymap, sorry)
20:18:16 <SimonRC> "LATIN SMALL LETTER O WITH DOUBLE ACCUTE"
20:18:28 <ehird`> it's Godel with an umlaut
20:18:37 <jix> and by the first know i'm talking about my previous math teacher and a good friend of mine
20:18:39 * ehird` is watching his markov chain-erator go
20:18:45 <SimonRC> no, it's a double-acute, not a double-dot
20:18:55 <ehird`> SimonRC: that is not how godel is spelled
20:18:59 <oerjan> not double acute, that would be Erdos
20:19:59 <oerjan> godel-numbers are something _quite_ different from erdos-numbers
20:20:19 <ehird`> my godel-number is somewhere in the reigon of 34872398723423!
20:20:24 <ehird`> so is my erdos-number, incidentally!
20:20:48 <oerjan> ehird`: you've never written a collaboratory paper?
20:21:01 <jix> the 2nd person in that row is heinz-otto peitgen ... the doktorvater(??) of my previous math teacher
20:21:19 <oerjan> i don't quite recall my erdos-number but i think it was around 5-6
20:21:31 <ehird`> yow, i have something close to exponential time increase per line
20:21:43 <ehird`> because the table gets bigger every time
20:21:48 <ehird`> there's a bigger table to check in
20:22:03 <jix> and he runs the company my previous math teacher works at now
20:22:36 <SimonRC> ooh: http://mathcentral.uregina.ca/QQ/database/QQ.09.98/tyler1.html
20:22:40 <ehird`> why is the brown corpus so damn big? ;)
20:25:13 <SimonRC> ehird`: what is this thing?
20:25:37 <ehird`> i'm feeding the brown corpus into my markov chain generator
20:25:47 <ehird`> the brown corpus is, basically, a shitload of american english text
20:26:03 <ehird`> http://en.wikipedia.org/wiki/Brown_Corpus
20:26:35 <ehird`> i just found a plaintext version
20:26:38 <ehird`> and am feeding it through
20:27:41 -!- ehird` has quit (Read error: 104 (Connection reset by peer)).
20:28:53 -!- ehird` has joined.
20:34:08 * ehird` just does the first 5000 for now :P
20:37:41 <ehird`> damn, my program is slow
20:52:46 <ehird`> i'll just try 500 lines
20:55:38 <ehird`> "I'm willing to stake my political career goes back to his election to city council in 1923."
20:55:42 <ehird`> the first thing it produced
21:03:50 <ehird`> my markov chain produces intelligable stuff!
21:06:08 <ehird`> some text it's produced (i think the brown corpus starts with political stuff, so that'd explain these):
21:06:08 <ehird`> "Sam Caldwell, State Highway Department public relations director, resigned Tuesday to work for Lt. Gov. Garland Byrd's campaign."
21:06:19 <ehird`> "I didn't smell a drop of liquor, and we didn't have a son, William Berry Jr., and a doctor, medical intern or extern be employed at the State Welfare Department's handling of federal funds granted for child welfare services in foster homes."
21:06:19 <Arrogant> it'll probably spit out a 3-page-long improvised speech given at a party with no provocation
21:06:51 <Arrogant> Okay that last one doesn't quite work
21:07:12 <oklopol> neither means anything, really
21:07:39 <oklopol> unless you can "stake" an even
21:07:54 <ehird`> "I'm willing to stake my political career goes back to his election to city council in 1923." makes perfect sense
21:08:12 <ehird`> it effectively means "I'm willing to bet that my political career goes back to his election to city council in 1923."
21:08:19 <ehird`> anyway, oklopol, your bot was worse
21:08:21 <oklopol> so you can "stake" an event
21:08:47 <ehird`> also, mine is a second-order markov chain
21:09:07 <ehird`> yours maps word=>(next,next...), correct?
21:09:25 <ehird`> second order chains map (word1,word2)=>(next,next...)
21:09:29 <ehird`> "hello world abc" would be
21:09:55 <ehird`> it makes it a lot more intelligable, and doesn't increase the parroting much
21:10:04 <oklopol> yeah, plagiarizing a bit more will produce a bit better text, but that's no real enhancement
21:10:15 <oklopol> there's nothing fundamentally better there
21:10:17 <ehird`> it is a dramatic improvement
21:10:22 <ehird`> everything >2 parrots far too much
21:10:28 <ehird`> everything <2 (well, 1) produces garbage
21:10:31 <ehird`> but 2 produces semi-coherent text
21:11:29 <Arrogant> what's the use of making these chains, anyway?
21:11:39 <oklopol> doubt a "first-order" chain would produce any less coherent stuff given that input.
21:11:43 <ehird`> what's the use of using esoteric languages, Arrogant? :)
21:11:49 <oklopol> Arrogant: there's no use, ever
21:11:56 <ehird`> oklopol: prove me wrong
21:12:22 <ehird`> feed the first 499 (NOTE: not all of it) of http://dingo.sbs.arizona.edu/~hammond/ling696f-sp03/browncorpus.txt to your markov chain generator
21:12:28 <ehird`> make sure to do it line-by-line not all at once
21:12:35 <ehird`> then produce some text
21:12:59 <oklopol> you didn't time your program yesterday, so i'm not gonnado this ;)
21:13:30 <ehird`> but fine, you're wrong anyway (sheesh, what a childish reason to stop the proving that markov order-2 chains are better)
21:14:01 <oklopol> the real reason is i don't care
21:14:17 <oklopol> markov chains are too trivial to be interesting
21:14:26 <ehird`> so what do you suggest instead?
21:14:43 <oklopol> i suggest we drop the subject :D
21:15:01 <ehird`> no, really, what do you suggest instead of markov chains?
21:16:08 <oklopol> waiting for someone to invent something that works?
21:16:41 <ehird`> "markov chains are too trivial! i won't use them.
21:16:41 <ehird`> i'm just waiting for someone to invent something TOTALLY AWESOME"
21:17:04 <oklopol> err... what's funny about not thinking markov chains are interesting?
21:17:25 <ehird`> because you call them "trivial" yet state you are waiting for someone to invent something less trivial
21:17:33 <ehird`> if they're the least-trivial thing out there, then they're hardly trivial
21:17:55 <oklopol> i'm saying if you insist on getting something better, just wait; i've never claimed to be at all interested in making good parrots
21:18:40 <oklopol> and why the fuck would i need to think markov chains are interesting just because they're the best we've got?
21:19:02 <oklopol> if they're trivial, they're not interesting
21:19:18 <oklopol> since there's nothing interesting you can do with them
21:20:11 <ehird`> i was talking in the context of making parrots
21:20:20 <ehird`> i didn't say anything about you in particular being interested in them
21:21:02 <oklopol> i just said i don't wanna do the test, because i didn't see a reason to do that
21:21:16 <ehird`> meh, whatever, i'm going to ask in #lojban
21:21:25 <ehird`> about my generate-and-filter idea
21:31:19 <lament> what's so special about markov chains?
21:32:08 <lament> they're neither the most trivial predictive model, nor the most interesting one
21:32:19 <ehird`> what's the most trivial then?
21:32:24 -!- Arrogant has quit (Read error: 104 (Connection reset by peer)).
21:32:49 <lament> bag of words model is simpler
21:32:51 <oklopol> you can't think of anything simpler than a markov chain? :P
21:33:05 <ehird`> well, pick random word, repeat
21:33:09 <ehird`> but apart from that? no
21:33:10 <oklopol> random line noise is simpler
21:33:42 <ehird`> that's not a predictive model, though
21:33:52 <lament> ehird`: i actually study this at university at the moment
21:34:15 <lament> ehird`: the simplest model, given some data, would be to pick the most common word in the data and always produce that
21:34:47 <oerjan> i think it would be even simpler to pick the first word
21:35:02 <ehird`> if so, that's... trivial, but uh how shall i put this
21:35:11 <ehird`> what's something trivial that doesn't involve repeating yourself over and over?
21:35:36 <lament> pick a random word, preferrably with a probability distribution influenced by what you see in the data.
21:36:22 <ehird`> so, "hello world hello" -> "hello hello world world hello hello hello hello world hello world world"?
21:36:31 <ehird`> ok, the next step up in triviality from that
21:37:13 <lament> well, this is as far as you can get with words being considered independent
21:37:30 <lament> so the next step would be to consider them in relationship to one another
21:37:37 <ehird`> which brings you to markov chains?
21:37:54 <ehird`> what about one step up in triviality from markov chains (of all orders)
21:38:53 <lament> i don't know, is there anything more complex than a markov chain of infinite order?
21:39:10 <lament> i would think not, you'd just never get enough data to train it
21:39:36 <lament> that's not joking, you can't just combine markov chains of different orders like that :)
21:39:56 <oklopol> lament: you can make it better by also introducing the concept of "topics"
21:39:56 <ehird`> if you pretend markov chains don't exist
21:40:02 <lament> but i think the next step is just optimizations
21:40:09 <lament> like topics, teaching it the english grammar, etc
21:40:37 <lament> (if you know it's english and there're topics, which doesn't have to be the case)
21:41:05 <lament> ehird`: a markov chain is just a particular kind of a dependence graph
21:41:20 <oklopol> hmm... wonder if it could ever find any meaning for anything given just the irc interface
21:41:26 <lament> http://en.wikipedia.org/wiki/Bayesian_network
21:41:34 <oklopol> i mean, in theory, given a perfect ai
21:41:49 <oklopol> we already know that's possible with full world interaction
21:41:58 <oklopol> but irc just has ops, kicks and privmsg's
21:41:59 <lament> ehird`: read that article and consider how markov chains tie in :)
21:42:05 -!- ehird` has quit (Read error: 104 (Connection reset by peer)).
21:42:16 <oklopol> "nah, i don't feel like it"
21:42:34 -!- ehird` has joined.
21:42:41 <ehird`> i missed the last few messages
21:42:45 <lament> ehird`: http://en.wikipedia.org/wiki/Bayesian_network
21:43:03 <lament> that article is way too technical and messy and doesn't have enough pictures
21:43:31 <ehird`> anything else? maybe something that doesn't make my head hurt? :-)
21:43:46 <ehird`> (Maybe even something I could have a go at implementing, heh.)
21:44:25 <ehird`> certainly, something that doesn't say "im" ;)
21:45:45 <ehird`> well let's put it this way
21:45:52 <ehird`> i highly doubt i could implement a bayesian network
21:46:15 <lament> a markov chain is a bayesian network that looks like this
21:46:33 <ehird`> i use a 2nd-order markov chain
21:46:36 <ehird`> what does it look like then?
21:46:47 <ehird`> (more importantly what's another example of a bayesian network)
21:47:00 <lament> harder to draw, w1 now has an arrow to w3, etc
21:47:39 <lament> a->b just means "b depends on a"
21:47:41 <ehird`> ok so i'm not exactly sure of the implications
21:49:05 <lament> your knowledge of w2 is influenced by the knowledge of w1
21:49:20 <lament> and, since there're no more arrows in the 1st order graph, nothing else
21:49:36 <oklopol> lament: isn't that just what a 2nd order markov chain is?
21:49:39 <lament> so it's some kind of a probability distribution with one parameter
21:49:49 <lament> and for a 2nd order chain, two parameters
21:51:22 <lament> and for an infinite order chain, infinitely many parameters, so you can't really get any more complex
21:51:40 <ehird`> so bayesian network is basically a markov chain
21:51:45 <ehird`> is there anything really seperate?
21:51:58 <lament> wow, i really suck at explaining this.
21:52:48 <oklopol> well, blowjob is a blowjob
21:53:34 -!- bsmntbom1dood has changed nick to bsmntbombdood.
21:53:47 <oklopol> python has really killed my brain
21:54:34 <lament> bsmntbombdood: there's always the first tiem
21:54:40 <oklopol> doing C is hard when everything that requires manual memory allocation seems too complicated to be a good solution...
21:55:01 <ehird`> for simple programs you don't need malloc, oklopol
21:55:07 <ehird`> just choose some fixed size and make an array of it
21:55:56 <oklopol> but yeah, that's what i'm doing for the first version
21:56:14 <ehird`> nobody. has. done. my. sierpinski challenge
21:56:30 <oklopol> also, all the time i'm trying to use C++ stuff :P
21:56:35 <oklopol> then i realize i decided not to
22:00:13 <oklopol> bsmntbombdood: yeah, we need to do something about your bad sucking capabilities, lament's experience will be most helpful
22:02:14 <oklopol> well yeah, but it's actually quite hard getting the ...whatsit open
22:02:24 <oklopol> you might puke without training
22:03:27 <oklopol> i already have a small orgy planned for the weekend, actually
22:03:53 <lament> that's disgusting, young man
22:04:39 <oklopol> the least disgusting number of people
22:06:12 <oklopol> well, there's wankfest, sex, threeway, fourway... orgy
22:06:25 <oklopol> you might say there's a least weird one there
22:06:48 <oklopol> hehe, -fest always makes me laugh
22:06:51 <lament> you skipped "sanity", right after "fourway"
22:07:27 -!- RedDak has quit (Read error: 104 (Connection reset by peer)).
22:07:42 <oklopol> wankfest, sex, threeway, fourway, sanity... orgy
22:10:37 <oklopol> that should be the motto of our orgy
22:13:02 <oklopol> ARE YOU STEALING MY TIMES?
22:13:39 <oklopol> hmm, perhaps i should eat a pizza ->
22:15:12 -!- jix has quit ("CommandQ").
22:17:19 <ehird`> with knuth's literate programming system iirc
22:26:22 -!- SimonRC has quit (Remote closed the connection).
22:26:27 -!- SimonRC has joined.
22:31:44 <ehird`> i think the root of things like text generation
22:32:17 <ehird`> "produce something that looks like it could be in set S"
22:32:41 <ehird`> Maybe Sim(S,C) where C(a,b) is a function for ranking (from 0 to 1, say) how similar a and b are
22:32:56 <ehird`> of course it'll be very hard to figure out an algorithm for Sim that works pretty well, even more so for C
22:34:52 <ehird`> Unless, of course, there's some new development i'm unaware of?
22:41:56 <ehird`> Sim brute forcing isn't very nice, though
22:42:05 <ehird`> Especially if you're generating nonsense text or the like
22:42:36 <ehird`> (levenshtein distance wouldn't work very well, because it's the more "mechanical" kind of "simularity" - you wouldn't get "new" text like e.g. markov chains)
22:43:07 <bsmntbombdood> i want to know how google's translate agorithm works
22:43:24 <ehird`> you can also suggest a translation
22:43:37 <ehird`> (which is why a while back "sarkozy sarkozy sarkozy" translated to "Cheney defends Bush")
22:45:00 <ehird`> if anyone produces an algorithm (without hand-waving) for Sim and C, i would be mightily impressed (that is if it actually worked ;))
22:45:42 <ehird`> oklopol: google.com/translate_t it's the best machine translation out there basically [it still sucks]
22:45:48 <ehird`> bsmntbombdood: of course a perfect Sim and C would be AI
22:46:26 <oklopol> i kill cows -> Ich tten Khe
22:46:33 <oklopol> i need no further testing...
22:46:53 <ehird`> try translating a foreign language article to english, you can make out what they're saying most of the time
22:47:20 <oklopol> it's just wrong... form of the verb
22:47:57 <oklopol> also, oerjan, don't know everything!
22:48:33 <oklopol> yeah, but that such a simple thing... it's always ich + {verb}e
22:48:48 <ehird`> it's neural-net or similar-based, though
22:48:56 <ehird`> you don't just tell it something mechanical
22:49:04 <ehird`> you teach it over time, like neural nets
22:49:33 <ehird`> blame the whole world, for misteaching it
22:50:55 <oklopol> i apologize to anyone not a part of this world!
22:51:02 <oklopol> hmm, i'd better do some sleeping -
22:51:59 <oklopol> i've actually been having them again
22:52:08 <oklopol> like last night, i dremt i was in my friends room
22:52:21 -!- oerjan has quit ("Good night").
22:52:24 <oklopol> and i knew i only hard 50 seconds before it would burst into flames and i'd die
22:52:41 <oklopol> so i thought "i'll pee on the floor"
22:52:49 <oklopol> (no, didn't wet my bed ;))
22:53:02 <oklopol> that was so absurd, i lolled my ass off when i woke up
22:53:34 <oklopol> think, you're about to die, and the first thing that pops to mind is "i'll ruin my friends carpet, mwahaha"
22:53:53 <oklopol> okay, that may not have been a good example
22:54:12 <oklopol> it was gonna be like 5000 degrees
22:54:17 <oklopol> instant kill, and i knew that
22:54:42 <oklopol> i just wanted to pee on the floor, because no one could blame me for anything
22:55:07 <oklopol> was also interesting when i actually died, but that's hard to explain really
22:55:45 <oklopol> hmm, no i feel bad for wanting to sleep
22:56:01 <ehird`> my dreams are so naive
22:56:04 <bsmntbombdood> i was thinking you could hide flammable liquids in your bladder
22:56:05 <ehird`> dying just makes them go black
22:56:25 <oklopol> well, most people wake up when they die
22:56:35 <oklopol> i did too, but i woke up in another dream
22:56:45 <ehird`> i wake up but only after a second or so of black
22:56:50 <bsmntbombdood> the half-dreams where you fall and then when you hit the ground and jerk irl are cool
22:57:39 <bsmntbombdood> when you're half asleep, just before really falling asleep
22:57:43 <oklopol> yeah, but i've had dreams where i fall from a mountain top, and through concrete, and i can't breath
22:58:08 <ehird`> one of my fears is finding myself inside a solid object
22:58:19 <ehird`> but actually completely immersed in a solid object
22:58:25 <ehird`> like, say, an FPS with noclip on
22:58:33 <oklopol> one time i had a dream my hand was carved empty, and when i woke up, for a moment i actually hallucinated it really was
22:59:04 <oklopol> almost all these great dreams occurred about 2 years ago when i didn't sleep nearly at all
22:59:15 <oklopol> guess i should start sleeping less again
22:59:15 <ehird`> also, another of my fears is reality's texture mapping fucking up and me suddenly finding myself pasted onto the floor
23:00:04 <ehird`> like, the texture of my body, would be mapped onto the floor instead
23:00:09 <ehird`> you'd like, feel the floor
23:00:14 <ehird`> because you'd be the floor
23:00:22 <ehird`> and all objects would morph in size and have textures mapped on
23:01:09 <oklopol> hmm, i'll really go now, i already slept through one school day this week
23:21:05 -!- staplegun has joined.
23:22:16 -!- sebbu has quit ("@+").
23:22:37 <ehird`> anyone want to write that sierpinski program challenge?
23:29:38 <pikhq> Wait for Oerjan to return. . .
23:29:48 <pikhq> And then our Norwegian friend can help you. ;)
23:30:06 <bsmntbombdood> http://www.itavisen.no/sak/493810/-___Norge_f%E5r_OiNK-__sak/