00:00:04 well, i know no haskell or erlang. but i was just lumping them together to be silly anyway ;_) 00:00:21 fizzie, what about a top-speaker-per-month-list? 00:00:25 Well, Erlang and Prolog are kind of, sort of related 00:00:29 fizzie, with nick aggregation of course 00:00:34 There is another, what about TeX syntax? And then, there are also many others 00:00:40 Such as XML 00:00:41 what kind of smiley is ;_) 00:00:49 * Sgeo|FsckThatCli hits zzo38 with a backslash 00:00:50 Sgeo|FsckThatCli, only in appearance, not in behaviour. 00:00:57 IN ANY CASE, my list was not meant to be all-inclusive, just a few ideas to start you thinking 00:00:58 alise, one with a 90° rotation in the middle 00:01:04 i'm serious though. i'll look at haskell 00:01:21 quintopia: think of how you want to write some code 00:01:23 make it work 00:01:27 Haskell is a language that everyone should learn. FOr some weird reason, I really, really want to just TEACH it to someone 00:01:29 don't look at other people's shit :P 00:01:30 Anyone 00:01:39 Sgeo|FsckThatCli: i very much doubt you'd be a good teacher 00:01:43 i've already done that alise. and i hit some writer's block. hence, asking for ideas. 00:01:48 i also doubt you know haskell enough to anyway :P 00:01:58 Sgeo|FsckThatCliACTION OUCH 00:02:00 alise, the basics? 00:02:02 quintopia: implement it, write stuff that would use it, do whatever comes first to mind 00:02:05 zzo38: what. 00:02:11 alise: sounds lame 00:02:13 zzo38: did you really just use CTCP in the middle of a message and expect it to work 00:02:14 zzo38 typoed his /me 00:02:18 quintopia: yeah writing a language you like how weird 00:02:21 Sgeo|FsckThatCli: he doesn't have a /me 00:02:24 Exactly 00:02:25 it's technically valid what he did 00:02:29 He typoed manually doing a /me 00:02:30 it's just that... 00:02:32 alise: that's the kind of attitude that resulted in Perl 00:02:33 nothing supports that. 00:02:36 oh 00:02:36 quintopia: lol 00:02:40 quintopia: i never said make something inconsistent 00:02:44 Sgeo|FsckThatCli: No, I meant to type it that way 00:02:46 i just said write it how seems to be nicest when you run into it 00:02:48 then implement that 00:02:52 zzo38: no client supports that properly. 00:03:11 I can vouch for this client not supporting it 00:03:16 well, i know what would BE nicest to me in an abstract sense, just not what specifics match my abstract feelings. 00:03:27 OK. I just wanted to see what happened to other people when I did do like that, regardless of what client support it. I know I think it is not a proper command 00:04:23 quintopia: try meditating 00:04:30 zzo38: it's valid by the ctcp spec 00:04:33 but nothing implements it 00:04:55 * Sgeo|FsckThatCli should make the Factor library support ut 00:04:56 *it 00:05:05 alise: i can't walk the line between meditation and sleep >.> 00:05:09 MORE UNSOLICITATED COMMENTS ABOUT FACTOR! 00:05:11 quintopia: ur doin it rong 00:05:19 alise: i know 00:05:21 (note: i have never meditated) 00:05:22 teach me? 00:05:24 okay then 00:05:27 alise: O, it is valid? Then everyone's client is just broken in that case. 00:05:30 i meant in the more introspective intellectual sense 00:05:31 :P 00:05:42 zzo38: Or just not taking a crappy standard literally. 00:05:46 I've lied in bed eyes closed dead tired but fully awake 00:05:49 Sgeo|FsckThatCli[0001]ACTION OUCH[0001] ? 00:05:50 i'm practicing group meditation right now 00:05:53 zzo38, what was that about 00:05:56 also known as "crowd-sourcing" 00:06:18 quintopia: not going too well, is it:P 00:06:25 NOT WITH YOU NO 00:06:38 You should use PSOX syntax! 00:06:38 quintopia: About nothing 00:06:44 but at the very least i've helped distract people from the logs 00:06:47 that's worth something 00:06:51 zzo38: *Vorpal: 00:06:54 PSOX > * 00:07:00 And PSOX2 will be turing-complete 00:07:15 Vorpal: You can do "top speaker per month" pretty easily if you want everyone on the list, but "top N per month" is a lot messier, since you can't easily apply a per-month limit expression. (Nick aggregation is just a matter of terribly ugly conditional expression in place of the nick column.) 00:07:20 turing-completeness is boring 00:07:37 Not when it makes HQ9+ +PSOX2 be turing-complete 00:07:47 (Note: This was a real idea. It's vaguely scrapped) 00:07:47 fizzie, a sub query for limiting it per month maybe 00:08:05 what i want to see is more languages in the "computes all functions that can be guaranteed at the outset to halt" family 00:08:11 quintopia: impossible 00:08:15 fizzie, or maybe not 00:08:15 mathematically 00:08:23 aiise: wrong 00:08:24 functions can halt but have no proof of halting 00:08:30 i guess your definition sorta excludes it 00:08:30 but anyway 00:08:33 alise, and the language can't do thsoe 00:08:37 you're pretty much limited to primitive recursive functions 00:08:40 + a few extras 00:08:44 and such languages already exist 00:08:44 fizzie, something with select distinct perhaps? 00:08:46 e.g. Coq 00:08:49 yes i know such exist 00:08:52 datalog is one 00:08:59 but i don't think there are enough 00:09:03 Coq is more powerful by far, I think. 00:09:31 * Sgeo|FsckThatCli pokes alise with one of Factor's predicate classes 00:09:47 Although any similarity with dependent types is superficial 00:10:10 coq doesn't seem to be for application purposes...if you're talking about this automated prover thing 00:10:14 -!- FireFly has quit (Quit: swatted to death). 00:10:15 Vorpal: I tried to write something like that; it's... well, non-trivial. When you apply a subquery per each row, you can only return one value; and if you're joining subqueries, you don't get a listing of months. You can do it either by unioning manually N queries for N months, or possibly N "select Nth person" queries. 00:10:16 quintopia: it is 00:10:36 quintopia: a common method of using it is "write Coq code to implement some data structure, algorithm, etc.; prove it correct; extract the code (Coq can do this) to Haskell/O'Caml" 00:10:40 tada, formally-verified library 00:10:55 Well, that's interesting 00:11:01 Maybe I should learn about Coq 00:11:03 but there are many proof-only developments in it, yes, like Goedel's first incompleteness theorem and the four-colour theorem 00:11:18 An aggregate function that would concatenate all returned rows into one string would work, but doesn't (at least standardly) exist. 00:11:24 ouc 00:11:25 oic 00:11:27 I vaguely feel like I'm switching my love back to Haskell 00:12:04 The thing that drove me away from Haskell last time was things like zippers. It should not be a headache to implement things that are easy imperitively 00:12:22 PL/SQL and its ilk (PL/pgSQL) can of course do it easily, but that's no longer SQL. 00:12:46 What's the equiv for SQLite? 00:13:05 -!- augur has joined. 00:13:24 none 00:13:29 Sgeo|FsckThatCli: "Do it in the client app" is, I think, pretty much it. 00:13:37 fizzie: pah it's easy with LISPABASE 00:13:59 I think SQLite has triggers though>? 00:14:14 alise: In LISPABASE everthing's sunshine and rainbows and unicorns, right. 00:14:55 Someone should work out if it's TC 00:16:03 Sgeo|FsckThatCli: It has a bit nonstandard implementation of triggers; you could get top-N-per-month with those, assuming the triggers are allowed to write to other tables. It still won't be a single query, though. 00:16:21 Can triggers trigger triggers? 00:16:27 of course! 00:16:29 I think it can be possible to use TeX to make music synthesis, and then you have a program that reads the DVI file to play the music 00:16:41 as long as those triggers don't retrigger the original triggers! 00:16:45 fizzie: LISPABASE: the conventionalness of Prolog combined with the easy-understandibility of Lisp code! 00:16:58 Chrono Triggersl 00:17:04 Triggersl 00:17:04 quintopia, are you guessing? 00:17:36 I don't get it. is alise suggesting that Lisp isn't easy to understand? 00:17:53 Well, not for most people. 00:17:53 I think I can read Lisp better than Factor 00:18:15 Triggers only run on modifications, too; a "more pure" solution would be completely read-only w.r.t the db. 00:18:33 I'm not thinking about solutions to the problem 00:18:47 I'm wondering if programming in SQLite/Triggers would be interesting 00:21:14 . 00:22:11 Vorpal: I tried to write something like that; it's... well, non-trivial. When you apply a subquery per each row, you can only return one value; and if you're joining subqueries, you don't get a listing of months. You can do it either by unioning manually N queries for N months, or possibly N "select Nth person" queries. <-- hm.. or move the picking of top N names to an application program 00:22:16 or procedural SQL 00:22:54 Yes, that's the usual cop-out. 00:24:01 i has been pinged! 00:24:34 alise: In LISPABASE everthing's sunshine and rainbows and unicorns, right. <-- but unicorns are vicious bastards! 00:24:47 look at the classical unicorn 00:24:55 i love how easy it is to segfault gprolog 00:25:03 | ?- append(X,Y,[X,Y]). 00:25:03 Segmentation fault 00:25:34 ^style ct 00:25:34 Selected style: ct (Chrono Trigger game script) 00:25:42 fungot, chrono triggers 00:25:42 Vorpal: i must ponder this turn of events, it can have a powerful effect on time. ask the one to bring back lost loved ones... it's what that guy in medina, a village near the mystic mountain" 65,000,000 b. c.? yes, i'd have done something very brave! he's probably up north, to guardia!!! let's toast our land! now we'll have some peace! magus is a tad on the spooky side. our only hope. 00:25:51 fungot, and what about the sword 00:25:52 Vorpal: there! there it is! but by the time we're through with you, you'll be in danger. open hatch. 00:25:53 Yesterday, I win! 00:26:01 fungot, no... the sword I said 00:26:02 ./09.01.19:10:15:40 fungot, will I marry me? 00:26:02 Vorpal: the masamune! ride again! 00:26:02 ./09.01.19:10:15:41 KajirBot: unfortunately i guess you would have a lot to do with 00:26:02 alise: but, we are far outnumbered! dalton! leave her alone! 00:26:02 alise: these unique items make us invincible! could the heavens. burrowing deep into the the world's life energy! it's shut tight... great weather. groovin' on life... hey! did you people just come out of the closet? get out?! hey! is that for us! the chef's in a snit, trying to get food to the front lines. heard a spell to energize the sword takes immense evil! indeed! this thing. what you have? transform! this trading house. 00:26:29 fungot, the sword what can't it stop alone? 00:26:29 alise: i'd like to see that mystical sword for myself! geez! i'll be right there! 00:26:39 what 00:26:43 fizzie, bug in fungot 00:26:44 Vorpal: these unique items make us invincible! after them! 00:26:49 that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! 00:26:53 Vorpal: Not a new bug. 00:26:53 fungot, the sword what can't it stop alone? 00:26:53 alise: i'd like to see that mystical sword for myself! geez! i'll be right there 00:26:53 Vorpal: yes, it's been awhile prometheus! go!! this is for you guys are a lot! the hero has made the epoch, your wings! now this is a treasure there... found the treasure!! probably filled with mystical forest powers. we are immortal! we shall hold this position to the last man! big fire where lavos fall from sky! we no can call that the chrono trigger. it is r66-y? cool? who knows what would become of my mystics? i must win! 00:26:53 Vorpal: these unique items make us invincible! how can i do for you? thanks! it's about the missing queen. she still looks so much like leene, that they will take you to your place of execution?! strange, but!? 00:26:58 that is completely wrong 00:27:15 fizzie: did you make the bug fix for fungot's pseudorandom selections permanent? 00:27:15 oerjan: your majesty! and queen. he and a friend left on a journey 10. all functions are down... got the terra arm and the crisis arm! found a dreamstone! i'd forgotten how beautiful they are the evildoers? magus's lair! you brave! he's probably up north, to guardia!!! let's toast our land! now we'll have some peace! magus is a tad on the spooky side. our only hope. 00:27:21 Vorpal: no, i pinged fungot 00:27:21 alise: you, with you standing around! the trial! what has gotten away with this! a top secret document has been left behind the floor to fight the mother brain, yet! 00:27:23 before you 00:27:27 or? 00:27:29 oerjan: I believe I did. 00:27:34 alise, yes you did but count the number of times 00:27:47 alise, it replied one more than the number of times you pinged it 00:27:52 . 00:27:54 Anyway, the "answers to the wrong person" is a known thing. 00:27:58 and directly after I highlighted it 00:28:09 fungot, tell me about the damn sword now 00:28:09 Vorpal: we are looking, but well behaved! crono!! i owe you one! you saved the day of lavos"... go to " mystic mountain" 65,000,000 b. c.? yes, i'd have done something very brave! he's probably up north, to guardia!!! let's toast our land! now we'll have some peace! magus is a tad on the spooky side. our only hope. 00:28:12 fungot, sigh 00:28:12 Vorpal: we are looking, but well behaved! crono!! they're escaping! 1b is escaping! so! that cathedral to the west?... yes! well then rest and relax! huh? 00:28:19 fungot, ... 00:28:19 Vorpal: frog will do. well, i better! whoosh! i wonder how everyone! humans! they're my friends! you'll see, you'll be in danger. open hatch. 00:28:24 Vorpal: stop 00:28:24 the 00:28:25 spamming 00:28:26 of 00:28:26 fungot 00:28:26 alise: that no one was allowed to use the crane, enter any two of these letters, a b y. 00:28:27 fungot, fine, suit yourself :P 00:28:27 now 00:28:27 Vorpal: you! take! we find! 00:28:28 it 00:28:29 is 00:28:30 even 00:28:31 alise, indeed 00:28:32 more 00:28:34 than 00:28:36 my 00:28:38 HackEgo 00:28:40 spamming 00:28:43 i'd say hypocrite but that's wearing kinda thin today! 00:28:45 alise, stop the reciting of the Befunge-98 spec 00:28:56 alise, indeed it is 00:29:02 Sgeo|FsckThatCli: Vorpal did that word-per-line thing earlier today. 00:29:09 also, stop freaking out about that 00:29:33 I also take offense on "completely wrong"; it generated the text just fine, it just directed to the wrong person. That happens to the best of us! (Like cpressey; not that I'm implying anything about relative goodness levels here.) 00:29:34 Sgeo|FsckThatCli: Vorpal did that word-per-line thing earlier today. <-- hm, which one. or did you mean fizzie that did the graphs for that 00:29:42 Vorpal: i mean you talked 00:29:43 like 00:29:43 this 00:29:44 to 00:29:44 stop 00:29:44 me 00:29:45 using 00:29:46 a 00:29:48 bot 00:29:49 alise, ah 00:29:54 not graphs of words per line 00:29:57 W 00:29:58 e 00:29:59 l 00:30:00 l 00:30:00 00:30:02 I 00:30:03 00:30:04 c 00:30:05 a 00:30:06 n 00:30:08 no, that is worse 00:30:13 * Sgeo|FsckThatCli stops 00:30:36 Oh 00:30:37 man 00:30:42 char per line is only okay if 00:30:42 d 00:30:42 i 00:30:42 a 00:30:42 g 00:30:43 o 00:30:45 n 00:30:47 a 00:30:47 That's going to hurt your line-length distributions, you know. :p "Won't anyone think of the line-length distributions!" 00:30:47 i thought this channel was getting easy to read for a moment 00:30:49 l 00:30:51 (sorry for that one, couldn't resist) 00:31:08 * Sgeo|FsckThatCli wonders how long Befunge-98 would be 1 char/line 00:31:17 I mean, timewise 00:31:19 Sgeo|FsckThatCli, run wc -c on it 00:31:21 oh 00:31:22 well 00:31:24 calc it 00:31:28 wc -c * what? 00:31:32 what about convert the ascii to binary 00:31:34 then one bit per line 00:31:47 what about unary? 00:31:53 Vorpal: people here have _low_ self-discipline 00:31:53 (yes, had to go there) 00:32:02 oerjan, err yes, what about it 00:32:22 Vorpal: regarding your resist 00:32:30 vorpal: as a result, the chan moves faster 00:32:31 Might the unary be longer than what the universe can store? 00:32:33 oerjan, ah 00:32:37 quintopia: well "hi!" would be 6,842,657 lines long. 00:32:54 !haskell do [1..20]; "Haskell! " 00:32:55 alise, how did you arrive at that number 00:33:02 Vorpal: manually 00:33:04 "Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! Haskell! " 00:33:07 if everyone had high self-discipline, they'd be doing something productive rather than calculating what 24 bits is in unary digits 00:33:13 alise, I mean, what is the formula for calculating it 00:33:14 and thus the channel would be slow 00:33:24 Vorpal: look at the ascii in binary 00:33:27 interpret that number as unary 00:33:27 done 00:33:39 alise, I'm trying to remember what the unary encoding is 00:33:44 Umm 00:33:48 just 1 repeated N times 00:33:50 and now 00:33:50 prepare 00:33:53 for my work of art 00:33:54 which I title 00:33:55 My idea is to add PicoC into Enhanced CWEB 00:33:57 "Unary '.'" 00:33:59 1 00:33:59 1 00:33:59 1 00:33:59 1 00:34:01 1 00:34:02 macrobreak 00:34:03 1 00:34:03 oh no 00:34:05 1 00:34:07 1 00:34:08 .... 00:34:09 1 00:34:11 1 00:34:11 good we abort it 00:34:13 1 00:34:14 -!- ChanServ has set channel mode: +o oerjan. 00:34:15 1 00:34:17 1 00:34:19 1 00:34:19 go oerjan! 00:34:19 yes 00:34:21 kick time 00:34:21 1 00:34:21 ! 00:34:23 1 00:34:31 -!- alise has joined. 00:34:33 1 00:34:34 oerjan, strange kick reason 00:34:35 1 00:34:35 Ok, so: What did alise actually print? 00:34:37 1 00:34:37 and kickban him 00:34:39 1 00:34:41 1 00:34:42 for now 00:34:43 1 00:34:44 2 00:34:45 1 00:34:47 1 00:34:49 1 00:34:51 1 00:34:52 oerjan, !!!! 00:34:53 1 00:34:54 -!- oerjan has set channel mode: +b *!*alise@91.104.246.*. 00:35:02 Aww, don't worry, Bender. There's no such thing as 2 00:35:03 oerjan++ 00:35:04 it was going to be '.' obviously 00:35:18 quintopia: but at the point he was kicked, what's the result 00:35:39 something unprintable 00:35:47 what's ascii 27? 00:35:55 space is 32 00:35:55 -!- oerjan has set channel mode: -bc *!*alise@91.104.246.*. 00:35:55 -!- ChanServ has set channel mode: +c. 00:36:02 oerjan, why -c 00:36:19 er i have no idea why that -c happened 00:36:26 because 00:36:28 i just did /mode -b *!*alise@91.104.246.* #esoteric 00:36:42 what is c anyway 00:36:46 -!- alise has joined. 00:36:50 that was my client 00:36:53 i said nothing after the first kick 00:36:54 oerjan, #esoteric first ? 00:36:54 color stripping, I think 00:37:08 alise, maybe, use /flushq in xchat 00:37:11 alise: you said it all before. you pasted it all at once right? 00:37:12 i did, tyvm. 00:37:14 * Sgeo|FsckThatCli wants to watch stripping in color >.> 00:37:14 i didn't realise 00:37:37 Vorpal: maybe, i couldn't remember the order but thought it would just give an error if it was wrong 00:37:38 Sgeo|FsckThatCli: is there anything you can't turn into an inane sexual reference? 00:37:44 -!- oerjan has set channel mode: -o oerjan. 00:38:13 It's weird, there was a time that I'd never make a sexual reference 00:38:15 oerjan, unlikely. That would be far too easy to use 00:38:20 -!- augur has quit (Read error: Connection reset by peer). 00:38:23 Vorpal: :D 00:38:50 -!- augur has joined. 00:38:58 when i do sexual references, i like to use surgical redaction. 00:39:04 Sgeo|FsckThatCli: is there anything you can't turn into an inane sexual reference? <-- no, but I invoke rule 34 on inane sexual references 00:39:42 Vorpal: Once upon a time, "Your mom said that in bed last night" and "That's what SHE said!" did it. 00:39:43 The end. 00:39:50 A naked superintelligent shade of the color blue? 00:39:52 alise, hah 00:40:37 oerjan, strange kick reason <-- i didn't provide any 00:40:42 oerjan, ah 00:40:44 i was talking to a friend about netbooks, and the following innocuous line emerged: 00:40:47 12. "ah well, yes, if you insist on 10 inches and lots of hard di[redacted]k, it will be $300, but that's only a hundred dollars more!" - quintopia 00:40:55 xD 00:40:57 that's surgical redaction in action 00:41:37 haha 00:41:48 night → 00:41:56 baibai 00:41:58 * Sgeo|FsckThatCli watches Vorpal run into the userlist 00:42:22 Vorpal: hàihài. 00:42:31 jejejejejejeje 00:42:36 BOTTELISP WILL BE SO SUPRA 00:42:40 IT WILL BE SUPRAINDENTENDT 00:42:48 OF THE POLEESE FORCE OF LISP 00:42:52 True fact. 00:43:21 * Sgeo|FsckThatCli feeds alise a cursed fortune cookie 00:43:26 -!- augur has quit (Ping timeout: 245 seconds). 00:43:59 hey, since i'm still new and all, can i get some brief bios maybe? some introduction type thingies? 00:44:14 2010-09-17 21:30:52 (95.5 KB/s) - `logs.db' saved [248597504/248597504] 00:44:16 so happy i could query for a duck 00:44:34 quintopia: i can give you two bios for the price of one! 00:44:41 SWEET 00:44:43 what's the price of one? 00:44:46 but Vorpal's would basically be offensive, so 00:44:49 quintopia: nothing at all! 00:44:57 well, just tell me who you are then 00:45:23 hi, i'm 15 and despite the nick -- which originated as a joke and ... stuck -- i'm male. until recently i was in a mental institution 00:45:30 I guess that isn't the greatest bio ever. 00:45:48 also you're american, as you have previously indicated 00:45:51 that's really enoguh 00:45:51 err 00:45:52 wat 00:45:56 i'm british 00:45:58 Vorpal: oh from various messages in my windows it appears that irssi interpreted /mode -b *!*alise@91.104.246.* #esoteric as first an unban, and then _all_ of #esoteric as a list of flags 00:46:06 oh nvm then 00:46:15 oh right he left 00:46:19 ooh i remember when otpbot ran rule 110 in the topic 00:46:22 thems were fun times 00:46:29 * Sgeo|FsckThatCli plays flash Portal 00:46:34 *modes 00:46:52 i never did finish playing through that 00:47:13 got to the place where it was like WTF HOW IS IT EVEN POSSIBLE TO MOVE THROUGH THIS LEVEL and gave up 00:47:27 actually, i never finished playing through the main manufactoria levels either :/ 00:47:50 <- 40 male norway 00:47:55 the one where you have to shoot yourself over the glass? 00:48:06 otpbot = wrlang? 00:48:14 oerjan: ethnicity? 00:48:23 *erlang 00:48:43 coppro: no the one with like a bajillion rooms all with sliding doors that are seemingly only openable from the inside...level 45 or something like that 00:48:56 hey oerjan turned 40 00:48:59 HAPPY OLD OERJAN 00:49:14 oerjan: what do you do? 00:49:15 quintopia: oh yeah and i've been here since 2007/2008. 00:49:23 oh 00:49:26 thus completing the worst bio in history 00:50:05 * oerjan always _thought_ norwegian was an ethnicity 00:50:34 scandinavian? 00:51:37 oerjan: well, i hear there are Lapps in there too, rarely 00:51:40 oerjan #esoterics for a living 00:51:52 if you ever get bored, try and understand one of his papers 00:52:00 you'll instantly feel a strong urge to do something else 00:52:02 ohho, an academic 00:52:11 sounds hawt 00:52:15 xD 00:52:23 we're all academic, it's just that most of us are too crazy to get published! 00:54:44 yes but he's 40. it takes some...uh...stamina to be an academic that long 00:55:03 if there are lapps among my ancestors they're probably quite a bit back 00:56:04 i'm not sure he academises any more. 00:56:16 he's also not talking about it 00:56:17 so i won't either 00:56:52 -!- augur has joined. 00:57:53 -!- Sgeo|FsckPuppy has joined. 00:58:06 Ubuntu can save its stuff to USB too, right? 00:58:22 very much 00:58:36 Sgeo|FsckPuppy: automatically? 00:58:38 uhh, i'm not sure. 00:58:39 this shall be the new official comic of #esoteric: http://i.imgur.com/jacoj.jpg 00:58:41 yes 00:58:42 yes it can 00:58:43 i think 00:58:44 it saves ~ 00:58:47 if you write the USB right 00:58:50 (using ubuntu's application) 00:59:01 quintopia: oh come on we never actually program 00:59:08 even if we do we never have users 00:59:30 with that attitude, of course you have no users! 00:59:45 -!- Sgeo|FsckThatCli has quit (Ping timeout: 252 seconds). 00:59:49 how do you personally define "esoteric programming language"? 01:00:35 No good attitude could have gotten PSOX users 01:00:35 let me guess, you're going to go on about how our research can pave the way for new practical languages of wonderfulness! 01:00:37 THIS GUY ASKS TOO MANY QUESTIONS 01:00:38 of course it could 01:00:48 we just don't give a shit about things /that/ boring :) 01:01:16 quintopia: uselessness is _nearly_ part of the definition 01:02:35 okay, just wondering 01:02:48 * quintopia had no intention of going anywhere with that 01:03:16 Would Freenode complain if my nick was Fuck? 01:03:32 Sgeo|FuckPuppy Erroneous Nickname 01:03:35 but there are some practical languages that i, for one reason or another, consider esoteric. 01:03:54 -!- quintopia has changed nick to You. 01:03:55 quintopia: i recall BancStar was (at a time anyhow) disqualified from the wiki for actually being seriously intended 01:04:00 -!- You has changed nick to quintopia. 01:04:08 despite being utterly crazy, or so 01:04:19 There's someone I know on another IRC network called You 01:04:24 me too 01:04:27 or was that something else 01:04:30 Foonetic? 01:04:32 i just wanted to see if it were registered here 01:04:39 yeah, foonetic fuck yeah 01:04:41 -!- augur has quit (Ping timeout: 240 seconds). 01:04:56 that's my homebase network dog 01:05:30 -!- augur has joined. 01:07:15 -!- Wamanuz has quit (Remote host closed the connection). 01:08:08 quintopia: hm in fact it's still not on esolang, but wikipedia has it: http://en.wikipedia.org/wiki/BANCStar_programming_language 01:09:01 oerjan: i'm very familiar with it. i remember when it was on esolang. 01:09:10 also, i recall a handful of tdwtf articles 01:10:15 -!- augur has quit (Ping timeout: 252 seconds). 01:10:20 -!- oerjan has quit (Read error: Operation timed out). 01:12:47 things that suck more than usual: today's xkcd 01:14:47 Things that suck: xkcd 01:15:30 used to be good once upon a time 01:15:38 but today it has reached a new level of suckitude 01:15:59 before, when it sucked there'd still be a few fanboys that liked it and stuff 01:16:07 today, even the fans agree it sucks 01:16:29 yeah xkcd used to be wonderful 01:16:38 I've taken up not having an opinion 01:16:47 and from comic 400 to 500, it rapidly descended into utter shite 01:16:57 after showing inklings of imminent collapse beforehand 01:18:44 alise, I assume I'll be much happier on Ubuntu 01:18:51 Is this a safe assumption? 01:19:12 -!- Oranjer has joined. 01:19:13 Yes. 01:19:16 -!- Oranjer has left (?). 01:19:20 I told you just to use Ubuntu from the start. 01:19:22 -!- EOF has joined. 01:19:23 Even as a temporary OS 01:19:29 It should be fine from a USB stick. 01:19:30 alise, I still haven't tried Knoppix 01:19:33 yay 01:19:34 Don't. 01:19:35 It's crap. 01:19:41 knoppix is :/ 01:19:47 Sgeo|FsckPuppy: I'd burn an Ubuntu CD, then use that CD's USB creator. 01:19:52 This lets you set it to preserve ~ on shutdown. 01:19:54 no 01:20:00 Who is EOF? 01:20:09 ^D 01:20:11 Burning a CD is a bit of a tricky proposition right now 01:20:12 he is a more experienced linux user 01:20:30 * EOF > u 01:20:42 Anything I can do from Parted Magic or Puppy Linux? 01:20:52 Or, well, I guess I can save the .iso to USB 01:20:54 zomg 01:21:00 fail 01:21:07 ? 01:21:10 are you on windows? 01:21:27 Right now, I'm on Puppy Linux 01:21:39 usb? 01:21:47 Managed to burn Parted Magic from my ancient computer that I'd rather leave off 01:21:55 Burned Puppy Linux from Parted Magic 01:21:58 -!- augur has joined. 01:22:12 Screwed up a few disks trying to get the saving to DVD right, eventually gave up 01:22:35 Now I might have a USB drive available soon, so 01:22:39 That might help save my sanity 01:22:46 you need to burn the image to the disk, don't just copy the file 01:22:59 can you open up a terminal? 01:23:00 I _know_ that 01:23:14 But when downloading the thing, I need a place to put it 01:23:20 Ramdisk probably won't cut it 01:23:27 put it somewhere on your hard drive 01:23:38 I have no hard drive in this computer 01:23:44 /. 01:23:50 uhm 01:24:16 so, you are running for a usb or a cd? 01:24:22 CD right now 01:24:32 do you have a flash drive? 01:24:35 Puppy Linux puts itself in RAM, so I can burn more CDs 01:24:46 Not right now, but there might be one in this house, according to my dad 01:24:49 i haven't even gotten halfway through the latest subnormality and it's already obvious the punchline is going to involve the waitress being the spy. try harder dude. 01:25:01 quintopia: EEEEH wrong. 01:25:20 hmm 01:25:31 so the puppy image is COMPLETELY in ram? 01:25:43 Yes 01:25:51 EOF: Yes, it loads into RAM from CD and then unmounts the CD. 01:25:59 he got me 01:26:02 how much ram do you have 01:26:10 and how much are you using? 01:26:15 At least 128MB! 01:26:22 ... 01:26:37 Below that, Puppy won't load into RAM. 01:26:54 Puppy, BTW, is *fucking tiny*. 01:26:58 but how much does he *have* 01:27:14 At least 128MB. 01:27:27 STFU 01:27:37 i'm asking Sgeo|FsckPuppy 01:27:49 Þou firſt 01:27:57 I have 1gb 01:28:05 hmm 01:28:19 how big is your iso? 01:28:24 I don't like the \outer command in TeX 01:28:33 Not sure, I think Ubuntu is ~600mb? 01:28:34 hmm 01:28:39 ohh 01:30:23 I can use my old computer if I have to 01:30:27 Which does have a HD 01:30:46 Also, I'd need to find a blank CD 01:30:46 kk 01:31:02 I have a bunch of blank DVDs, but old comp doesn't have a DVD drive 01:31:04 -!- augur has quit (Ping timeout: 276 seconds). 01:31:07 This computer does 01:31:11 do you live in canada, if so i can mail an ubuntu install disk to you :) 01:31:38 * Sgeo|FsckPuppy is too impatient for that. And I'm somewhere vaguely south of you 01:31:46 lol 01:32:15 then just request a free one from canonical 01:32:30 * Sgeo|FsckPuppy reemphasizes impatient 01:32:37 I'll just use the old comp 01:32:46 Or, again, when I have the flash drive... 01:33:03 * Sgeo|FsckPuppy vaguely wonders if solid-state 1TB drives exist 01:33:10 yes 01:33:19 .... 01:33:21 * Sgeo|FsckPuppy wants 01:33:32 and they're getting cheaper than 2000 bucks 01:33:37 ... 01:34:31 i lied 01:34:35 http://www.newegg.com/Product/Product.aspx?Item=N82E16820227515 01:34:42 i bit over 2k 01:35:15 but it's on sale 01:35:23 350 bucks off 01:35:26 he is a more experienced linux user 01:35:26 * EOF > u 01:35:28 so who are you? 01:35:37 apart from irritating 01:35:41 :) 01:35:45 indeed 01:36:48 a bit more specification would be appreciated 01:36:49 i'm a canadian(eww a canadian, i know right :) ) with knowlege of unix-like systems and a passion for hardware and sexual superiority 01:37:01 you wouldn't happen to know Quadr*scence, would you? 01:37:39 only whenever #esoteric turns into #mention-my-sexual-awesomeness-at-every-single-turn, it's usually his fault. or someone he dragged along. 01:37:44 Quadrescence ? 01:37:50 great; now you've pinged him. 01:38:16 indeed 01:38:21 you mentioned him 01:38:38 omg 01:38:57 he's got no ip or host name :/ 01:39:18 it's called a cloak 01:39:19 he may well be a bot :) 01:39:29 you don;t know 01:39:36 so is there any actual reason you mentioned your passion for sexual superiority? 01:39:56 tell me if you think 01:40:09 this guy's pic is kinda cool 01:40:10 http://profile.ak.fbcdn.net/hprofile-ak-snc4/hs353.snc4/41669_790499184_93_n.jpg 01:40:23 i have no idea who you are or what you're trying to achieve, but shut the fuck up. 01:40:33 kay : 01:45:57 -!- HackEgo has quit (Ping timeout: 271 seconds). 01:46:04 -!- HackEgo has joined. 01:50:53 * EOF is somewhat knowledgeable about GNU/Linux 01:51:35 -!- Leonidas_ has joined. 01:51:44 How did you find out about this channel? 01:51:54 -!- coppro_ has joined. 01:51:57 -!- SimonRC has joined. 01:52:43 -!- SimonRC_ has quit (Write error: Broken pipe). 01:52:44 -!- coppro has quit (Write error: Broken pipe). 01:52:51 -!- Leonidas has quit (Remote host closed the connection). 01:52:53 i was talking in a channel about constructed languages, and somebody mentioned the esotericity of Esperanto 01:53:34 and i first tried esolang (common short for for esoteric language) 01:55:15 PicoC seems to be broken, a bit 01:55:29 "typedef" always causes a parse error 01:55:56 -!- augur has joined. 01:56:04 If you define something "void zzz(void)" it expects an argument and then says it can't set void 01:56:16 Statements like "0;" fail 01:56:50 hmm 01:57:05 zzz is the name of the function 01:57:55 and void is both an argument and the value(or lack therof) returned... 01:58:29 change void in the parentheses to voidarg 01:58:40 and replace calls to the arg 01:59:07 -!- augur_ has joined. 01:59:27 why did you mention this tho? 01:59:34 zzo38 ? 01:59:47 EOF: Because the GNU C compiler doesn't work like that 02:00:14 Changing "void" to "voidarg" says bad type declaration 02:01:24 or just call it with an argument 02:01:46 -!- augur has quit (Ping timeout: 276 seconds). 02:01:53 void zzz(viodarg) 02:02:07 **voidarg 02:02:15 methinks 02:02:15 That doesn't work 02:02:28 is your code open? 02:02:33 And calling it with an argument makes the error message that it cannot set void 02:02:39 -!- augur_ has quit (Read error: Connection reset by peer). 02:03:04 then don't use void as an arg 02:04:00 void zzz(){grab values from variables and constants in the main function 02:04:06 } 02:04:25 email me your code at hamiltonham9@hotmail.com 02:04:35 -!- SgeoN1 has joined. 02:04:58 I think Puppy is playing some kind of sick joke on me 02:05:10 yeah 02:05:14 it will do that 02:05:29 Unless Firefox with some 10.x old version of Flash really is that crashtastic 02:06:01 Geoffrey, knoppix or Ubuntu? 02:06:09 Erm eof 02:06:38 debian 02:07:02 or fedora if you want a user friendly distro with good repos 02:07:12 I'm not looking for a permanent installation 02:07:13 -!- Sgeo|FsckPuppy has quit (Ping timeout: 252 seconds). 02:07:30 fedora on a usb, 02:07:35 Just a LiveCD that's usable for now 02:07:40 hmm 02:07:54 why can't you do a perm install 02:08:04 dual booting is a blast 02:08:11 No hd 02:08:15 ... 02:08:49 iunno 02:08:58 EOF: I don't have email 02:09:03 ... 02:09:38 how do people with no hdd, and no email; make it into irc 02:09:58 Believe it or not, neither is necessary for TCP/IP access. 02:10:51 It's not my fault I don't have an HD! 02:10:55 Well, it is, kind of 02:12:44 Actually, I do have an HD, it's just removed from my computer. Shall I put it back in? 02:13:32 I need to find a good C interpreter which I can embed into CTANGLE. It needs to be small, and it must be a proper subset of C (so that any codes here can also be compiled with a C compiler), but it doesn't need to support comments, because CTANGLE strips out comments. 02:14:20 zzo38: http://typewith.me/TXT4kmXm2O < put your code in here so i can see what you were doing 02:14:23 Also, the #define command should be supported. 02:14:53 Was about to suggest a nop 02:15:09 Now it just needs to do simple term rewriting? 02:15:26 EOF: You do not even understand at all what I am doing. 02:15:33 nope 02:15:34 Do you know about Enhanced CWEB? 02:15:35 plz 02:15:39 explain 02:15:49 i'm half asleep 02:15:54 i can't infer 02:16:05 zzo38: TinyCC. 02:16:11 It's even a library. 02:16:26 Hooray, Fabrice Bellard. 02:16:31 yes 02:16:35 Enhanced CWEB is a program I have written, based on CWEB, used for literate programming. It can combine C codes and TeX codes. There are other features, as well. 02:16:47 it can interperet the linux kernel 02:17:24 EOF: Compile quickly. 02:17:29 As a bootloader. 02:17:51 It also needs to be able to compile partially and call those functions before compiling the rest of the program 02:18:07 zzo38: Yeah, TinyCC can compile a single function and get you a pointer to it. 02:18:23 It's really an amazing piece of work. 02:18:35 http://bellard.org/tcc/tccboot.html 02:18:44 mmm 02:18:46 saucy 02:18:50 I can't use TinyCCb because it is compiler, it compiles for x86. I need it to be platform-independent interpreter (no compile) 02:19:06 compile the linux kernel in 15 seconds on a p4 02:19:08 Oh, you want it platform-independent. 02:19:11 Um. 02:19:15 Your best bet is Clang. 02:19:30 cool 02:19:35 LLVM based 02:19:49 Yeah, not platform *independent*, but ported to a lot more things. 02:20:12 No! It needs to be platform independent. It needs to interpret only, not compile. 02:20:20 Nobody's written one. 02:20:27 So, a c like language? 02:20:42 It wouldn't be hard, it's just not been done. 02:20:42 Sounds easy enough 02:20:45 sudo apt-get install clang 02:20:57 The only preprocessor command that is needed is #define I don't need #if or #include 02:20:58 * SgeoN1 slaps ^D 02:21:05 And I don't need support for comments either 02:21:10 http://clang.llvm.org/ 02:21:22 Can llvm be interpreted? 02:21:26 zzo38: Basically, you don't need the preprocessing pass outside of macros. 02:21:29 lol 02:21:36 SgeoN1: Hypothetically, yes. 02:21:39 SgeoN1: Yes, a C like language, and it has to be a proper subset of C and the syntax like C 02:21:45 llvm on tinycc running CLang 02:22:08 zzo38: I suggest you write a C interpreter. Sorry. 02:22:34 Can it be full C? Any reason for the "proper"? 02:23:32 -!- augur has joined. 02:23:46 SgeoN1: It can be full C, but I don't need it. Programs generated by CTANGLE are compiled using a normal C compiler anyways, I only need the interpreter so that you can write functions for generative programming in your program. 02:24:37 So don't call it proper subset? 02:25:12 SgeoN1: What I mean is not all features of C are needed 02:25:36 Subset, without proper, allows for that 02:25:57 SgeoN1: Yes, sorry, I just made a mistake 02:26:10 However, I still don't need it to support full C 02:26:16 you want an interpereter that can, in real time, run your code like a script 02:26:45 and i first tried esolang (common short for for esoteric language) ;; we're not idiots. 02:27:03 EOF: Yes, like that 02:27:17 email me your code at hamiltonham9@hotmail.com ;; we know how to use pastebins 02:28:28 O, and it should be able to call a function that was parsed before, even during the parsing of another function. 02:30:33 My idea is that if CTANGLE finds a identifier "abcdef" and the function "abcdef$" exists (defined by interpreting), then it will call abcdef$() to generate the C codes for that (which does not necessarily have to be interpreted; the generated code might be used only by the C compiler). 02:30:50 EOF: No offense or anything, but are you familiar with the Dunning-Kruger effect? 02:31:43 (And by favorite pastebin is sprunge) 02:31:53 ^^men and out of place 02:32:00 ima leave 02:32:07 men? 02:32:19 Confusing use of ^ there, ^D. 02:32:51 it's ^^ = 0x0015 02:32:56 s/it's // 02:36:05 two ^s mean two lines up 02:36:26 I'm still not sure what "men" is supposed to mean. 02:36:46 mean 02:36:59 Ah 02:38:10 the humor of that was that it's directly proving a link to the Dunning-Kruger effect, and so was that. 02:39:55 -!- augur has quit (Ping timeout: 276 seconds). 02:39:58 Please specify "that" further. 02:40:00 I think being here tempers that effect in me 02:40:17 Firefox crashed again for absolutely no reason 02:40:37 Maybe Chrome will be better 02:40:47 Chromium 02:41:45 Although I still think I'm better than the idiots who surround me irl 02:41:46 -!- mtve has quit (Ping timeout: 358 seconds). 02:42:04 -!- mtve has joined. 02:51:16 Maybe IceWM will be less crasshy 02:53:03 It looks like with TinyCC I might be able to make it generate codes for the virtual machine 02:53:18 If I can do that, then I can use it. 02:53:53 SgeoN1: Which OS? 02:54:50 Puppy Linux 02:55:06 try ubuntu :P 02:55:17 I will asap 02:55:52 I just sent an apology to someone for being an asshole when I was in 7th grade 03:01:56 I will try libtcc. 03:02:14 SgeoN1: Seriously? 03:02:24 How assholish? 03:03:56 How do I open a tab on this thing 03:04:13 are there not many languages that are entirely event-driven? the wiki doesn't seem to have an event-driven category... 03:04:23 -!- augur has joined. 03:04:25 quintopia: there are one or two, I think 03:04:30 and the idea entertained me for a while -- and still does 03:04:38 i believe there is one on the wiki. good luck finding it 03:05:35 I use Ubuntu at Free Geek. It is not the best but it is a full Linux distribution and it works. I use the command-line mostly. 03:06:49 LSL... 03:06:57 Whatever it is: no. 03:07:12 quintopia: http://esolangs.org/wiki/Special:Search?search=event&fulltext=Search 03:07:16 our search does work quite well 03:07:21 It doesn't do them in a nice way, but... 03:07:23 alise: 2iota 03:07:26 of course there's beta-Juliet and Portia and 2iota, yes 03:07:29 cpressey's langs 03:07:32 but i recall another too 03:08:06 the rest on that search appear related to ABCDXYZ 03:08:14 which doesn't really count 03:08:23 Hmm, I swear I remember another. Oh well. 03:08:35 quintopia: you forgot beta-Juliet and Portia (2iota's predecessors, although utter failures) 03:10:10 alise: i didn't have to mention them since you already had 03:10:11 * alise gets inspired by Biota 03:10:15 quintopia: right :P 03:10:31 It'd be nice to have a language like Biota, except instead of having a boring data counter, the program is expected to evolve itself. 03:10:43 Like Befunge if the self-modification was integral, not an tacked-on thing. 03:11:01 what's the name of the original OOL again? starts with a P? 03:11:02 The results of a conditional, say, moves a code path to the right where the IP will hit it, or something. 03:11:17 Programs restructure themselves gradually with algorithms, turning calculations into "output" data... 03:11:20 quintopia: object oriented, you mean? 03:11:59 yes 03:12:06 the one that was entirely picture based 03:12:12 came before actual digital computers 03:12:13 quintopia: Uhh... Modula-2, successor to Pascal? 03:12:16 ... What? 03:12:28 Are you not thinking of the lambda calculus or something? 03:12:36 no 03:12:36 I know of no object-oriented language predating computers. 03:12:53 well, i should have thought you would, since this is esoteric 03:13:06 this should be the one place on earth that remembers 03:13:13 it's coming back now 03:13:15 We're not omnipotent, and it's a distinct possibility that you're hallucinating. 03:13:19 the name ends in kalkul 03:13:27 That rings a bell... 03:13:34 ...nope. 03:13:40 Maybe... 03:13:49 quintopia: wait -- what's your native tongue? 03:14:02 second result for *kalkul on google is the pi calculus 03:14:05 pi-kalkül 03:14:08 german 03:14:27 that doesn't predate computers though, afaik 03:14:57 * SgeoN1 will probably be eating dinner at the pizza place tomorrow. Or maybe the day after 03:16:04 found it 03:16:09 what's it called? 03:16:19 Plankalkl 03:16:28 yes that's what came to mind 03:16:33 but i couldn't place it 03:16:37 it's not OO though 03:17:55 you're right that it doesn't predate digital computers 03:18:01 apparently the Z3 used binary logic 03:18:14 but it definitely predates transistors... 03:20:02 quintopia: you're job is to make my language i just thought about 03:20:55 alise: working on it. will have it uploaded in a minute. 03:21:01 good 03:22:25 quintopia: do you like roguelikes? 03:25:16 alise: wtf are you doing on #haskell ? 03:25:40 Mathnerd314: err, what? 03:25:51 i've been visiting #haskell since... early 2008 03:26:19 * Mathnerd314 searches logs 03:26:30 you may want to grep "ehird" 03:27:20 How exactly do I use libtcc anyways? Will it do what I need it to do? 03:28:17 well, it compiles... not interprets 03:28:19 is that okay? 03:28:28 * pikhq is amazed at how many accents the UK has 03:28:42 hmm, revenantphx wasn't in #haskell the last time 03:28:50 he seems terribly irritating 03:29:23 There's even accents retaining the T-V distinction. 03:29:23 he's acting like he's crossed the desert of enlightenment and found the truth and is now lecturing hopeless peasants :P 03:29:28 pikhq: T-V distinction? 03:29:38 alise: It looks like I can write a code generator, if I can write one that compiles to a virtual machine code in memory, I could do that. Can it work with partially compiled programs? 03:29:48 alise: In English, the difference between "thou" and "you". 03:29:57 Specifically, the formality bit. 03:30:14 It's a feature in most languages in Europe (sprachbund, not from PIE) 03:30:20 What is the difference between "thou" and "you"? 03:30:29 Oh, this has nothing to do with Haskell. 03:30:30 precisely. 03:30:30 ... 03:30:41 pikhq: Err, nobody says "thou", so I'm assuming that's what you mean. 03:30:42 zzo38: Thou is the second person singular informal. You is the second person singular formal and the second person plural. 03:30:47 alise: Yes. 03:30:51 zzo38: Degree of politeness. 03:30:52 They still use "thou". 03:31:12 pikhq: *not what you mean. 03:31:15 pikhq: Surely not. 03:31:16 Nobody does. 03:31:21 Thou shall not use the word "shall". It is outdated. 03:31:23 alise: A handful of accents do. 03:31:26 I know people and they do not do this. 03:31:30 pikhq: Maybe as "tha" or something. 03:31:47 It's Oop North, so that's probably what it comes out as. :P 03:31:51 pikhq: O, so "you" is supposed to be plural. 03:32:01 zzo38: Yes. 03:32:07 pikhq: Wow just look at your #haskell tab's scrollback for a bit and admire how irritatingly loud revenantphx is. 03:32:22 At least when I dominate all the discussion here I'm not /too/ idiotic. 03:32:26 Before the development of the T-V formality thing (it started in French), it was just a singular vs. plural distinction. 03:32:43 In... Pretty much all Indo-European languages. 03:32:44 ah, the french. experts at making things stupider. 03:33:00 Yup. 03:34:06 And of course, now in *most* accents "thou" is some sort of hyper-formal. 03:34:10 kmc: I have homework for AP CS in hava actually 03:34:12 Because that's how you refer to god! 03:34:12 what up yo 03:34:16 AP -- isn't that high school in the US? 03:34:41 i'm pretty sure this guy is an irritating rubyist who found out about haskell on, i dunno, twitter, p=.99 03:34:43 alise: It's a high school class that can become college credit upon passing the relevant AP test. 03:35:02 Typically for the more menial college courses. 03:35:12 pikhq: right. so the veteran of the desert of enlightenment, two-week master of haskell, dominator of all #haskell discussion is complaining about his high-school homework 03:35:20 US education being what it is, high schools treat them as the holy grail of intelligence. 03:35:26 RIP #haskell 1997 or something -- 2010 03:35:28 "ZOMG YOU TOOK AP CALCULUS!" 03:36:09 He wont even let me align variable assignments D: 03:36:16 MY CODE IS SO UGLY WITHOUT A BUNCH OF POINTLESS WHITESPACE IN IT 03:36:22 WHY IS MY TEENAGED LIFE SO IMPOSSIBLE ;__; 03:36:51 It amazes me that the common perception of a lot of college classes is that they're impossible. 03:37:05 For instance: calculus is considered by many (idiots) the hardest math class in existence. 03:37:31 * alise wonders if there's a shorter way to get the sign of a number in Python than x/abs(x)... 03:37:40 ideally i want to map, say, negative to 0 and positive to 1... 03:37:43 Yes, the calculus of derivatives and integrals. 03:39:00 alise: cmp(x, 0) 03:39:12 bsmntbombdood: very nice 03:39:18 has 0 == 0, but oh well, what can you do 03:39:33 bsmntbombdood: (I'm golfing and trying to get myself an (-inf,-inf) to (inf,inf) list.) 03:39:40 maybe i'll just write a class for it 03:39:53 but that's *so* *much* *code*! 03:40:22 wait i can just use a hash table 03:40:23 duh 03:40:24 isn't (inf, inf) = {inf}? 03:40:49 bsmntbombdood: (inf,inf) being a coordinate tuple there, or what do you mean? 03:41:06 (inf, inf) being a range 03:41:10 no, coordinates here 03:41:22 i just want an infinite 2d plane centred at (0,0) 03:41:26 but again, a dictionary works 03:42:04 so i got a job programming php 03:42:09 i miss python :( 03:46:07 I can also try PicoC but it says very small, but various things don't work 03:46:42 or even real languages in general 03:50:29 def D(x,y): 03:50:29 PicoC seems to work badly. After a variable "x" is declared, then using "#define x 4" results in "'oUB' is already defined" 03:50:29 for Y in range(y,y+80): 03:50:29 for X in range(x,x+24): 03:50:29 s.addch(Y-y,X-y,w.get((Y-40,X-12),46)) 03:50:32 now why doesn't this work... 03:51:46 Leaving the house, bleh 03:52:23 If 'y' is undefined, a code such as "void qq() { y++; }" will not result in an error until you try to call "qq();" 03:53:01 "X-y" lol 03:54:07 I also got the message "macro is not a function - can't call". But it is a macro that takes parameters. 03:55:56 -!- SgeoN1 has quit (Ping timeout: 245 seconds). 03:56:27 -!- SgeoN1 has joined. 03:56:40 I think PicoC is broken. 03:58:51 If 'y' is undefined, a code such as "void qq() { y++; }" will not result in an error until you try to call "qq();" 03:58:52 lolwut 03:58:57 Look at the defects in PicoC listed here: http://code.google.com/p/picoc/issues/list 03:58:58 static typing ftw 03:59:44 aha! i knew there were people here who actually programmed for a living! 03:59:47 -!- EOF has quit (Ping timeout: 252 seconds). 03:59:57 I suppose it is OK that the qq() function given there causes an error only when you call it, but it isn't OK that "#define x 4" results in "'oUB' is already defined" 03:59:59 Should I have an evil plan to save the world? I think so. 04:00:00 quintopia: err, who? 04:00:03 oh, bsmntbombdood 04:00:07 yeah, bsmntbombdood never comes here any more 04:00:08 case in point 04:00:10 Also \ at the end of a macro line is not accepted 04:00:15 i think cpressey is the only regular who develops professionally 04:00:31 and, really, it's his channel, we just live in it :) 04:01:13 quintopia: I used to. 04:01:25 I'm currently unemployed. And a college student besides. 04:01:37 what level? 04:02:32 Undergrad 04:02:40 "PicoC is a very small C interpreter for scripting. It was originally written for scripting a UAV's on-board flight system and it's also very suitable for other robotic, embedded and non-embedded applications too. 04:02:42 The core C source code is around 4000 lines of code. It's not intended to be a complete implementation of ISO C but it has all the essentials. When compiled it only takes a few k of code space and is also very sparing of data space. This means it can work well in small embedded devices. It's also a fun example of how to create a very small language implementation while still keeping the code readable." 04:02:51 zzo38: i think it's not aiming for absolute standards-compliance 04:02:53 pikhq: i thought you sysadminned 04:03:11 alise: And developed. Mostly patches to make programs comply. 04:03:44 if you can call that programming :D 04:03:50 alise: That's OK it is doesn't do it absolutely, but it has to work good, instead of working badly 04:04:00 I spent a month doing that to a program once. It counts! 04:04:08 (23, 79, 46) 04:04:13 Why would this break curses addch()?! 04:04:45 wtf with +79 it yields no error 04:04:49 what's up with that. 04:05:24 Isn't there any better program than this??? 04:05:25 -!- augur has quit (Ping timeout: 255 seconds). 04:05:38 zzo38: it's not bad, it's just not designed for your purpose 04:05:44 you could probably buy one of the proper c interpreters 04:05:54 http://www.softintegration.com/ 04:05:56 Ch for instance 04:06:32 I can't do with buying one, it has to be free software 04:07:09 I will try to use PicoC anyways, though, to see if it might work 04:07:20 Maybe I can modify it a bit, if necessary 04:09:54 Ch is not even platform-independent 04:10:52 it's portable 04:10:56 so yes it is 04:13:06 PicoC also does not seem to supprt typedef (although I do not get an error about 'typedef' being undefined) 04:13:08 http://sites.google.com/site/redcodenl/patent-infringement 04:13:33 "You wrote some code using an FFT to identify music files. We also do that, and make money off it. Therefore, stop it. It's illegal or something!" 04:14:19 -!- augur has joined. 04:15:31 heyo 04:15:46 hi 04:16:53 pikhq: I challenge you to a COMPETITION 04:17:01 hows life, alise 04:17:04 augur: lifey 04:17:07 awesome 04:17:32 someone posted a picture of some random boy claiming it was you 04:17:33 :| 04:17:45 Yes. 04:17:58 If you didn't notice that was the same person whose trolling you were encouraging the day before. 04:18:14 i dont really pay attention to this channel so :D 04:18:14 Hahawtf? 04:18:27 Gregor: wat 04:18:27 but i will say, the boy in the pic he posted was a total hotty 04:18:37 and if you had become that boy i would've been all "zomg wat" 04:18:44 So I've avoided rape. 04:18:50 That counts as a good day in my opinion. 04:18:59 It would only start as rape. 04:19:10 Perhaps I can just fix PicoC. 04:19:19 alise: well, it wouldn't've been rape 04:19:20 YES PICOC LET US TALK ABOUT PICOC. 04:19:21 trust me 8D 04:19:24 augur: PICOC! 04:19:47 something like that :D 04:19:54 ... 04:20:03 but seriously now 04:20:12 good night 04:20:27 Really, if you want a "C interpreter", you'd probably be best to munge something around a C compiler. 04:21:06 nah, you'd be better off having a simulated machine that you compile to! 04:21:17 Gregor: No, it has to be totally portable, apparently. 04:21:21 ok night 04:21:23 so tired 04:21:50 alise: How totally portable? Windows? OS X? Hardware is easy in this case, software is not. 04:22:08 Gregor: Anything that runs C I'd assume. 04:22:28 Right, so I suppose munging something around 20 C compilers is no fun :P 04:22:40 Although I'm pretty sure zzo38's code flagrantly violates the C standards in more ways than one. 04:24:12 All C code flagrantly violates C standards :P 04:25:03 Gregor: Rather, all "C" "code" is not actually C code :P 04:25:59 Fair enough :P 04:26:11 $ ./picoc -i 04:26:13 starting picoc 04:26:15 picoc> a 04:26:17 picoc: lex.c:500: LexTokenise: Assertion `ReserveSpace >= MemUsed' failed. 04:26:19 Aborted 04:26:19 xDDD 04:26:21 I'm impressed. 04:27:15 Gregor: 04:27:17 #include 04:27:17 #include 04:27:17 int main(void) { return puts("Hello, world!") == EOF ? EXIT_FAILURE : EXIT_SUCCESS; } 04:27:23 I'm pretty sure that's bona-fide C code. 04:27:34 In what way is that not bona-fide C code? 04:27:37 That's fine. 04:27:44 Well, it's the first piece ever written! 04:28:02 Ohohoh 04:28:05 :P 04:28:18 -!- SgeoN1 has quit (Ping timeout: 240 seconds). 04:29:08 "Given a a series of integers print φ(n) for each of them." 04:29:10 Oh come on. 04:29:24 Next anagolf problem: "Implement a computer algebra system!" 04:29:48 -!- SgeoN1 has joined. 04:30:06 wd@>5&p:".;._2(1!:1)3 04:30:09 21 characters of J including IO 04:30:12 Like everything else 04:33:35 NOTE TO SELF: install J tomorrow. 04:34:23 Goodnight. 04:34:24 Bye. 04:34:27 -!- alise has quit (Quit: Leaving). 04:38:07 -!- iGO has joined. 04:54:27 [19:37.21] Quadrescence ? 05:05:30 PicoC does not even support the "goto" command. 05:22:27 -!- zzo38 has quit (Remote host closed the connection). 05:35:57 damn 05:36:36 http://esoteric.voxelperfect.net/wiki/1mpr0mp2 05:36:54 too practical to be on esolang? 06:03:07 quintopia: "CP0, CP1, etc.: The C prefix indicates that the output pin should be driven low. Again, this is only legal for" (cuts off) 06:07:06 quintopia: To me, that doesn't look too practical... 06:11:40 How many languages are there that if you learn, then walk away from, become unreadable? 06:11:45 J is one 06:22:26 Ilari: thanks 06:29:04 Ilari: it seems obvious to me that it is Turing-complete. what would be the shortest way to prove it? 06:33:59 -!- SgeoN1 has quit (Quit: Bye). 06:36:31 06:38:17 -!- cheater99 has joined. 06:41:00 -!- cheater99 has changed nick to cheater00. 06:44:32 -!- GreaseMonkey has joined. 06:44:32 -!- GreaseMonkey has quit (Changing host). 06:44:33 -!- GreaseMonkey has joined. 06:50:40 quintopia: Implement algorithm that takes subset of brainfuck (+, -, <, >, [ and ]) and outputs 1mpr0mp2 program that halts if and only if the original brainfuck program halts? And then note that there is obiviously nothing in the language that's super-turing... That's at least one way. 07:50:36 -!- GreaseMonkey has quit (Remote host closed the connection). 07:59:59 -!- clog has quit (ended). 08:00:00 -!- clog has joined. 08:11:16 -!- iamcal has joined. 08:11:17 -!- cal153 has quit (Read error: Connection reset by peer). 08:33:00 -!- augur has quit (Remote host closed the connection). 08:48:26 -!- Leonidas_ has changed nick to Leonidas. 09:19:12 -!- MigoMipo has joined. 09:23:37 -!- Phantom_Hoover has joined. 09:24:03 Is it just me, or does it look like a Dinosaur Comics/Doctor McNinja crossover is headed our way? 09:28:18 -!- Phantom_Hoover has quit (Ping timeout: 240 seconds). 09:31:30 -!- Phantom_Hoover has joined. 09:31:59 -!- Phantom_Hoover has quit (Client Quit). 09:32:18 -!- Phantom_Hoover has joined. 09:36:35 -!- mjmscott has joined. 09:37:14 -!- mjmscott has quit (Read error: Connection reset by peer). 09:42:01 -!- Phantom_Hoover has quit (Ping timeout: 240 seconds). 10:24:06 -!- tombom has joined. 10:37:32 -!- iGO has quit. 10:42:26 have added a new category suggestion to the talk page. quickly approve it so its absence will stop annoying me 11:15:21 -!- Wamanuz has joined. 11:35:16 -!- FireFly has joined. 11:42:31 fizzie, doing some quick benchmarking on complex queries (such as the ? ratio ones) it seems that postgres is faster than sqlite 11:45:57 hm not doing case insensitive for nick. Could affect it 11:47:23 ah there is "ilike" 11:50:48 hm also on queries not using the nick column at all 11:52:19 select type,count(*) as cnt from logs group by type order by cnt desc; takes 1519 ms in postgres according to itself. Not sure how to get such exact value in sqlite, but wall clock based timing indicates about 13 seconds there 11:54:15 hm it seems based on explain that sqlite creates a temporary table? And postgres does a seq scan on the table then a hash aggregate 11:54:25 both sorts at the end 12:04:36 -!- olsner has quit (Quit: Leaving). 12:15:20 -!- tombom has quit (Quit: Leaving). 12:41:37 -!- BeholdMyGlory has joined. 12:44:31 -!- MigoMipo has quit (Read error: Connection reset by peer). 12:49:36 fizzie, eh found a bug with your script: it doesn't handle modes properly at all! 12:50:31 as in, it doesn't capture the actual mode change 12:50:37 just that a mode change was done 12:53:26 -!- BeholdMyGlory has quit (Remote host closed the connection). 12:53:42 -!- BeholdMyGlory has joined. 12:57:39 fizzie, btw to speed up full import: create indexes after importing data. This is faster for both sqlite and postgres 12:58:14 *much* faster 13:23:13 -!- yiyus_ has quit (Read error: Connection timed out). 13:26:40 -!- Zuu has quit (Ping timeout: 276 seconds). 13:28:10 sqlite> select (select count(*) from logs where type = 3) - (select count(*) from logs where type in (4,5,6)); 13:28:10 308 13:28:12 strange 13:28:22 that is joins - (parts + quits + kicks) 13:28:34 the value ought to be equal to number of people in here 13:29:05 that is around 45 or so (since the last log fetched is not from this very instant, but about an hour ago or such) 13:30:23 hm clog losing connection probably resulted in some mismatch 13:32:30 fizzie, oh and: initial data import into sqlite is faster than for postgressql, but queries that aren't able to use indexes directly are significantly faster in postgres. 13:33:20 -!- Zuu has joined. 13:33:21 -!- Zuu has quit (Changing host). 13:33:21 -!- Zuu has joined. 13:33:45 using the COPY construct would probably be faster for importing into postgres, but that would require parsing the logs and storing them into a stream first. Would mean a major rewrite of the program. 14:21:59 -!- alise has joined. 14:26:01 01:29:03 I miss the time I was a "real" student; those get a 50 % discount. 14:26:02 http://www.phdcomics.com/comics/archive/phd072709s.gif 14:26:22 02:42:26 have added a new category suggestion to the talk page. quickly approve it so its absence will stop annoying me 14:26:26 *very* unlikely 14:26:33 Graue is touchy about categories. 14:27:21 04:41:09 In the raw clog format, around 117 MB. 14:27:24 only 50 MB not long ago :) 14:27:27 fizzie, current versions with fixed parsing of mode changes and putting new nick for nick changes into target column. To make full import fast: move the db.commit() call to top level and add the indexes (apart from primary keys) after the data has been imported: http://sprunge.us/TEhQ http://sprunge.us/QMgO 14:30:48 "The computer needs to restart to finish installing updates. Please save your work before continuing." 14:30:55 for kernel headers?! oh wait, no, kernel update 14:31:01 * alise ponders ksplice 14:31:48 alise, ubuntu? 14:31:54 yeah 14:31:59 alise, if so yeah the last update would have worked fine with ksplice I think 14:32:07 checking the changelog 14:32:18 yeah i don't really care enough though, i'll just not reboot 14:32:32 Vorpal: btw, i now have strong empirical evidence that this glossy screen is /far/ superior to other glossy screens 14:32:39 I saw someone with a Dell Latitude laptop 14:32:47 alise, what brand is your laptop 14:32:49 and it was more than twice as glossy as this one 14:32:52 Vorpal: Toshiba 14:32:54 ah 14:32:57 well, could be 14:33:18 It's glossy but reflections are very, very subtle and if you focus on the screen, completely invisible most of the time. 14:33:37 Of course, in daylight you can see them slightly, but then again, it's usable in sunlight, unlike matte displays. 14:34:00 Actually the part that accounts for almost all perceived reflections is the plastic border around the screen; that's pretty reflective. 14:34:17 Basically I always forgot I was using a glossy display when I first got this. 14:34:23 The Dell screen was /awful/. 14:34:30 * Vorpal wonders if "select count(*) as cnt,length(body) as len from logs where type = 0 group by len order by cnt desc;" will work 14:34:50 hm numbers seems plausible 14:34:51 seem* 14:35:20 length = 3 is most common at 54081 lines of that length 14:35:38 second place is length 2 at 48851 14:35:59 there is one message of length 0?? 14:36:47 Vorpal: is there a way to limit 1 for each group? 14:36:56 that is, only show the first one for the first count thingy? 14:37:12 alise, you mean, one "example" message of each length? 14:37:18 i guess so 14:37:25 7|409 14:37:26 7|414 14:37:26 7|417 14:37:26 -> 14:37:29 7|409 14:37:45 eh, what is the 7 here? 14:37:57 Urgh, I hate SQL so much. 14:37:58 oh wait, you are using sqlite 14:38:01 Vorpal: I just ran your query. 14:38:21 alise, right, I was doing it in postgre, so I got confused by the different output format 14:38:26 Also, when I'm doing a query that isn't selecting for me, just returning results that involve me, or you, how can I group things? 14:38:33 That is, "treat X, Y and Z as the same for grouping". 14:39:30 alise, you mean nicks? well (nick like 'alise%' or nick like 'ehird%' or nick like 'tusho%') is what I used so far. Then separate queries for me and you 14:39:39 which is indeed cumbersome 14:39:46 alise, maybe create a few with them merged :D 14:39:51 a view* 14:39:55 not "a few" 14:39:56 yes ut 14:39:58 *but 14:39:58 weird typo 14:40:00 sqlite> select count(*), nick from logs where type = 0 group by nick order by count(*) desc limit 10; 14:40:07 how can I merge the ehird, alise, tusho groups into one? 14:40:20 i'm not very good at sql, mostly because i hate it 14:40:48 alise, I can't think of a way. Maybe some string function to replace all ehird and so on with alise and then select on that sub query? 14:41:02 See now in Prolog... :) 14:41:14 alise, yeah right, so do that then 14:41:19 So basically I replace the logs table with a subquery? 14:41:22 Also, yeah, yeah, I'm on it... 14:41:41 alise, anyway those 7 things there, they were "7 times have we seen a string of length Okay, so X == (select * from X). (I've never sub-frommed before.) 14:41:44 * 14:41:49 Vorpal: ah 14:41:51 alise, so merging those seems a bit strange 14:41:57 plot it! 14:42:00 plot the result 14:42:13 so most of our messages are 3 characters long xDD 14:42:23 alise, would have been done already if you hadn't been buggering me about SQL the past few minutes :P 14:44:20 sqlite> select count(*), nickx from (select *, replace(nick,"ehird","alise") as nickx from logs) where type = 0 group by nickx order by count(*) desc limit 10; 14:44:22 not sure this works :( 14:44:37 oh, it does 14:44:53 -!- BeholdMyGlory has changed nick to Behold. 14:44:53 hmm 14:44:59 is there a replace() that operates on "foo%", I wonder... 14:45:01 -!- Behold has changed nick to BeholdMyGlory. 14:45:39 seems not 14:45:41 stupid sql 14:46:09 * alise just replaces ehird`/alise`/ehirdiphone/aliseiphone/iEhird 14:46:26 Vorpal: you've never used a nick other than AnMaster or Vorpal for any significant length of time, right? 14:47:49 sqlite> select count(*), nickx from (select *, replace(replace(replace(replace(replace(replace(replace(replace(nick,"iEhird","alise"),"aliseiphone","alise"),"ehirdiphone","alise"),"alise`","alise"),"ehird`","alise"),"AnMaster","Vorpal"),"tusho","alise"),"ehird","alise") as nickx from logs) where type = 0 group by nickx order by count(*) desc limit 10; 14:47:49 371209|alise 14:47:49 220357|Vorpal 14:47:53 Vorpal: Ha ha ha observe my dominance. 14:48:01 lament is on that list now, haha 14:48:20 Re creating indices only after a large import, that's a very common speedup trick; didn't think about it yesterday, though. 14:48:36 fizzie: Is there any reason other than disk space not just to add an index on fucking everything?# 14:48:39 *everything? 14:48:46 371209|alise 14:48:46 220357|Vorpal 14:48:46 101196|ais523 14:48:50 Look at that. 14:48:55 (I did scarf -> ais523.) 14:50:25 hm 14:50:41 alise, I used AnMaster|ipv6 and so on 14:50:43 and a few more 14:50:45 Incidentally, is there some particular reason why you do that with a subquery and not just select count(*), replace(...) as nickx from logs where type = 0 group by nickx order by count(*) desc limit 10? 14:50:48 alise, but far less than you 14:50:54 "Moreover, some widely used database systems include ideas and algorithms developed for Datalog. For example, the SQL:1999 standard includes recursive queries, and the Magic Sets algorithm (initially developed for the faster evaluation of Datalog queries) is implemented in IBM's DB2." 14:51:00 Datalog, oh yeah. 14:51:04 so we are talking about 2000 messages instead of 20000 or such 14:51:11 fizzie: Um... because Vorpal was stupid and told me to use a subquery. 14:51:13 fizzie, do you know how to get log y scale in gnuplot? 14:51:16 Vorpal: So basically I WIN. 14:51:20 -!- EOF has joined. 14:51:29 sqlite> select sum(length(body)) as s, nickx from (select *, replace(replace(replace(replace(replace(replace(replace(replace(replace(nick,"scarf","ais523"),"iEhird","alise"),"aliseiphone","alise"),"ehirdiphone","alise"),"alise`","alise"),"ehird`","alise"),"AnMaster","Vorpal"),"tusho","alise"),"ehird","alise") as nickx from logs) where type = 0 group by nickx order by s desc limit 10; 14:51:29 13946929|alise 14:51:29 8420495|Vorpal 14:51:29 5119631|ais523 14:51:35 That's pretty impressive. 14:51:37 alise, I said I couldn't think of any other way than subquery 14:51:47 alise, I'm no SQL expert either 14:52:00 iHATE SQL 14:52:11 EOF, who are you? 14:52:21 Vorpal: You can simply do "using 1:(log($2))" to get the proper log-scale shape, but then the ticks and labels and such will have those raw log numbers. 14:52:24 internal string array databases FTW 14:52:36 EOF, for logs of this channel? 14:52:53 fizzie, hm 14:53:00 yeah 14:53:01 "set logscale y" (or "set logscale y N" for base N) if you want a "proper" log-scaled plot. 14:53:02 carry on 14:54:48 fizzie, what about hiding legend? 14:55:00 Vorpal: set key off 14:55:02 ah 14:55:20 Or maybe "unset key". Or maybe both work. 14:55:23 hmm 14:55:24 Error: no such function: mean 14:55:27 Well...fuck...you? 14:55:36 fizzie, set key off worked 14:55:40 Oh, avg(). 14:55:41 now to figure out sizing the thing 14:55:43 How wonderfully imprecise. 14:55:49 could you guys give me an indication of WTF you are talking about? 14:55:50 alise: It's also just sum/count, anyway. 14:55:54 Error: no such function: mean Well...fuck...you? <-- how mean 14:55:58 alise: There's a soundex approximate-matching function in sqlite, you could use that for nick-folding. 14:56:00 fizzie: aka mean :P 14:56:05 fizzie: haha wow 14:56:18 EOF, we are generating useless statistics on the logs for this channel (from 2004 to now) 14:56:26 yay 14:56:30 Whoa, sorting by average message length changes EVERYTHING. 14:56:34 are the logs open data? 14:56:35 454.0|fizzief 14:56:36 325.5|ElMexicano 14:56:36 325.5|Linuxiano 14:56:36 325.0|Yst 14:56:36 284.5|Toyd 14:56:36 241.0|ololobot_ 14:56:38 238.0|mycroshift 14:56:41 221.0|ghostless 14:56:42 203.0|fizzieg 14:56:44 fizzief? 14:56:45 198.666666666667|Mathematica 14:56:46 EOF: no extremely secret 14:56:48 nobody can have them 14:56:50 what about me? 14:56:51 Hey, that Mathematica was my bot. 14:56:54 EOF: nope 14:56:57 EOF, did you check topic? 14:56:58 you have to pay for them 14:57:21 05:23:55 fizzie, hm I looked at Gregor's hg repo for logs. why is there a fizzie directory in https://codu.org/projects/esotericlogs/hg/index.cgi/file/0ad952ce6894 ? 14:57:26 don't check that out without asking him first 14:57:32 it fucks with his server 14:57:37 due to the traffic 14:57:58 cool logs 14:57:59 Vorpal: I can't quite recall what "fizzief" was for; I usually append a letter that's somehow related to the client or computer or something, but I don't know what "f" might stand for. 14:58:17 i'm going to merge them all 14:58:18 fizzie, also egobot somehow managed a line of length = 0 14:58:21 I have no idea how 14:58:32 14:58:38 Vorpal: perhaps a control char 14:58:40 that the client stripped 14:58:42 but not the server 14:58:45 client = clog 14:58:48 alise, hm, perhaps 14:59:00 did you see any empty line from me ther? 14:59:01 there* 14:59:05 no 14:59:06 ./printall > fulllog 14:59:09 not \0, presumably 14:59:15 but perhaps \1 or \255 or something like that 14:59:18 alise, ah, was a single color-code start 14:59:18 EOF: seriously? 14:59:20 Vorpal: Did you filter for type=0? There's one zero-length "act" there, at least. 14:59:20 cat * > log 14:59:25 yeah 14:59:25 how can you not think of that? 14:59:31 fizzie, yes I filtered 14:59:39 fizzie, since length(NULL) seems useless 14:59:43 select * from logs order by random() limit 10; is fun 14:59:47 315078|2007-07-22 13:36:26|blahbot`|||0|Reloaded. 14:59:49 my reloadable bot! 14:59:52 but i would have to cd into the workign directory :? 14:59:59 EOF: cat logs/* > full 15:00:06 and i is MAJOR NUVB 15:00:08 writing a script is utterly pointless 15:00:15 yeah 15:00:18 you said you were experienced with unix-like systems and gnu/linux especially 15:00:29 Aw, [the soundex() function] "is only available if the SQLITE_SOUNDEX compile-time option is used when SQLite is built", and Ubuntu's sqlite doesn't do that. :( 15:00:35 i don't use cat so much now 15:00:39 fizzie: bugger 15:00:43 EOF: Why the hell not...? 15:00:54 ./07.07.17:05:42:48 I am also from ~do, but with the new command;command form. 15:00:58 oh yeah, the gaping security hole feature 15:02:54 alise, here is the plot in question: http://sprunge.us/XGbZ 15:03:08 * Vorpal runs 15:03:14 Vorpal: gtfo 15:03:32 it's not like it's encoded perfect mathematical curves there either 15:04:21 alise, here anyway: http://sporksirc.net/~anmaster/tmp/msglen.svg 15:04:28 -!- yiyus_ has joined. 15:04:40 alise, and indeed it doesn't 15:04:53 Hmm, how best to render that plot? I know, I'll set the font to Comic Sans. 15:04:56 Ahh, thank god it's in svg. 15:05:01 Now it's much more readable. 15:05:07 alise, what XD 15:06:21 alise, anyway reuploaded it to http://sporksirc.net/~anmaster/tmp/msglen.svg with less square size 15:06:26 to make it easier to read 15:06:49 ./07.07.19:10:39:55 ^y^x$^x^y$^x^y^x$^x^yx$xx$^x$^x^yx$xx^x$^x^yx$xx$$^x^y^x$^x^yx$xx$^x$^x^yx$xx^x$^x^yx$xx$^x^y^x$^x^yx$xx$^x$^x^yx$xx^x$^x^yx$xx 15:06:51 Okaaaayyyy... 15:08:20 !bf ++[>+++<-]>+. 15:08:25 :) 15:08:30 that was what gave the one char thingy 15:08:31 before 15:08:36 * Vorpal waits for EgoBot to react 15:08:41 !help 15:08:42 help: General commands: !help, !info, !bf_txtgen. See also !help languages, !help userinterps. You can get help on some commands by typing !help . 15:08:55 hm maybe it it doesn't pass the ircd any more 15:09:13 ^bf ++[>+++<-]>+. 15:09:22 hm presumably 15:10:10 14:34:45 how many people here pronounce "char" as "care" and how many pronounce it as in "charred"? 15:10:13 "car", actually... 15:10:58 ... 15:11:01 k 15:11:17 EOF, so who are you? Someone relatively new here or someone who just changed nick? 15:11:46 he's new. 15:11:51 right 15:11:54 someone who learned of this channel yesterday 15:11:55 Vorpal: btw I /msg'd you 15:12:07 alise, you didn't? 15:12:09 did you now 15:12:17 Vorpal: i did a while back 15:12:20 not sure if you saw 15:12:33 alise, err, I think I had a bogus setting in my client 15:12:38 Vorpal: I can resend. 15:12:42 alise, yes please 15:13:13 Vorpal: done 15:14:16 :) 15:15:55 Vorpal: any ideas on how to count occurences of a substring in a string in sql? 15:16:01 maybe sqlite is lacking here 15:16:06 but it implements the standard... 15:16:42 That doesn't sound like something there would be a function for. 15:16:55 alise, you mean like find all messages containing a substring? Or like total number of times that substring appears (counting duplicates in a given message as several, instead of one) 15:16:55 ? 15:17:24 just grep your dir :) 15:17:25 I assumed the latter; the first is just like/glob. 15:17:25 alise, the former is easy, the latter I have no clue how to do 15:17:37 Vorpal: "abbbbacca" on "a" -> 3 15:17:41 yeah the first is body like '%foo%' 15:17:44 the latter hm... 15:17:44 yeah 15:17:52 sqlite> select count(*) as n, replace(replace(replace(replace(replace(replace(replace(replace(replace(nick,"scarf","ais523"),"iEhird","alise"),"aliseiphone","alise"),"ehirdiphone","alise"),"alise`","alise"),"ehird`","alise"),"AnMaster","Vorpal"),"tusho","alise"),"ehird","alise") as nickx from logs where type = 0 and body like "%fuck%" group by nickx order by n desc limit 10; 15:17:52 3806|alise 15:17:52 1250|pikhq 15:17:52 735|oklopol 15:17:56 not an accurate measure, but 15:18:15 i'm pretty sure i'll come on top for fucking in the end. pun not intended yet no effort was taken to avoid. 15:18:19 alise, you could perhaps output the result lines instead and then grep that? 15:18:25 Vorpal: xD 15:18:45 alise, I'm pretty sure that sql fans would claim doing this in sql would be against some normalisation principle or such :P 15:18:46 Vorpal: i could just grep :P 15:18:49 Explode all possible substrings of each line into another table, then it's just a count there. 15:18:55 fizzie, haha 15:18:57 fizzie: I'll get right on it! 15:19:03 fizzie, how many GB? 15:19:14 Very many, is my guess. 15:19:14 369|instead of ">", "I fucked your mother last night"|lament 15:19:21 selecting text there too produces a wonderful flavour-chart of fucks 15:19:27 302|but I /think/ that's a working dtou script in Brainfuck|ais523 15:19:30 how predictable 15:19:35 hmm, does like have a word boundary thing? 15:19:44 alise, personally I would create a view for that replace madness 15:19:52 body like "fuck%" or body like "% fuck%" 15:19:52 there 15:20:01 Vorpal: i'm too lazy :D 15:20:08 yay now ais523 is no longer on it 15:20:11 that filthy non-curser 15:20:31 Vorpal: grepping the output means i might as well just grep my hardcopy logs 15:20:37 but that doesn't order it for me 15:21:03 alise, well depends on exactly what you want to do. I mean, you aren't supposed to write the entire client program in SQL generally :P 15:21:33 alise, from postgres I get: 15:21:35 ERROR: column "scarf" does not exist 15:21:36 LINE 1: ...lace(replace(replace(replace(replace(replace(nick,"scarf","a... 15:21:36 ^ 15:21:44 lol postgre fails 15:21:55 alise, sure this is completely standard? 15:22:00 oh wait 15:22:05 I think I'm in the wrong db 15:22:06 maybe 15:22:09 it's standard. 15:22:17 maybe nesting so much isn't :) 15:22:24 Vorpal: and pah, with LISPABASE it's easy 15:22:35 hm 15:22:38 you write Lisp-datalog queries seamlessly integrated with Lisp! 15:22:48 alise, link to this wonderful thing 15:23:40 alise: Incidentally, PostgreSQL can do regexp_replace() (and regexp_split_to_table, and then you can get the table length for count) to text. 15:23:46 sorry, ~/mind is chmodded solely to me 15:23:58 fizzie: POSTGRES POSTGRES POSTGRES 15:24:01 I'M TIRED OF HEARING ABOUT POSTGRES 15:24:06 *snifff* 15:24:14 alise, actually it is postgres doing it right 15:24:24 NOTHING'S RIIIIGHT 15:24:24 alise, it implements " as the sql standard says I think 15:24:30 which is how? 15:24:31 you want ' around strings 15:24:35 whatever 15:24:45 alise, " quotes a column or table in case it has the name of a key word 15:25:01 -!- Flonk has joined. 15:25:57 ./07.07.31:14:46:15 Errno::ENOENT: (eval):1:in `initialize': No such 15:25:57 file or directory - BIG GAPING HELLO.JPG-STYLE SECURITY HOLE 15:32:09 -!- olsner has joined. 15:32:28 Why is it that the best applications demand reinventing everything first? 15:33:41 I need someone to poke with questions about a Prolog database in Lisp. :| 15:33:47 alise, http://sprunge.us/bfQW 15:33:59 thuso 15:34:02 oops 15:34:15 so anyway who wins 15:34:22 alise, was it "tusho"? 15:34:23 er wait 15:34:25 Vorpal: yes 15:34:27 yours is crap 15:34:30 it doesn't do the useful thing 15:34:34 i.e. count individual fucks 15:34:55 Vorpal: does postgres have an eval("1+1+1+1") => 4? 15:35:07 if so, regex replace every non-fuck with nothing, then regex replace every fuck with 1+ 15:35:08 append 0 15:35:10 and eval() 15:35:11 to get fuckcount 15:35:20 alise: SICP includes a prolog-in-scheme, doesn't it? 15:35:24 alise, no function with the name eval to do that at least 15:35:37 olsner: it's more like datalog here though, with some sort of efficient algorithm (magic sets?) 15:35:41 and database-oriented 15:35:42 Vorpal: bah 15:35:50 Vorpal: use pl/sql :P 15:36:01 alise, yeah but I would have to learn pl/sql first :P 15:36:07 Vorpal: then you can just use regex to make afuckbcfuck -> xx for the two fucks 15:36:09 then run a procedure on it 15:36:11 that j... 15:36:11 wait 15:36:17 length(replace fuck with "x" and remove others) 15:36:23 something like that yeah 15:36:26 ... 15:36:30 can Postgres handle \0 in a string? 15:36:34 if so, since \0 can't be said on IRC, 15:36:45 alise: With postgres, you just need to regex-split all lines with "fuck" into an array, then take array-length minus 1 as the fuck-count. 15:36:49 alise, I assume it can, but maybe the client can't 15:37:01 alise: oh, ok... datalog seems to allow doing much more complicated and clever stuff than prolog, apparently 15:37:14 Vorpal: length(regex_replace(regex_replace(body,'fuck','\0'), '\0*', '')) 15:37:15 or something 15:37:25 olsner: yes, since it's sub-TC 15:37:36 olsner: but basically i want to ask questions about the lisp-facing api 15:38:08 alise, anyway you easily win in number of times saying "fuck" it seems. With (body ilike 'fuck%' or body ilike ' fuck%') you get 551 lines? Seems low hm 15:38:41 ilike for case insensitive 15:38:55 ARC NEWS: arc is still on version 3, exactly the same people still care 15:39:04 Vorpal: er, i got much more than that. 15:39:18 sqlite> select count(*) as n, replace(replace(replace(replace(replace(replace(replace(replace(replace(nick,"scarf","ais523"),"iEhird","alise"),"aliseiphone","alise"),"ehirdiphone","alise"),"alise`","alise"),"ehird`","alise"),"AnMaster","Vorpal"),"tusho","alise"),"ehird","alise") as nickx from logs where type = 0 and body like "%fuck%" group by nickx order by n desc limit 10; 15:39:18 3806|alise 15:39:24 not much worse with your modification i'd wager 15:39:25 oh wait 15:39:28 I forgot a % 15:39:33 -!- zzo38 has joined. 15:39:33 Vorpal: can you try that regexp thing for me? 15:39:37 i dunno if \0 will escape properly 15:39:38 but worth a try 15:39:58 alise, with (body ilike 'fuck%' or body ilike '% fuck%') you get 3092 lines of having said that work 15:40:00 word* 15:40:04 and hm 15:40:13 Vorpal: what went wrong the first time, then? 15:40:27 alise, ' fuck%' != '% fuck%' 15:40:28 :P 15:40:48 alise, simply a typo in other words 15:40:56 alise, anyway what regex? 15:41:14 Vorpal, alise: here's a fuck-counter in postgres: 15:41:15 fis=> select length(regexp_replace(regexp_replace('foofuck, bar fuck baz, fuck fuck and so on', 'fuck', E'\n', 'g'), E'[^\n]', '', 'g')); 15:41:15 length 15:41:15 -------- 15:41:15 4 15:41:25 Works for all strings with no embedded newlines, which is all IRC messages. 15:41:31 hm 15:41:49 (Replaces all fucks with newlines, then all non-newline characters with nothing, then counts the string length.) 15:42:12 Possibly \0 would work too. 15:42:40 It needs the E'...' string for strings-with-escapes, though. 15:42:47 fizzie: Yeah, that's basically what I said. 15:43:00 You could do that in sqlite with an unholy nesting of 254 replaces. 15:43:05 ERROR: column "logs.body" must appear in the GROUP BY clause or be used in an aggregate function 15:43:05 LINE 5: length(regexp_replace(regexp_replace(body, 'fuck', E'... 15:43:05 ^ 15:43:06 huh 15:43:10 And maybe embedded control characters. 15:43:14 that's stupid 15:43:30 did you have body like fuck anywhere? 15:43:33 remove that obviously 15:43:34 Vorpal: Obviously sum() that if it's a group-by thing. 15:43:38 hm 15:43:39 or that 15:43:42 fizzie, ah yes 15:43:49 "as fucks" 15:43:58 sum(length(...)) as fucks, order by fucks... 15:44:28 "query is running"... 15:44:33 not a fast one 15:45:06 hm first filtering on body ought to help 15:45:14 hm that seems very wrong 15:45:24 you have fewer "fuck" than number of lines with it 15:46:26 alise, 3521 fucks, and 3806 lines with that in it. Something is off here 15:46:54 try replacing \bfuck\b 15:46:55 fizzie, what does the E in front of those strings do? 15:46:59 otherwise "brainfuck" might weird it 15:47:01 Vorpal: escaped string 15:47:02 he said 15:47:11 ah missed that 15:47:50 alise, anyway I set it to do all such substrings, without regard to where in the string 15:48:07 which is the issue 15:48:09 since "brainfuck" 15:48:13 and it still ends up as fewer occurrences of the word than number of lines 15:48:14 *because of "brainfuck" 15:48:28 alise, yes but the line thing I compared to would include that too 15:48:38 both would 15:48:43 with current query 15:50:10 actually 4098 lines vs. 3521 occurrences of the word 15:50:14 something is definitely off 15:50:21 [[At Lulu's suggestion I have changed the rating to "teen" so it is now possible to search for and buy Prime Intellect when you are logged in at the default settings. I'm not sure whether it's the zombie sex or the incest scene which is most teen-oriented, but people have complained about the rating so if the publisher doesn't think it needs the shrink wrap, who am I to argue?]] --Roger Williams 15:50:37 Vorpal: like is case-insensitive, while the regex only counts completely lowercase fucks; that's one possible reason. 15:50:46 aha 15:50:49 yeah put 'i' as the last argument? 15:50:49 yeah 15:50:53 and make sure to use \bfuck\b 15:51:05 Yeah, changing the 'g' to 'gi' would help. 15:51:13 fizzie, yep it did 15:51:34 alise, is \b really posix regexp? 15:51:38 thought it was pcre? 15:51:49 eh 15:51:50 alise, and this thing is posix extended according to docs 15:51:52 (^| ) 15:51:55 ( |$) 15:51:56 job done 15:52:08 indeed 15:52:28 bind '"\eOR": " \C-ucd */\e*\n"' 15:52:52 alise, you end up at 1336 then, probably due to excluding stuff like "fucking" 15:53:02 Vorpal: It's actually posix-extended with some non-POSIX additions, but I don't think \b is one of them. \y is "matches at the beginning or end of word", though you'd probably need \\y in a E'...'. 15:53:20 * Vorpal tries (^| )fuck(|ing|er|ed)( |$) 15:53:26 Vorpal: fuck(er|ed|wit|ing)? 15:53:34 alise, will do in a sec 15:53:43 with the one I just said: 2663 15:53:47 (mother|)fuck(|er|ed|wit|ing) 15:54:09 i guess by the carlin measuring post motherfucker != fucker 15:54:15 alise, the last one: 2674 15:54:31 soupdragon at 806 as number 2 15:54:36 hmm 15:54:41 i wonder what pushed pikhq off 15:54:47 pikhq fucks a lot. uhh... we need new terminology 15:54:48 then followed by oklopol at 419, and pikhq at 408 15:54:54 Vorpal: elide (mother|) 15:54:56 then me at 175 and augur at 120 15:55:23 alise, right. lets see: same ordering in the top 6 15:55:48 At FreeGeek, I have an account there now. I managed to make a shell function to select the next subdirectory of the parent directory of the current directory, it uses a pipeline with 'pwd' and 'cat' and 'cd' and 'ls' and 'sed' 15:55:51 I forgot what it was for the remaining of the top 10 compared to last query, and I'm running this in a GUI tool for easier editing 15:55:57 (pgadmin3) 15:56:01 (awesome tool) 15:56:09 "Can you feel yourself tripping ALL the balls??" "W-" "We're leaving no ball left untripped here, Dromiceiomimus." --Dinosaur Comics 15:56:59 -!- EOF has quit (Ping timeout: 252 seconds). 15:58:56 06:34:19 Because there's one instance where there's just "* EgoBot" in the logfile. 15:59:00 the explanation 15:59:04 ah, i see how that works 15:59:10 ^AACTION^A 15:59:10 or 15:59:11 ^AACTION ^A 15:59:14 length 0, isn't it? 15:59:17 length 0 /me 15:59:34 -!- abcde has joined. 15:59:40 * abcde 15:59:44 * abcde 15:59:52 The latter was ^AACTION ^A. 15:59:58 ^AACTION ^A 16:00:03 More 0-length messages! 16:00:04 -!- abcde has changed nick to Guest92520. 16:00:10 I assume they appear as type=0. 16:00:11 -!- Guest92520 has quit (Remote host closed the connection). 16:00:12 Perhaps not. 16:01:04 ah created that view 16:01:11 makes the whole thing a lot less horrible-looking 16:01:20 moving all the replacing to the view 16:02:13 the definition for the view is http://sprunge.us/BTdI (note: I didn't enter it like that, that is what I get back when asking postgres for the sql code for it!) 16:05:29 of course that view is useless for nick changes 16:05:33 but lets do some kick stats on it 16:07:08 lament kicked most people, followed by oerjan and fizzie 16:07:29 most often kicked was alise followed by dixon (who was that?) 16:07:43 dixon is Quadr*scence's trolling buddy 16:07:43 alise was kicked 6 times 16:07:45 from the Day-Long Troll 16:08:13 OF YORE 16:08:18 wait, dixon, asiebot, oklopol, Quadr* and bsmnt_bot share second place 16:08:27 at 2 kicks each 16:08:30 compared to your 6 times alise 16:08:36 alise, I guess you are just hated :P 16:08:44 actually 16:08:46 most of those would be one of 16:08:52 oh? 16:08:58 - when i asked to be kicked until i was kicked 16:09:03 (I have no idea why) 16:09:05 and possibly 16:09:08 - accidental floods 16:09:19 i don't think i've been kicked for actually being an awful person before 16:09:27 select the kick messages 16:10:08 alise, you were kicked twice by lament, once by fizzie and three times by oerjan 16:10:22 Vorpal: select the messages if you can 16:10:37 two of the oerjan kicks were from yesterday if you've updated since then 16:10:40 What TeX format should I use for binary numbers in Enhanced CWEB? 16:10:41 alise, well, I would need to select on timestamps around those kicks or such 16:10:49 or do you mean kick message? 16:10:51 (when i tried to paste the unary '.', and then again when I forgot to /flushq before rejoining) 16:10:52 Vorpal: the latter, yes 16:10:56 zzo38: any. 16:11:42 alise, http://sprunge.us/dDJd 16:11:52 alise: I know I could use any, but I meant one format that can be used specific for binary numbers, like how it formats octal and hexadecimal numbers in their special format, then what format should be used for binary numbers (0b00110110011)? 16:12:09 the two lament ones i asked for, i think 16:12:10 ehh 16:12:12 i'll just look i up 16:12:18 Vorpal: select tusho and ehird too :P 16:12:22 alise, you have the UTC timestamp 16:12:30 alise, that is logs_na, the nick normalised view 16:12:31 oh 16:12:33 i was thinking 16:12:35 i didn't use alise in 2008 :P 16:12:47 alise, why would I create a view and not use it? 16:13:23 alise, btw I found that sqlite is way faster when doing a full import of all the data, but way slower to do complex queries that have to scan the table (as opposed to using indexes) 16:13:38 I think almost everything on logs_na require full table scan 16:13:38 15:14:46 im too fucking badass to get kickbanned motherfuckas! 16:13:52 15:15:34 15:24:49 one tyme my profesor did dat 2 and i was like holy shat i got like 4 routersa that are MIPS 16:13:52 15:15:34 15:30:20 augur: im lookin fo sometin i can drink durin class i dun rly wanna pop any pills durin class 16:13:52 15:15:35 15:30:26 campus securiy canna b liek waddat 16:13:52 15:15:35 15:31:15 BUT I ALSO WANT TO FUCKIHN BURP IN MY PROFESSORS FACE AND IT NEEDS TO SMELL LIEK ENERGY 16:13:52 not sure if postgres supports materialised views 16:13:54 15:15:35 15:34:30 randall is betta dan spinelli olol 16:14:39 alise, what? 16:14:48 Vorpal: lines from that day 16:14:50 ah 16:14:52 it appears this Phenax person was a little insane 16:14:59 alise, "spinelli"? 16:15:01 some talk about a kickban of me, though, so it might have been the day before 16:15:04 Vorpal: who knows 16:15:54 alise, hm the dates are normalised to UTC 16:16:03 you need to correct for that, could cross log dates 16:16:13 yeah 16:16:18 see with botte... 16:16:31 alise, well the thing is, if you use the db you don't need to 16:16:43 you can just apply a timezone to the stuff 16:16:47 i know 16:16:50 the issue is when going back to the clog text log files 16:17:00 15:47:19 "The GNU Compiler Collection. Includes C/C++, java compilers, pie+ssp extensions, Haj Ten Brugge runtime bounds checking" is about max lenght 16:17:03 what a hopeless description! 16:17:08 alise, this is easier in postgres since the timestamps are of the type timestamp 16:17:13 rather than just textual data 16:17:18 "Compilers. Includes some languages, flargle bop doogit, and asiudha4w98y." 16:17:35 15:47:19 "The GNU Compiler Collection. Includes C/C++, java compilers, pie+ssp extensions, Haj Ten Brugge runtime bounds checking" is about max lenght <-- max length of what? 16:17:50 portage description field 16:17:51 i think 16:17:55 ah could be 16:17:58 yeah 16:18:01 what with the pie+ssp stuff 16:18:58 15:49:07 --- mode: ChanServ set +o lament 16:18:58 15:49:16 tusho: what's a good hostmask? 16:18:58 15:49:22 lament: for whom? 16:18:58 15:49:38 *!*@67.15.72.46 ? 16:19:05 Vorpal: sqlite can do time maths with the textual data (that's why I put it in there in that format); it can even assume UTC and convert it to local times for displaying. What it doesn't do is explicit timezone support, so you can't convert it into an arbitrary timezone within sqlite. (Again something you're supposed to do in the client app side.) 16:19:06 this is in 08.06.18 leading up to the kick 16:19:22 yeah then me and oklopol ask to be kickbanned 16:19:45 15:51:26 okay 16:19:45 15:51:30 --- mode: lament set +b *!*n=tusho@91.105.124.* 16:19:45 15:51:31 --- quit: Slereah_ (Read error: 104 (Connection reset by peer)) 16:19:45 15:51:34 --- kick: tusho was kicked by lament (lament) 16:19:45 15:51:37 :D 16:19:46 15:51:46 was fun knowing him 16:19:48 15:52:19 o.o; 16:19:50 15:53:12 lament: i don't think he wanted a permanent ban :P 16:19:52 15:53:43 well, we'll never know 16:19:54 15:53:49 :D 16:19:56 Immediately after: 16:19:58 15:54:16 lament, please please unban it 16:20:00 UNBAN IT 16:20:07 fizzie, postgres can do timezones, oh and it's timestamps take less space than the textual data of sqlite 16:20:24 gahahaha Phenax still isn't kicked by this point 16:20:33 alise, how stupid I was, asking him to unban you 16:20:40 Unban it, you mean. 16:20:42 Vorpal: Sure, that was just a rationale for why they're in the format they are, and that sqlite can in fact do stuff with them. 16:20:56 alise, I think you had previously claimed to be genderless or something silly like that 16:20:59 a few days before or such 16:21:05 i don't think so 16:21:15 could be wrong though 16:21:16 since augur immediately corrected you 16:21:20 15:55:41 --- join: someguy (n=ehird@91.105.124.212) joined #esoteric 16:21:20 15:55:53 you guys just banned my brother? 16:21:20 15:55:58 fucktards 16:21:20 15:56:10 someguy: big/little brother? 16:21:20 15:56:19 oklopol: i'm like 23, I lose count 16:21:29 alise, hm 16:21:34 so hard to keep track of your other personality's age 16:21:44 15:56:28 lament, you *could* have kickbanned Phenax 16:21:44 15:56:43 AnMaster: that actually seems like a good idea. 16:22:05 hm 16:22:14 Vorpal: 15:56:52 http://www.bigzaphod.org/taxi/ 15:56:54 HAHAHAHA 15:56:59 at that language 16:22:17 you had no taste 16:22:29 well okay it's a nice esolang but not HAHAHAHA nice 16:22:41 15:57:15 lament: yeah motherfucker unban him or i'll kill you and your kids 16:22:48 such support 16:22:49 I think they redesigned the page 16:22:58 I remember it as way more "low-tech" 16:23:08 15:59:06 lament: unban my brother would you 16:23:09 15:59:14 someguy: why? 16:23:09 15:59:21 lament: because it's midnight, duh. 16:23:42 alise, what was the logic for that 16:23:55 i have no idea 16:24:23 16:08:36 you're all nuts 16:24:23 16:08:42 that's my opinion 16:24:23 16:08:51 am i nuts? 16:24:23 16:08:55 lament, "all"? you mean me too? 16:24:23 16:09:00 No. You're all nuts except oklopol. 16:24:24 16:09:01 I'm not nuts! 16:24:26 16:09:05 !! 16:24:49 16:09:29 why would I be nuts? 16:24:49 16:09:47 AnMaster : Becayse you're dangling between my legs. 16:24:49 16:10:01 Slereah, sod off 16:25:06 mhm 16:25:17 (Epilogue of this little archaeological dig: lament: TAKE IT OFF!! um. Not in public.) 16:25:35 16:12:19 tusho: here's a better idea 16:25:35 16:12:21 --- mode: lament set +o tusho 16:25:35 16:12:25 AWESOME 16:25:35 16:12:25 there, now you can ban yourself 16:25:35 16:12:31 --- mode: tusho set -o lament 16:25:36 16:12:32 --- mode: tusho set +b lament!*@* 16:25:38 16:12:38 --- kick: lament was kicked by tusho (tusho) 16:25:41 Note to everyone: never give me +o. Ever. 16:25:47 indeed 16:26:13 16:13:17 lament's brother lameguy is gonna be so pissed. 16:26:18 this log is amazing 16:26:44 16:17:01 FREE UNBAN TO THE PUBLIC 16:26:44 16:17:04 --- mode: tusho set +b AnMaster!*@* 16:28:33 what 16:29:02 fizzie, there? 16:29:45 fizzie, did you see above wrt having to reimport the data again? (Due to not storing what modes changed in mode changes.) 16:30:00 Vorpal: you were asking for ops so you could spread a new regime of unbannedness to everyone 16:30:13 so i nipped the anti-communist western lies in the bud 16:30:34 16:30:36 hm 16:30:34 16:30:48 after last time, i think i learned not to trust you with ops 16:30:34 16:30:59 --- mode: ChanServ set +o dasf 16:30:34 16:31:00 lament: i was posessed by demons last time 16:30:34 16:31:09 damn straight 16:30:38 -!- Flonk_ has joined. 16:31:33 alise, who was dasf? 16:31:35 me 16:31:38 mhm 16:31:51 i don't know why i got opped again, since i was developing a bot to keep giving me ops forever :) 16:32:12 16:51:02 --- mode: tusho set +o ihope 16:32:13 16:51:07 Yay! :-) 16:32:13 16:51:12 --- mode: ihope set -oo oppiebot tusho 16:32:17 -!- Flonk has quit (Ping timeout: 265 seconds). 16:32:18 Darn rebels. 16:32:20 -!- Flonk_ has changed nick to Flonk. 16:33:20 -!- jix has joined. 16:33:45 -!- jix has quit (Client Quit). 16:34:05 -!- jix has joined. 16:34:25 alise, lets see how many times you have been banned hm 16:35:29 that "Nice try" was a request-ban too 16:35:34 of a sort 16:35:49 i asked for it then tried to change it into a coup when +o happened, but oerjan just ruthlessly killed me anyway 16:35:59 so basically only the second-last kick was for doing something bad (pasting a lot of "1"s) 16:36:06 the last one was when i forgot to /flushq before rejoining 16:36:13 select * from logs where body like '+b%' and (body ilike '%ehird%' or body ilike '%tusho%' or body ilike '%alise%') and type = 9; 16:36:21 7 times 16:36:26 hee 16:36:39 2 by oerjan, one by fizzie and 4 by lament 16:36:52 alise, that query won't work with the db you downloaded from fizzie yesterday btw 16:37:02 alise, since it didn't properly parse mode changes 16:38:00 alise, it isn't possible to actually take over a channel with services, unless someone give you the flags there, or you find out a password or such 16:38:01 * alise installs J 16:38:05 I need a ~/local. 16:38:10 actually, I wonder why people don't use ~/.local/. 16:38:15 alise, cause a mess with +o yes, but takeover? no 16:38:18 Vorpal: i just relied on apathy 16:38:28 and it was for chaos, anyway 16:38:48 alise, also I doubt oerjan killed you. He isn't an oper 16:38:57 In other news, I made a construction to convert an arbitrary Turing machine into sqlite3 triggers, and ran through it manually a 4-symbol 13-state machine that reverses a string. 16:38:58 Witness the elegance of it: http://p.zem.fi/sqlite-tm.sql 16:39:00 kicked perhaps 16:39:18 The best way to reverse a string since sliced bread. 16:39:19 fizzie, haha 16:39:44 fizzie, does it work on any string length or just that length? 16:39:56 Vorpal: killed in the literalmetaphorical sense 16:40:14 fizzie, could it reverse bcaa for example? 16:40:30 alise, ah 16:40:30 Vorpal: Any string, though you'll probably hit the maximum trigger nesting depth (1000) pretty easily. 16:40:48 fizzie: So SQLite is turing-complete? Awesome! 16:40:50 Vorpal: (Also, the manual SELECT statements at the end only print the first six tape values.) 16:40:58 Vorpal: Any string, though you'll probably hit the maximum trigger nesting depth (1000) pretty easily. fizzie: So SQLite is turing-complete? Awesome! <-- um 16:41:05 Vorpal: and Python has a nesting stack limit 16:41:10 doesn't mean the language itself isn't TC 16:41:14 in that way 16:41:15 hm 16:41:46 alise, actually I think it may have documented limits on upper size of db and such 16:41:57 implementation defects 16:41:58 of course if that is specced or documented 16:42:03 is a tricky question 16:42:16 meh, grammar fail 16:42:52 Vorpal: why don't people use ~/.local? 16:42:55 as opposed to ~/lcoal 16:42:56 *local 16:43:01 alise, for what? 16:43:06 Indeed I already have a ~/.local/share... 16:43:12 For some reason it actually worked without recursive_triggers enabled, but only for exactly 16 steps; don't know what's up with that. 16:43:13 Vorpal: for the unified-tree version of ~/bin etc. 16:43:23 ehird@dinky:~$ ls .local/share/ 16:43:23 applications desktop-directories gvfs-metadata rhythmbox totem ubuntuone 16:43:23 desktop-couch Empathy pitivi tomboy Trash webkit 16:43:27 So there's precedent... 16:43:35 alise, I use local for ~/local/llvm/2.7/{bin,lib} and so on 16:43:48 stuff like that 16:44:02 (It also takes around six seconds to reverse "abc", so...) 16:44:07 yes, but that clutters ls ~ 16:44:13 so why do I never see ~/.local? 16:44:16 apart from... just now 16:44:41 alise, you never see it because presumably you don't use ls -a 16:44:59 (or ls -A) 16:45:04 so? 16:45:08 uh 16:45:10 you totally misinterpreted 16:45:11 i mean 16:45:14 why do you never see people using it 16:45:16 as opposed to ~/local 16:45:16 ... 16:45:28 alise, yes I know, I selected the literal interpretation :P 16:45:42 I hereby reask the question to fizzie. 16:45:58 alise, and I don't have an answer to why it isn't used 16:46:29 Vorpal: what i mean is do you have any personal arguments against it 16:46:50 alise: I don't know; for .local in particular, ~/.local/share is already used by a big pile of applications for user-specific stuff. I keep a ls-visible ~/local/ myself too; I guess I somehow wouldn't want to "hide" something like that, but I don't really have a rational argument here. 16:47:19 I guess so. 16:47:27 It's just that I like to have ~ be non-system stuff, you know? 16:47:48 watasi no hôhầkurahuto ha unakì tè i'hąi tèsu. 16:47:50 >_> 16:48:00 anyway if it isn't split in projects it would be rather irritating to upgrade anything with more than one or two files installed in there 16:48:33 alise, ~/local is not system stuff, it's your own stuff 16:48:43 as in, not done with package manager 16:48:56 it's still linux stuff 16:49:00 operating system stuff 16:49:17 i'm sure everyone else can understand the distinction 16:49:19 nah it's user land, so maybe gnu/linux stuff :P 16:49:36 and I know what you meant 16:49:47 pikhq: tèsu tèsu tèsu tèsu tèsu tèsu tèsu tèsu tèsu tèsu tèsu 16:49:57 alise: "My hovercraft is full of eels". 16:49:59 augh 16:50:09 My hovercraft is full of desu. 16:50:33 pikhq: Quick, translate "my cousin is a meerkat of strange angles". 16:50:52 That would be 私のホーバークラフトはですで一杯です。 (watashi no hōbākurafuto wa desu de ippai desu) 16:52:11 Uh oh; for QI to complete, Stephen Fry will have to continue hosting it until he's 72. 16:52:12 私の従兄弟は変な角のミーアキャットです。 16:52:17 God I hope the man has dedication. 16:52:26 pikhq: Is that the meerkat one? I guess not. 16:52:34 "Watashi no itoko ha hen na kado no miiakyatto desu." 16:52:43 I guess so! 16:52:52 * alise tries to find the relevant clip 16:53:15 It doesn't help that I don't know even what series it's from. 16:53:16 (watasi no itoko ha henn na katò no mîakiȳa'to tèsu.) 16:53:39 aha 16:53:46 I think I know 16:53:55 * alise finds the bit 16:54:15 Load quicker, video! 16:55:34 alise, QI? 16:56:05 Vorpal: a hilarious quiz show where all the questions have absolutely obvious and completely wrong answers, and points are usually awarded simply for being "quite interesting". 16:56:24 Penalties are "awarded", complete with flashing lights and a siren, for giving obvious answers. 16:56:24 alise, what about the one about about your aunt? Another such common translation trope. 16:56:28 Hosted by Stephen Fry. 16:56:34 Usually the majority of scores end in the negatives. Especially Alan Davies'. 16:56:35 Well now, *that* is interesting-looking: behold the cumulative since-2004 activity in logscale: http://zem.fi/~fis/actl.png 16:56:36 alise, ah. 16:56:50 alise, and what is that about "complete"? 16:56:54 "My cousin is a meerkat of strange angles" is a quote from an episode that also has the hovercraft quote, so I'm trying to find it. 16:57:01 Vorpal: Each series covers topics starting with one letter. 16:57:05 Series A, B, C, etc. 16:57:13 One series per year, ergo... 16:57:19 ah 16:57:22 I see 16:57:28 We've just started season H. 16:57:35 fizzie, wtf? 16:57:48 fizzie, what's with the stripes 16:57:54 alise: I think you could make it sound completely awesome while describing exactly one bit. "Hosted by Stephen Fry". 16:58:22 pikhq: Think I've almost found the time... 16:58:26 fizzie, or the blue parts that don't touch zero 16:58:30 Vorpal: Ask gnuplot, not me: all I did was "set logscale y". I guess filledcurves + zeros (-inf in logscale) don't really work so well. 16:58:39 fizzie, haha 16:59:18 I do like how it's gone outside the graph borders, though. 16:59:20 fizzie, try setting them to 0.000001 or such and see if it helps 16:59:39 fizzie, I would suggest running valgrind on that thing, just in case 17:00:27 No errors from valgrind (despite a few "definitely lost" bytes); trying the non-zeroising now. 17:01:07 It does lines with zeros in them just fine in logscale (puts in vertical lines that go below the chart (but are clipped at the bottom) and then leaves gaps), but filledcurves, no so much. 17:01:26 fizzie, and the non-zeroizing did not help? 17:01:34 Haven't tried yet. 17:01:46 http://zem.fi/~fis/actll.png is the same data but "with lines". 17:02:12 fizzie, I suggest storing the crazy graph rather than overwriting it (maybe renaming it to actl-crazy.png or such 17:02:24 pikhq: http://vimeo.com/7868344; might as well start watching at 18:00. The actual exchange is at at 21:11, but it's out of context, so. 17:02:27 pikhq: I did all that work for you! 17:02:35 All for strangely-angled meerkats. 17:03:32 alise: Vimeo does not let you skip beyond where has been downloaded. 17:03:40 pikhq: Indeed. I suggest utilising the pause button. 17:03:43 It buffers quite quickly. 17:03:53 Vorpal: Yeah, it's about the zeros; http://zem.fi/~fis/actlf.png where zeros were replaced with 0.01 is a normal, boring sort of plot. 17:03:54 Or, just watch it all. :P 17:03:54 Which is what I'm doing. 17:03:56 alise, on pikhq's connection? 17:04:03 Vorpal: he's in a metropolitan area now. 17:04:08 Vorpal: Reasonable broadband. 17:04:13 fizzie, so what does it show? 17:04:14 pikhq, ah nice 17:04:18 Better broadband than mine, I think. 17:05:23 fizzie, is it number of lines but logarithmic? 17:05:27 Vorpal: Just the usual cumulative activity; the top shape is number of messages/month centered at that point of time. Since it's logscale, though, you can't really read the relative activity of different people, so the colors are a bit useless. 17:05:33 The point is, though: it doesn't suck. 17:05:54 fizzie, yeah those down near the bottom get most space 17:06:50 fizzie, try inventing a plot that you can do as a 3D plot with colours to indicate a forth value on the surface 17:07:03 need to be reasonably flat to be usable as a static image hm 17:07:19 if there are dips it would be hard/impossible to see all 17:12:12 fizzie, can you think of a query to check that timestamp always goes up when the serial does? 17:12:18 to check for jump backwards 17:12:29 jumps* 17:13:10 fizzie, I can't think of one that does not involve procedural code 17:13:16 though it would be easy with that 17:13:19 Oh, that should be doable; just a moment. 17:13:49 For the reference, I named my row-id column "idx"; you probably used something else. 17:14:08 fizzie, I used serial 17:14:19 * alise starts reading Fine Structure 17:14:23 Finally! 17:14:27 Finally! 17:15:33 Vorpal: Funny, my db does in fact have some backwards jumps. 17:15:47 I still can't decide where J should go! 17:15:49 Vorpal: select l1.idx, l1.tstamp, l2.idx, l2.tstamp from logs l1 join logs l2 on l2.idx = l1.idx+1 where l1.tstamp > l2.tstamp limit 10; -- just s/idx/serial/ and maybe remove the limit if you want them all. 17:15:56 188904|2006-10-29 09:01:03|188905|2006-10-29 08:12:58 17:15:56 1202220|2009-07-30 22:10:01|1202221|2009-07-30 22:09:40 17:15:59 I have those two, and more. 17:16:00 fizzie, what is the query 17:16:05 ah 17:16:07 there it is 17:16:19 fizzie, l2? 17:16:30 l1 and l2, just arbitrary names. 17:16:44 It just joins the logs table with itself on the condition that left.idx = right.idx+1. 17:17:32 right 17:17:36 So it only checks all such pairs where serial1 = serial2+1, i.e. two successive rows. 17:17:46 fizzie, assuming serial has no gaps that works 17:18:12 You could easily cross join logs with logs to check all pairs, but that'd involve N^2 lines in the intermediate join, which might be a bit problematic. 17:19:01 hm 17:19:57 fizzie, all the months looks like dst ones 17:20:19 well almost 17:20:29 the 2009-07-31 one looks strange 17:20:56 but most are around oct/nov broder 17:20:58 border* 17:21:07 There's a DST-looking backwards jump in http://tunes.org/~nef/logs/esoteric/09.10.17 -- it might be that clog's clock and pytz's rules don't exactly agree on when the DST switch happens. 17:21:43 * Vorpal adds order by l1.serial 17:21:47 pytz ignores the manual "ooh, clock jumped back, handle this properly" logic if it thinks the time specified is not ambiguous, i.e. is outside that one hour when DST switching is legal. 17:22:04 fizzie, hm 17:22:26 I guess it's even borderline possible that the clog machine's been switched from DST to non-DST manually at approximately the right time. 17:22:43 fizzie, do you pass it the previous timestamp so it can compare? or? 17:23:03 ~/local/j64-602? ~/local/j64? ~/local/j? ~/.local/j64-602? ~/local/j64? ~/local/j? ~/j64-602? ~/j64? 17:23:13 For updatedb.py? Yes, it does notice if the clock jumps backwards in two messages. 17:23:36 alise, ~/local/j64/602? assuming 64 is not version number 17:23:48 if part of version number ~/local/j/64-602 might work 17:24:05 64 signifies 64-bit 17:24:17 j32/j64, although on 32-bit i think it's just called j602 17:24:17 (Away for some light entertainment now; back in ... three hours? Maybe four. Something like that.) 17:26:39 fizzie, a few of those look like they could be ntp step style adjustments 17:30:28 Vorpal: any other suggestions in light of that new knowledge? :P 17:30:46 j hasn't updated beyond 602 in... well, ages. i've never known it as anything other than 602 17:33:17 alise, well, do you plan to install both versions side by side? 17:33:38 no. 17:33:54 alise, and if no updates are to be expected I guess subdir for different versions wouldn't be needed. So ~/local/j might be fine 17:34:07 But it's so short! 17:34:15 alise, isn't that an advantage? 17:34:34 Well, it just seems so wrong. 17:35:07 alise, mhm then why not ~/local/languages/array/j/x86-64/602 :P 17:35:14 >_< 17:35:16 I just mean the /j/ part. 17:35:26 alise, I know, I was joking.... 17:36:49 hm should start importing irc client logs 17:36:57 same basic idea should work 17:37:45 since most were logged in client rather than bouncer I think I need to somehow mask non-channel stuff that showed up in the same tab 17:38:36 also hm, should probably not ignore logging started/ended stuff in clog apart at the start and end of the file 17:38:48 would make spotting gaps due to malfunction easier 17:40:37 oh hm 17:41:06 I found CINT could be used for C interpreter, but CINT is not C. CINT is C++ instead. 17:43:39 Although how PicoC works the parser has to wait for input sometimes, maybe I can work around 17:44:52 fun I need some aug -> 08 mapping 17:46:51 | ?- assertz((path(X,Y):-edge(X,Y))). 17:46:51 yes 17:46:51 | ?- assertz((path(X,Y):-edge(X,Z),path(Z,Y))). 17:46:51 yes 17:46:51 | ?- assertz((edge(a,b))). assertz((edge(b,c))). assertz((edge(b,e))). assertz((edge(c,d))). assertz((edge(d,c))). assertz((edge(d,a))). assertz((edge(d,e))). assertz((edge(e,f))). 17:46:53 (lots of yeses) 17:46:55 | ?- path(a,f). 17:46:57 Fatal Error: local stack overflow (size: 8192 Kb, environment variable used: LOCALSZ) 17:46:59 Useless! 17:48:08 and year is not on the lines hm 17:48:55 Even does it with a ! in the middle of that second path clause. 17:52:30 * Vorpal changes his date format for personal logs 17:52:43 now I still need to parse the old hm 17:53:03 "Newton OS running on an iPad" 17:53:10 brb, buying an iPad. 17:53:39 Oh my god it can even run at the full iPad resolution. 17:54:10 Vorpal: Duude ^ 17:55:04 mhm 17:55:08 http://www.panic.com/blog/2010/09/newton-never-dies/ 17:57:33 alise, what does r in front of a python string mean? like r'foo' 17:58:18 no escaping 17:58:19 used for regexps 17:58:27 r'\n\(x' etc. 17:58:55 ah 18:03:32 -!- trinithis has quit (Quit: Leaving). 18:04:56 alise, python only seems to document {m,n} not {n}, does it not support the latter? 18:05:45 -!- trinithis has joined. 18:06:02 Vorpal: it probably does. 18:06:03 try it. 18:06:15 yeah will shortly 18:06:49 * Vorpal wonders if "if y is not None" will work.... 18:07:15 huh seems so 18:08:08 alise, wait, does it treat "is not" as one keyword but "not" as another? 18:08:21 Yes. 18:08:53 >>> not 1 18:08:53 False 18:08:53 >>> True is not 1 18:08:53 True 18:08:53 >>> True is (not 1) 18:08:54 False 18:09:02 I've only just realised how fucked up that is. 18:10:49 That's what she said!!! 18:11:01 alise, you mean you never thought about this before? 18:12:03 alise, it is like SQL. "foo not in (list, of, values)" doing the same thing as "not foo in (list, of, values)" 18:12:10 and it's fucked up syntax with not 18:12:51 also python's lack of switch, how did you index a dict now again 18:13:40 ah standard syntax 18:14:05 python is fucked up, period 18:14:34 nooga, indeed but I already had code for doing almost what I wanted in it 18:14:42 trinithis: Die. 18:14:49 also python's lack of switch, how did you index a dict now again 18:14:50 x[y] 18:14:55 or x.get(y, default) 18:14:57 alise: roll a 1 thru 6 please 18:15:00 alise, hm 18:15:02 if/elseif/else is generally used though 18:15:05 rather than a dictionary of functions 18:15:14 alise, well I need to look up value here 18:15:20 a mapping from month name to number 18:16:10 Then yeah, a dictionary. 18:24:15 You could parse the timestamps with strptime, but that's probably not significantly less ugly than regex + manual mappings, and might have some locale-derived breakage. 18:25:59 -!- myndzi\ has joined. 18:26:05 Vorpal: also needs estoppel -> ehird 18:26:07 erm -> alise 18:26:09 i used that for a while 18:27:03 mhm 18:27:33 sqlite> select count(*) as n, replace(replace(replace(replace(replace(replace(replace(replace(replace(replace(replace(nick,"estoppel","alise"),"alise_","alise"),"scarf","ais523"),"iEhird","alise"),"aliseiphone","alise"),"ehirdiphone","alise"),"alise`","alise"),"ehird`","alise"),"AnMaster","Vorpal"),"tusho","alise"),"ehird","alise") as nickx from logs where type = 0 and body = "fuck" or body = "Fuck" or body="FUCK" or body="fuck!" or body="Fuck!" or body="FU 18:27:33 CK!" group by nickx order by n desc; 18:27:33 45|alise 18:27:33 20|oklopol 18:27:35 5|lament 18:27:37 4|Slereah_ 18:27:39 fizzie, and yes I have locale issues. Plus old ones have year in "begin logging" line but not on each line 18:28:22 I have a pile of logs that don't even have the day/month in timestamps, those are pretty horrible to handle. There's usually a "day changed" line at midnight, but still. 18:28:23 fizzie, I now switched to an easier to parse format for the future. 18:28:33 fizzie, I use a global current_year here 18:28:45 new one has UTC offset at the end 18:29:18 fizzie, and mine has day but not such day changed line. 18:29:24 sqlite> select avg(length(body)) from logs; 18:29:24 40.8540693378636 18:29:27 In case you were wondering. 18:29:34 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx 18:29:38 That is an average messages for us. 18:29:40 It continues to amaze me how freaking *retarded* voting machines are. 18:29:41 Seems a bit short, but oh well. 18:29:48 fizzie, and I rotate monthly, or before that I rotated when the log dir started getting large 18:30:04 at least all files begin with log started lines 18:30:10 Longest line evar was 471. 18:30:11 alise: You probably want to filter type=0 there. 18:30:13 sqlite> select avg(length(body)) from logs; 18:30:13 40.8540693378636 18:30:13 sqlite> select avg(length(body)) from logs where type = 0; 18:30:13 41.6410379994674 18:30:13 -!- myndzi has quit (Ping timeout: 276 seconds). 18:30:20 fizzie: Well, phooey. 18:30:36 sqlite> select max(length(body)), body from logs where length(body) = max(length(body)); 18:30:36 Error: misuse of aggregate: max() 18:30:37 I think avg ignores nulls, though. But there's still quit messages and the like. 18:30:40 Well, fuck you! 18:30:46 fizzie, I'm not sure how to handle year wraparound 18:30:48 hm 18:30:58 guess I could detect dec -> jan 18:31:23 alise: sqlite> select length(body), body from logs where length(body) = (select max(length(body)) from logs); 18:31:24 471|Colorado Springs: A guy walked into a little corner store with a shot gun and demanded all the cash from the cash drawer. After the cashier put the cash in a bag, the robber saw a bottle of scotch that he wanted behind the counter on the shelf. He told the cashier to put it in the bag as well, but he refused and said "Because I don't believe you are over 21." The robber said he was, but the clerk still refused to give it to him because he didn't believe him 18:31:24 . At this 18:31:47 arke said that 18:31:48 god knows when 18:31:56 2005 18:32:12 the only 471-long message, god bless 'im 18:32:52 month = monthnum[m.group(1)] 18:32:52 if month < prev_month: # New year 18:32:52 current_year += 1 18:32:52 prev_month = month 18:32:54 XD 18:32:58 Hey, I have the second-longest message, 467 chars. 18:34:01 Can someone order-by-closeness-to-average-over-everyone? 18:34:07 i.e. who has the most average average line length? 18:34:28 Then one EgoBot, then *seven* fizzies, heh. (In the top-10 longest comments.) 18:34:43 fizzie: He definitely should've been carrying a 6-shooter instead. 18:34:48 (legal!) 18:36:52 The Most Unusual People: 18:36:56 454.0|412.358962000533|fizzief 18:36:56 325.5|283.858962000533|ElMexicano 18:36:56 325.5|283.858962000533|Linuxiano 18:36:56 325.0|283.358962000533|Yst 18:37:01 The latter is the difference between them and the average. 18:37:02 All higher... 18:37:18 alise: Top-5 averageness with a single statement: 18:37:19 sqlite> select nick, abs(avg(length(body))-(select avg(length(body)) from logs where type=0)) as k from logs where type=0 group by nick order by k asc limit 5; 18:37:19 alise_|0.0118991117972982 18:37:19 Xeanalyth|0.016037999467386 18:37:19 ehercd|0.0219903804197656 18:37:24 The averagest people: 18:37:29 41.6529371112647|0.0118991117972982|alise_ 18:37:29 41.625|0.016037999467386|Xeanalyth 18:37:29 41.6190476190476|0.0219903804197656|ehercd 18:37:29 41.5585106382979|0.082527361169511|mib_vvzkm4 18:37:29 41.54|0.101037999467387|osaunders 18:37:32 fizzie: Hey, thief. :| 18:37:49 Doing it first is not thievery. :p 18:37:52 TOTALLY IS 18:38:04 is ehercd me? I think he is. 18:38:11 I guess being the top speaker means I contribute a lot to the average. 18:38:22 433847|2008-01-23 18:54:48|ehercd|||0|it's emacslicious! 18:38:23 433848|2008-01-23 18:55:01|ehercd|||0|and perhaps a bit over the top, but hey, it'll beep for you.. 18:38:23 433849|2008-01-23 18:55:13|ehercd|||0|I wonder what would happen if i (dissociated-press)'d right now 18:38:26 That does sound like you. 18:38:55 yeah it's me mocking ERC 18:39:02 fizzie, I now return (is_utc, (year, month, day), (hour, minute, second), restofline) from my parsedate function, which seems to work on some test cases 18:39:05 now on to the rest 18:39:31 -!- MigoMipo has joined. 18:42:33 I have a pretty horrible set of regexps for irssi and bip logs. 18:43:35 http://p.zem.fi/tuks and so on. 18:44:00 fizzie, I gave up on purely regexp here 18:44:15 fizzie, and my logs are "xchat-style, but modified) 18:44:23 s/)/"/ 18:44:26 Anyway, away again. 18:44:29 cya 18:58:19 ~/local is ugly because the rest of my dirs are sentence-case :< 19:05:16 -!- Arty has joined. 19:05:59 -!- benuphoenix has joined. 19:06:24 -!- Arty has quit (Client Quit). 19:06:39 programming messups: int add(int in1, int in2){return in1+in2;} 19:06:52 -!- myndzi has joined. 19:06:59 benuphoenix: what a silly function 19:07:43 i almost put it on my java homework 19:09:09 http://stackoverflow.com/questions/184618/what-is-the-best-comment-in-source-code-you-have-ever-encountered 19:10:49 -!- myndzi\ has quit (Ping timeout: 265 seconds). 19:14:20 -!- augur has joined. 19:19:47 who is Graue? 19:20:02 God. 19:20:54 oh okay, then i might as well treat him like that other guy who had delusions of godhood 6000 years ago 19:21:00 * quintopia kills Graue and takes his place 19:21:29 he also administrates the esolangs server and wiki. 19:23:04 -!- Vorpal has set topic: Welcome to #esoteric, the international hub for esoteric programming language design and deployment | logs: http://tunes.org/~nef/logs/esoteric/?C=M;O=D. 19:23:10 (just a test for log parsing) 19:23:18 But there's no way that the Linear-Bounded Automaton complexity class cannot be approved. Already like half the languages on the list claim to be in that class. I have no idea why it does not appear already. 19:24:02 Vorpal: is your nick a reference to the Jabberwocky? 19:26:41 -!- calamari- has joined. 19:27:56 benuphoenix: As far as I know, Carroll invented that word, so every use is a reference to Jabberwocky, whether direct or not. 19:29:00 quintopia: Indeed. 19:29:11 Incidentally, the vorpal sword went snicker-snack. 19:30:07 only in the english 19:31:30 -!- calamari- has quit (Client Quit). 19:32:12 http://code.technically.us/post/1109586140/exchange-remote-wipe-is-a-terrible-terrible-bug Connect to exchange server on a mobile --> now your administrators can wipe your device 19:32:19 That's certainly an esoteric feature. 19:32:51 But there's no way that the Linear-Bounded Automaton complexity class cannot be approved. Already like half the languages on the list claim to be in that class. I have no idea why it does not appear already. 19:32:54 well, for a start, that's the wrong name :) 19:33:50 it'd be [[Category:Linear-bounded automaton]] 19:34:04 stop being pedantic 19:34:08 for one second 19:34:21 i'm stopping you getting killed by Graue. 19:35:24 I created a category without asking earlier this year. I didn't get killed and it's still there. He can't be as mean as you say. 19:35:38 he's inactive. just you wait 19:35:49 besides, that violates the policies regardless, so ais523 will probably delete it if he notices 19:36:06 quintopia: here is what happened the last time someone created a category and Graue noticed: http://esolangs.org/wiki/User_talk:Stux#Incident 19:36:39 that's for the year categories. which are now used. 19:37:18 alise: he did notice, told me off, and left them there. i think.\ 19:37:31 gah my log format is ambiguous 19:37:50 Vorpal: is your nick a reference to the Jabberwocky? <- yes, and nethack 19:40:45 * quintopia just read the incident. 19:41:07 I agree with Stux. Graue looks like a douche here. 19:41:15 -!- coppro_ has changed nick to coppro. 19:41:20 -!- coppro has quit (Changing host). 19:41:20 -!- coppro has joined. 19:41:30 quintopia: yes. his reaction was... over-the-top. 19:41:37 nonetheless, that's how strong the tradition is. 19:45:56 well, this time i followed procedure. how long will it takes? 19:49:50 probably forever 19:52:25 blergh @ java font rendering 20:01:21 error in: jijs demos button 20:01:21 20:01:25 could not locate dll GL: find dll 20:01:35 ('could not locate dll ',y) 13!:8]24 20:01:36 Thanks, J. 20:01:41 Real helpful. 20:03:36 (The answer is "install mesa development libraries".) 20:04:49 And yet OpenGL isn't even supported on J64, so it's pointless. 20:04:52 But you need to anyway. 20:05:37 haha 20:06:31 More fancy-graphical-esque REPL types should just be web-based. That is, run on a local port. 20:07:02 Also it would be *really* nice if there was a way to refer to something like http://localhost:python/ rather than http://localhost:894357345/ 20:08:35 alise, could you please send a notice to the channel? For regexp finding purposes. It seems to show different ones for myself and other people 20:09:41 -!- Leonidas has quit (Ping timeout: 240 seconds). 20:09:45 meh... anyone else? 20:09:55 was mine not good enough? 20:10:19 alise, I didn't get any ^_^ 20:10:31 alise, huh, I found the error, can you retry please 20:10:43 thanks 20:11:47 -!- Leonidas has joined. 20:21:07 I typed "/notice I have no idea what I am doing". What did I just do? 20:22:11 I kinda need to know so I can defend myself 20:27:57 such information will be disclosed on a need-to-know basis, not kinda-need-to-know :) 20:31:25 -!- MigoMipo has quit (Quit: Quit). 20:32:27 You most likely would have sent a notice to Mr I, but there seems to be no such person in freenode at the moment. 20:32:44 (Do they even allow one-letter names here?) 20:33:36 "Nick/channel is temporarily unavailable" -- maybe not, then. 20:34:37 alise, there's a command-line interface for J btw 20:34:37 -!- wareya has quit (Read error: Connection reset by peer). 20:35:33 FireFly: yeah but meh. 20:35:42 -!- wareya has joined. 20:35:44 I prefer it, at least 20:35:45 (Do they even allow one-letter names here?) 20:35:46 sort of. 20:36:53 QuakeNet's services have one-letter names; there's at least Q, R and S, maybe O too (I forget). 20:40:27 There was also an L before 20:43:58 alise: Here's a you-related statistic: http://p.zem.fi/48tk 20:44:04 -!- trinithis has quit (Ping timeout: 265 seconds). 20:44:14 fizzie: you need iEhird in there too :P 20:44:50 also, my goal is to one day break the 0.5 barrier 20:44:52 then 0.6 20:44:54 and then leave 20:44:55 forever 20:44:58 destroying the channel 20:45:10 Oh, I was assuming you'd go to something that's strictly greater than one, then leave. 20:47:17 wat 20:47:38 Well, you know, it's best to have some serious challenge in your goals. 20:52:52 -!- Vorpal has changed nick to Vorpal_. 20:52:53 -!- Vorpal_ has changed nick to Vorpal. 20:53:01 there, fixed the ambig in new logs 20:53:16 alise: Did you know you've written 15298175 bytes of messages here so far? With the usual "one megabyte per book" rule-of-thumb (and that's a *big* book; the h2g2 books are ~300k), you could already publish a 15-volume alisopedia. 20:53:16 fizzie, to parse my logs I need to switch to a new mode in the middle 20:53:26 fizzie, as in, a completely different set of regexp 20:57:55 fizzie: sweet 20:58:12 fizzie: I'm gonna actually do that 20:58:14 with lulu! 20:58:18 first person to buy it gets laughter 20:58:45 -!- Vorpal has changed nick to Vorpal_. 20:58:46 -!- Vorpal_ has changed nick to Vorpal. 21:00:27 -!- tombom has joined. 21:08:38 Also, here's where (some of) you people are: http://zem.fi/~fis/esomap.png 21:09:10 If you object to your neighbours, just remember: math doesn't lie. 21:12:44 what's that a map of? 21:13:08 map of what 21:13:47 nooga: yeah, must be 21:14:24 It's 32 different vaguely writing-style related features (I can provide a list) computed for all clog'd messages since 2004 for the people in the map; then the two first PCA components plotted as xy-scatterplot. 21:14:58 ok, what does that mean? :) 21:15:23 It doesn't really mean much. Besides, I see now that I've used a completely silly feature set; let's redraw a new map. 21:18:21 The default feature spec string seems to be really really boring: it's just a 25-item histogram of word lengths, then word length mean/variance, words per phrase mean/variance, phrases per message mean/variance and frequency of numerics. Even though the script does type/token counts and wordnet wordclass guesstimates and sort-of pronoun group frequencies and all. 21:24:15 FFT, equalize 21:24:26 save as wav and release a record 21:24:28 tadaaa 21:24:36 you're experimentam musician 21:25:14 fizzie, I don't think it is realistic to parse my irc logs before today, since while I could tell apart actions and parts and so on on screen due to colour, it was impossible in log file 21:25:17 fixed now anyway 21:25:59 experimental 21:29:49 Replotted that esomap with a more sensible settings. 21:30:01 -!- LeeWi_ has joined. 21:30:15 What, .fi *again*? 21:31:23 SRY! -.- 21:31:30 -!- LeeWi_ has left (?). 21:31:38 Heh, I scared someone off. 21:32:01 I wonder if I should, you know, /msg after or something. 21:32:22 I was just writing a "no, no, it's fine, it just makes me wonder" reply when he jumped ship. 21:38:49 msg 21:39:30 codzilla 21:49:16 Gah, octave's plotting functions are horreeble, even comparing to MATLAB, which is no picnic either. 21:59:07 whowas LeeWi_ 21:59:09 err 21:59:10 fail 22:00:51 I told em in a privmsg that I didn't actually mean e couldn't be here, but it didn't seem to help. Oh well. 22:00:56 -!- augur has quit (Remote host closed the connection). 22:01:11 -!- augur has joined. 22:06:55 Wow, that's a remarkably horrible plot: http://zem.fi/~fis/esomap-comp.png 22:08:40 I should start spewing more garbage so I get included in stats like these 22:08:53 random log quote: "* tusho Swhacks himself for being a naughty pirate." 22:08:55 alise, ^ 22:09:24 Wow, that's a remarkably horrible plot: http://zem.fi/~fis/esomap-comp.png <-- what does it show? 22:09:27 olsner: If I keep adding more people in that plot, it's going to be... uh, I can't say "unreadable", because it already is, but... "unreadabler?" 22:09:42 fizzie, besides being truncated at the side 22:09:48 Vorpal: Some of the (potentially more interesting) raw components used for esomap.png 22:09:59 Vorpal: You can guess most of the truncated bits, though. 22:10:08 fizzie, what is the esomap thingy 22:10:11 looks cool 22:10:13 but what is it 22:10:39 http://zem.fi/~fis/esomap.png -- "It's 32 [actually 60] different vaguely writing-style related features (I can provide a list) computed for all clog'd messages since 2004 for the people in the map; then the two first PCA components plotted as xy-scatterplot." 22:10:52 (I said that a few lines up there.) 22:11:00 fizzie, PCA? 22:11:20 Principal Component Analysis. 22:11:46 fizzie, also I can't guess at what "artion frequency" is, nor what "cy (she+her+hers+herself)" is 22:11:57 s/art/at/ 22:12:02 It basically finds the directions of largest variances for multidimensional data. 22:12:09 also "ses per paragraph"? 22:12:27 "ation frequency" -> "punctuation frequency"; it's part of the same set as the one under it, which does show more of it. 22:12:35 ah 22:12:46 fizzie, and what is the number at the end of each line 22:12:49 "cy (word1+word2+word3)" -> "something something group of words frequency (word1+word2+word3)" 22:13:08 Those are the actual numeric values at start and end. 22:13:18 ah 22:13:21 and the ses? 22:13:41 I tried to do it with a 16-part subplot, but it kept putting space everywhere and titles wherever, so I had to put them all in the same X axis; it's "physically" plotted for the [0, 1] range. 22:14:00 That's "phrases per paragraph"; it's a sort-of logical continuation to "chars per word" + "words per phrase". 22:14:09 huh 22:14:20 "paragraph" here means "message" in this context. 22:14:38 Oh, and the numeric values are probably normalized somehow, don't recall exactly how. 22:14:40 fizzie, what tool do you use for these 22:15:01 A really messy Perl thing I/we wrote for a previous course on natural language processing. 22:15:08 and those "PCA components", which ones are those? 22:15:12 -!- Sgeo has joined. 22:15:24 fizzie, as in: I have no clue what "Principal Component Analysis" means 22:15:29 Updating Flash makes a real difference to the browser's stability 22:15:36 It basically finds the directions of largest variances for multidimensional data. 22:15:37 I think I _can_ survive on Puppy for a while 22:15:40 That's sort of what it is. 22:15:46 If I had to 22:16:06 There's an illustrative plot in two dimensions at http://en.wikipedia.org/wiki/File:GaussianScatterPCA.png 22:16:27 fizzie, pretty 22:16:28 (The arrows there are the two principal components.) 22:16:42 fizzie, are the components always orthogonal? 22:16:51 random log quote: "* tusho Swhacks himself for being a naughty pirate." 22:16:52 alise, ^ 22:17:01 I'll give you whatever you want, just don't tell the tabloids! 22:17:07 Vorpal: Yes, IIRC it builds an orthogonal basis always. 22:17:09 alise, XD 22:17:36 "PCA is mathematically defined[2] as an orthogonal linear transformation that transforms the data to a new coordinate system such that the greatest variance by any projection of the data comes to lie on the first coordinate (called the first principal component), the second greatest variance on the second coordinate, and so on. PCA is theoretically the optimum transform for given data in least square terms." That's very succinctly put, quoted from wikipedia. 22:17:55 fizzie, so that means that the last axis will always be fixed to be orthogonal against the other ones. But I don't see how you can go from 64 dimensions to 2 22:18:01 fizzie, 2 <-> 2 is easy 22:18:12 05:43:13 Hmm. I haven't bought any albums, recently, apart from this one. 22:18:12 05:43:19 * tusho Swhacks himself for being a naughty pirate. 22:18:12 05:43:23 * tusho denies that Swhack. 22:18:12 05:43:30 (/me checks download status...) 22:18:12 Oh, I just only plotted the first two ones; those are, after all, where the largest variance is. 22:18:14 YOUR LIES HAVE BEEN EXPOSD 22:18:16 *EXPOSED 22:18:18 fizzie, ah 22:18:41 fizzie, these axis could be somewhere between the original ones I presume? 22:19:10 Sure, otherwise it'd be just "find variance for each component, select max". 22:19:19 swhick-swhack 22:19:21 05:59:16 and this mouse is very crappy, I suspect it's made of cardboard but am not sure 22:19:45 alise, huh? I didn't check the context, it was a purely random selection from actions said by you 22:19:52 Vorpal: I know 22:19:55 I was joking. 22:19:59 ah 22:20:07 Assuming you realise what the standalone message sounded like without context. 22:20:15 alise, yeah after I pasted it :P 22:20:22 alise, but before you responded 22:21:05 fizzie, fancy stuff, you need to make an n-dimensional explorer for it! Hm didn't glfunge do that kind of? 22:21:16 for funge-space 22:21:19 or was that some other one 22:21:40 It was on the todo list, but I don't think I got very far. Some other one might've done it better. 22:22:19 that non-compliant bequnge or whatever iirc did it 22:22:23 What I should do is to randomly divide those per-person datasets into N parts and compute the features separately for each one, then you'd get to see a bit how much intra-person variance there is. 22:23:13 . 22:23:19 Vorpal: I have formulated two SQL queries to generate micro-arguments. 22:23:20 Behold: 22:23:22 sqlite> select nick, body from logs where type="0" and (nick like "ehird%" or nick like "tusho%" or nick like "alise%") and (body like "AnMaster:%" or body like "Vorpal:%") order by random() limit 1; 22:23:23 sqlite> select nick, body from logs where type="0" and (nick like "AnMaster%" or nick like "Vorpal%") and (body like "ehird%," or body like "tusho%," or body like "alise%,") order by random() limit 1; 22:23:24 I love xlock 22:23:33 Read results after one another. 22:23:33 alise, hah 22:23:35 Sgeo: slock is far superior 22:23:37 :>P 22:23:39 *:P 22:23:44 slock? 22:23:52 Sgeo: Oh, you're here; re Turing-completeness of sqlite/triggers, see http://p.zem.fi/sqlite-tm.sql 22:23:54 http://tools.suckless.org/slock 22:24:11 Vorpal: 22:24:13 sqlite> select nick, body from logs where type="0" and (nick like "ehird%" or nick like "tusho%" or nick like "alise%") and (body like "AnMaster:%" or body like "Vorpal:%") order by random() limit 1; select nick, body from logs where type="0" and (nick like "AnMaster%" or nick like "Vorpal%") and (body like "ehird%," or body like "tusho%," or body like "alise%,") order by random() limit 1; 22:24:14 as a single command 22:24:15 alise, what is the word for "not realistic setting in historical work" now again 22:24:20 I forgot it 22:24:25 Vorpal: oh wait i got something wrong 22:24:29 Vorpal: err "fictionalised"? 22:24:33 i'm not sure what you mean 22:24:38 What happens if someone sees an xlock-ed or slock-ed screen and just presses Ctrl-Alt-Backspace? 22:24:46 Vorpal: 22:24:48 sqlite> select nick, body from logs where type="0" and (nick like "ehird%" or nick like "tusho%" or nick like "alise%") and (body like "AnMaster:%" or body like "Vorpal:%") order by random() limit 1; select nick, body from logs where type="0" and (nick like "AnMaster%" or nick like "Vorpal%") and (body like "ehird%,%" or body like "tusho%,%" or body like "alise%,%") order by random() limit 1; 22:24:56 alise, no, as in clothes from wrong period in a holy wood movie taking place in 1400 century 22:24:59 Sgeo: seeing an slocked screen == seing nothing :P 22:25:02 Vorpal: anachronism 22:25:04 alise, right 22:25:07 anachronistic, etc. 22:25:10 thus anacron 22:25:14 the daemon 22:25:19 alise, mixing ehird and vorpal in an argument is anachronistic 22:25:21 ehird|AnMaster: not the same thing 22:25:21 AnMaster|alise, terminal edition? I want F24! 22:25:22 that was what I wanted to say 22:25:35 it seems to always show as ehird/AnMaster 22:25:37 not sure why yet 22:25:47 ehird|AnMaster: if you want to use GUI subversion on OS X, http://versionsapp.com/ seems the most popualr way. 22:25:47 AnMaster|ehird, see the "unauthicated" at the top? 22:25:51 And arbitrary turing machines can be constructed in this fashion? 22:25:55 ehird|AnMaster: you're using the closed source virtualbox right 22:25:55 AnMaster|ehird, why? 22:26:07 alise, probably most common 22:26:14 AssertionError: invalid log body: 471 22:26:16 interesting 22:26:18 Sgeo: I'm pretty confident they can, though I haven't algorithmized the construction yet. 22:26:20 from importing own logs 22:26:22 Vorpal: i would have expected one alise or tusho by now, though 22:26:32 fizzie: You're awesome 22:26:34 ehird|AnMaster: Does bzr work on Windows? 22:26:34 Vorpal|alise, while I'm pretty sure it mentions - should work. Like it does for cat and so on 22:26:50 http://zem.fi/~fis/esomap-comp.png ;; so who wins eh 22:27:08 also, are people at opposite ends most-unalike for that? 22:27:20 hm 22:27:41 oh, alise finally 22:27:45 alise|AnMaster: GHC doesn't take much RAM to compile. I don't _want_ OpenOffice. 22:27:45 AnMaster|ehird, I have seen enough of the file format 22:27:46 alise: Yes, though the actual numeric values that it shows at both ends seem to have been normalized a bit. 22:28:10 Vorpal: Ha, the first actually-coherent argument: 22:28:12 tusho|AnMaster: the second sentence is irrelevant 22:28:12 AnMaster|tusho, you wouldn't have said that if I pointed out the error :P 22:28:31 ehird|AnMaster: that's as may be. 22:28:31 AnMaster|ehird, CRT? where? 22:28:40 randomised #esoteric would make a great automatically generated comic strip 22:28:48 It's incredible how rarely I use descriptive adjectives 22:29:18 fizzie, hm, those components, how do you get "verbs"? Do you use a dict? What about made up words then 22:29:34 Vorpal: does "not like" work in SQL? 22:29:53 alise, try it and see 22:29:56 :P 22:30:01 i don't need it actually, just realised 22:30:27 Q: why? 22:30:28 A: i think i might make an evolutionary AI for blahbot` 22:30:56 * ehird considers gobolinux+etoile 22:31:00 alise, what was etoile? 22:31:13 http://etoileos.com/ 22:31:19 they're basically taking GNUStep + Smalltalk 22:31:22 Vorpal: The word classes are very much guesswork. For each word, it takes all wordnet's senses that have frequency counts, selects the most frequent one, and classifies as that. It completely fails to take context into account, e.g. "red" is always an adjective, not a noun, since wordnet count for red (adj) is 43, for red (noun) 9. 22:31:26 "* oerjan hits himself ===\___/" 22:31:27 hm 22:31:29 and developing it into something pretty awesome 22:31:31 I wonder what he did 22:31:50 Vorpal: There's a pile of context-considering word-class disambiguators, but they all take a horrible amount of time to run. 22:31:51 What's this about Smalltalk? 22:31:53 * Sgeo insanes 22:31:56 Sgeo: No Smalltalk. 22:32:01 It's actually based on... C++. 22:32:04 fizzie, red as noun? 22:32:12 they're basically taking GNUStep + Smalltalk 22:32:14 And... and... PHP. 22:32:18 Sgeo: I was lying. 22:32:20 Joke. Ha ha! 22:32:21 You're just trying to scare me 22:32:22 oh yeah the name of the colour 22:32:26 Hilarious, yes. 22:32:29 NO SMALLTALK THER. 22:32:31 *THERE 22:32:32 AT ALL 22:32:33 GNUStep HEH 22:32:35 Vorpal: 1. (9) red, redness -- (red color or pigment; the chromatic color resembling the hue of blood) 22:32:42 right 22:32:42 GNUStep vs Apple's stuff 22:32:46 nooga: gnustep is shit but etoile are doing a surprising job at making it awesome 22:32:49 (they're not trying to clone OS X) 22:32:55 yeah i know 22:32:59 etoile looks promising 22:33:02 The webpage has code that looks very Smalltalk-y 22:33:07 Sgeo: nope it's C++ 22:33:14 but i think it's too early to assess it 22:33:47 You don't actually think I believe you right now, do you? 22:33:49 fizzie, anyway parsing my own logs turn out to be hugely complicated 22:33:51 :/ 22:34:02 Sgeo: I'm just hoping. 22:34:08 Vorpal: just use a heuristic for joins/parts 22:34:10 no biggie 22:34:20 "It is a port of the GoboLinux tools to Cygwin, allowing them to be used on top of Windows." But... why? 22:34:36 alise, "* alise action" "* Topic for #esoteric is" 22:34:40 alise, and so on 22:34:41 Vorpal: indeed 22:34:42 I fixed this now 22:34:45 but yeah it's a PITA 22:34:46 Vorpal: assume that nobody called Topic has ever been in 22:34:47 or at least 22:34:49 just use this for topics 22:34:56 /\* Topic for #esoteric is / 22:35:04 if it gets some false shit in, meh, who cares 22:35:06 alise, it will misclassify a lot. "* [alise] blah blah" is a whois result 22:35:10 Vorpal: so? 22:35:13 i've never been called [alise] 22:35:18 indeed 22:35:20 -!- Sgeo has changed nick to [Sgeo]. 22:35:20 so just filter out all [foo]s 22:35:25 it is valid though 22:35:26 in nicks 22:35:28 yes 22:35:28 * [Sgeo] screws with stuff 22:35:31 but it's old stuff 22:35:35 alise, anyway I made actions use * and non-actions use + from now on 22:35:41 so just parse the old stuff, re-save them and fix errors when you come across them 22:35:45 i.e. convert to new format with a heuristic 22:35:49 alise, oh and /exec without -o went to log without the timestamp 22:35:58 but some other stuff also lacks timestamp 22:35:59 and so on 22:36:37 <[Sgeo]> AFK 22:36:43 <[Sgeo]> Eating half of dinner at the pizza place 22:36:49 alise, and what I'm doing currently is writing those heuristics :P Which is hugely complicated 22:37:11 * [Sgeo] screws with stuff <-- it won't mess up stuff any more :P 22:40:34 alise, " I should become the U.S. archivist." 22:42:52 Naturally. 22:43:34 <[Sgeo]> . 22:43:42 <[Sgeo]> AFK 22:43:58 farfetched 22:45:42 * oerjan doesn't think people would be happy if he started using #esoteric for backup 22:45:45 I wonder how that would work 22:46:01 * oerjan suddenly has an idea for an #esoteric slogan <-- I think I need to read the context here 22:48:01 * ais523 quickly leafs through their #esoteric handbook for the bit on protecting oneself from an angry Slereah_$ 22:48:01 -!- [Sgeo] has quit (Ping timeout: 252 seconds). 22:48:02 heh 22:48:03 Vorpal: dump base64fied files and then restore them from logs 22:48:13 nooga, hah 22:50:44 -!- benuphoenix has quit (Quit: leaving). 22:51:03 Hey, that's actually a pretty nice spread: http://zem.fi/~fis/esomapn.png 22:51:04 awesome 22:51:19 yaay 22:51:26 fizzie, what are the colours 22:51:40 Just arbitrary colours 30 degrees apart in the hue wheel. 22:51:42 With some overlap. 22:51:43 i've got the biggest spread ad 6 22:51:51 fizzie, what does it depict 22:51:59 what does this mean? 22:52:23 It's again two first PCA components, except this time for each person's messages have been randomly split into ten parts. 22:52:24 fizzie, and why is pikhq so similar to oerjan 22:52:44 Based on that, given a large enough sample, even these simple features could be used to determine reasonably reliably which one out of these 13 people was speaking, except that pikhq/oerjan do get confused. 22:52:45 fizzie, oh nice 22:53:02 Using all the 60 dimensions could easily unconfuse those two too, this is just a 2-dimensional projection after all. 22:53:04 fizzie: so do I win? 22:53:10 also, i could guess who said who 22:53:12 with very high accuracy 22:53:14 *who said what 22:53:17 assuming it was just the regulars 22:53:18 fizzie, there is also some sgeo/deewiant overlap 22:53:27 i could easily distinguish sgeo/deewiant 22:53:35 fizzie, it is interesting to note both me and alise have very small spread 22:53:38 seriously, set up a game where it randomly gets a, say, 5-line chunk of logs 22:53:40 perhaps due to larger sample size? 22:53:42 alise: Sure, but you're not an algorithm. 22:53:42 and make me fill in the names 22:53:45 i will pwn 22:53:46 :| 22:53:50 fizzie, thus meaning it will be more uniform? 22:53:53 fizzie: make an animated 4d projection 22:54:01 Vorpal: Yes, larger datasets probably make for smaller variance there. 22:54:27 fizzie, might be interesting to split in equal sized bins, so everyone get n messages per dot 22:54:41 fizzie, to see if the spread size does vary much or not 22:54:43 kod kod kod 22:54:44 for us 22:55:07 pikhq: I sure hope Fine Structure becomes more than merely a collection of Hughes' best short stories as it goes on. 22:55:59 Vorpal: I don't have a script for that handy, but I can test-run that too. 23:00:27 -!- tombom has quit (Quit: Leaving). 23:02:31 fizzie, awesome 23:03:25 (But later; early morning tomorrow.) 23:03:37 ah 23:03:45 alise, * ais523 is confused that tusho seems to think that ais523 is a ChanServ expert 23:04:17 he's said similar things to you far many more times 23:04:18 :) 23:04:45 are you just randomly logreading? 23:07:49 Vorpal: I did one more slight prettification to the existing esomapn.png, though: added contour-ellipses at sigma and 2*sigma for gaussians fitted on those ten points. (Though ten is not a very large sample to estimate things from, so... Still, it looks undeniably more sci-fi now.) 23:07:54 alise, I'm taking last action by you or involving you in the screenful before a crash of the parser 23:08:05 or in some cases not involving you 23:08:39 fizzie, so it does :P 23:12:38 STAR TREKKIN' ACROSS #ESOTERICVERSE 23:12:54 ON THE STARSHIP IRC, UNDER CAPTAIN ... UM... ...!! 23:12:56 STAR TREKKIN' ACROSS #ESOTERICVERSE 23:12:58 etc. 23:13:16 fizzie: Weren't those lines before? 23:13:20 Or was it changed from that earlier? 23:14:18 "* ehird googles core i7 sucks" 23:17:26 -!- olsner has quit (Quit: Leaving). 23:17:59 -!- zzo38 has quit (Remote host closed the connection). 23:18:22 -!- Sgeo|web has joined. 23:19:01 The lines weren't in an earlier version, but I just sneakily updated in-place. 23:19:49 Random 32-bit number in J: #.?32$2 23:21:05 Random 32-bit number redux: ?2^32 23:21:05 :P 23:21:10 But the former has a different distribution, I think. 23:21:13 fizzie, do pikhq and oerjan differ more on some pther PCAs? 23:22:41 STAR TREKKIN' ACROSS #ESOTERICVERSE <-- too long to fit on the line. try "esoverse" to get the same number of syllables 23:22:49 true 23:22:53 who's the captain 23:22:59 um 23:23:08 who owns the channel? 23:23:11 "lament" is too long, "fizzie" is too long, "lilo" is too long and too dead 23:23:14 star trek HA! 23:23:21 andreou 23:23:22 hm 23:23:23 i've actually never watched it 23:23:24 Vorpal: founder is andreou, but 23:23:32 -ChanServ- 1 andreou +votsriRfAF [modified ? ago] 23:23:33 -ChanServ- 2 fizzie +votsriRfA [modified ? ago] 23:23:33 -ChanServ- 3 lament +votsriRfA [modified ? ago] 23:23:33 -ChanServ- 4 Aardappel +votiA [modified ? ago] 23:23:33 -ChanServ- 5 oerjan +votsriRfA [modified 30 weeks, 1 day, 01:00:41 ago] 23:23:33 who is andreou 23:23:38 nooga: our glorious founder 23:23:52 i don't know him, never talked with him 23:23:58 alise, also it needs a longer starship name than irc 23:24:03 enterprise is far longer 23:24:30 AssertionError: invalid log body: SOCKSProxy failed to connect to host (error 1). 23:24:31 hah 23:25:01 Vorpal: eye are cee 23:25:02 en ter prise 23:25:08 same syllables, fits perfectly 23:25:17 only stupid people say irk :) 23:25:33 j is so much fun to try and work out 23:27:08 " 23:27:08 *y is _1 if y is negative, 0 if it is zero, 1 if it is positive; more generally, *y is the intersection of the unit circle with the line from the origin through the argument y in the complex plane." 23:27:47 hmm 23:27:54 this etoile actually looks really cool 23:28:08 i always wanted to use Obj-C outside of OSX 23:30:12 Vorpal: http://zem.fi/~fis/esomap34n.png -- 3rd/4th principal components. These differentiate oerjan and pikhq a bit better. 23:30:47 fizzie: why a I always in a different place? 23:31:16 ha, alise has the smallest galaxy :D 23:31:50 nooga, not unexpected as discussed before 23:31:54 nooga, smaller is better 23:31:56 Is that "different" as in "different from others" or "different between images"? 23:32:47 so what does smaller actually mean? 23:32:59 A lot of those features are means, which tend to have smaller variance as the sample size grows. I'll see if I can do that same-sized sets thing. 23:33:45 With same sample sizes, a tighter grouping would mean the things it measures don't vary much in the person's text. 23:33:49 fizzie: different between images, i guess that's because of perspective 23:34:33 fizzie, you could use all axis to get some useful differentiation between everyone perhaps? 23:34:56 Vorpal: Any sensible classifier would use all 60 components, sure. 23:34:57 -!- augur has quit (Remote host closed the connection). 23:35:06 fizzie: DO IT. 23:35:19 DO IT! FA... 23:35:37 fizzie, btw apart from the components being orthogonal, are they also the same length 23:36:01 fizzie, forgot what "all axis having equal-sized 1" was called 23:36:18 Vorpal: I don't think it's orthonormal, no. 23:36:27 ah right that was the word 23:36:46 fizzie, brb 23:37:41 alise: I'll try out training an alise-detecting neural net tomorrow or later. :p 23:37:47 Not hard :P 23:38:21 huh 23:38:41 just map one, small part of my brain 23:39:04 nooga: what? 23:40:29 Extracting algorithms out of brains, now that'd be quite a neat trick. 23:40:49 yeah 23:41:14 alise: nothing, it's just that somehow i can tell that you is you even if you change nick 23:41:43 it's easy to tell who says something even if you don't look 23:41:46 the small part of my brain directs my fingers to type /whois and etc. 23:41:47 we're all very predictable 23:42:48 even if i will start producing textual white noise 23:42:55 ; 23:48:28 yes 23:49:28 fizzie, please include zzo in that graph too 23:49:35 fizzie, just to see where he ends up 23:50:50 Vorpal: the galaxy would appear outside of your screen's border 23:50:52 in midair 23:50:56 *mid-air 23:51:04 alise, hah. 23:51:44 hm how to best characterise zzo? Perhaps the almost agressive accepting of the other persons viewpoint after stating his own? (But is fine too if you prefer it that way) 23:51:48 alise: that was not a question 23:51:51 or his use of "source codes" 23:51:57 rather than "source code" 23:52:13 to me code in the sense of "source code" is uncountable 23:52:52 -!- distant_figure has quit (Remote host closed the connection). 23:53:46 -!- augur has joined. 23:54:54 i have lots of source codes 23:55:06 -!- distant_figure has joined.