00:00:07 quintopia, a) I was only mistaken for him by a drunk middle-aged unattractive man, and b) I'm asexual 00:00:38 And c) no 00:00:57 i don't see how the being mistaken for mark zuckerberg by such an outlier has any bearing on the present decision on whether to go out in a bowtie 00:01:44 Hmm, maybe you're right 00:06:54 -!- carado has quit (Ping timeout: 252 seconds). 00:13:01 -!- Bike has joined. 00:13:30 -!- realz has joined. 00:13:42 -!- realz has quit (Changing host). 00:13:42 -!- realz has joined. 00:35:35 do people actually use fibonacci heaps for anything 00:37:34 http://www.quora.com/What-is-a-Fibonacci-Heap?share=1 has an answer in the negative from "Adam D'Angelo, former algorithms TA" 00:41:50 kmc: I've used a heap library that had a fibonacci heap as one of the backends 00:41:57 but I'm not sure whether it used that backend or not 00:42:12 when I've written heaps manually, I typically use binary heaps 00:43:19 hmm… I remember that paper which had an article hitting all the heap complexity bounds in actual time 00:43:21 rather than amortized time 00:43:30 using a really crazy and complex heap impl 00:44:40 http://exoplanets.org/ attention astro nerds 00:45:21 btw, I didn't tell you about my latest programming project that I haven't told you about yet 00:45:36 ais523\unfoog: that's neat 00:45:37 it's not particularly eso, but it's about the same order of bad idea 00:45:42 I guess there is general research on de-amortization 00:45:57 basically, it's a deparser for text documents 00:46:00 -!- yorick has quit (Remote host closed the connection). 00:46:14 you know how you can format HTML or Texinfo or nroff or whatever as text? 00:46:26 this would go the other way, with no knowledge of the source format 00:46:31 trying to infer how various things were formatted 00:46:44 I can think of a huge many usecases for this 00:46:49 but it's likely to be quite difficult to write 00:47:06 there are some startups doing this, or similar 00:47:13 finding PDFs on the web and extracting them to data tables 00:47:30 that's not exactly what I want to do, but it's the same field 00:49:24 actually the motivating example was to write a program that could display all the random files lying around /usr/share/doc in whatever format the user wanted 00:49:30 and the motivating reason for /that/ was writing a better man/info/yelp 00:49:56 I was going to say "smoothsort", but it turns out that uses a completely different sort of heap. 00:50:38 The Leonardo numbers, apparently. 00:50:55 (anyone who hangs out in #acehack will know that I have trivial motivating reasons for really complex projects) 00:52:01 I think I'll start by converting everything to elastic tabs 00:52:09 (something that text editors should do anyway) 00:53:56 kmc: i feel like i should read _Purely-Functional Data Structures_ 00:53:58 did you read it 00:55:24 -!- Taneb has quit (Quit: Leaving). 00:58:11 -!- john_metcalf has quit (*.net *.split). 00:58:11 -!- FreeFull has quit (*.net *.split). 00:58:11 -!- ski has quit (*.net *.split). 00:58:11 -!- Lymia has quit (*.net *.split). 00:58:11 -!- rodgort has quit (*.net *.split). 00:58:11 -!- FireFly has quit (*.net *.split). 00:58:19 -!- ski has joined. 00:58:24 some of it 00:58:26 it's good 00:58:29 -!- FreeFull has joined. 00:58:38 -!- Lymia has joined. 00:58:47 -!- rodgort has joined. 00:58:57 -!- FreeFull has quit (Changing host). 00:58:57 -!- FreeFull has joined. 00:59:13 -!- Lymia has quit (Changing host). 00:59:13 -!- Lymia has joined. 01:00:42 -!- FireFly has joined. 01:05:26 :t flip (foldr id) 01:05:27 [c -> c] -> c -> c 01:07:08 > (flip (foldr id)) [succ, pred] 17 01:07:10 17 01:07:14 metal 01:07:26 > (flip (foldr id)) [(7+),(2*)] 17 01:07:27 41 01:08:44 @pl \f.\g.\h.\x.g(f(h(x))) 01:08:44 (line 1, column 3): 01:08:44 unexpected "." 01:08:45 expecting letter or digit, operator, pattern or "->" 01:08:53 @pl \f->\g->\h->\x->g(f(h(x))) 01:08:56 flip ((.) . (.)) . (.) 01:08:57 optimization suspended, use @pl-resume to continue. 01:09:06 bleh, lambdabot, why don't you work on Verity? 01:09:08 @pl-resume 01:09:14 is @pl still ill 01:09:14 flip ((.) . (.)) . (.) 01:09:14 optimization suspended, use @pl-resume to continue. 01:09:21 @pl-resume 01:09:24 oerjan: I think so 01:09:26 i've yet to see @pl-resume give a different answer 01:09:33 flip ((.) . (.)) . (.) 01:09:34 optimization suspended, use @pl-resume to continue. 01:09:55 anyway, that function seems to be the programming-language version of the functor \Rightarrow 01:09:57 which is kind-of important 01:10:00 so I was wondering if it had a name 01:10:47 @pl \f g h x -> g (f (h x)) 01:10:50 flip ((.) . (.)) . (.) 01:10:50 optimization suspended, use @pl-resume to continue. 01:10:52 @pl \(f,g)->\(x,y)->(f(x),g(y)) 01:10:52 uncurry (flip flip snd . (ap .) . flip flip fst . ((.) .) . flip . (((.) . (,)) .)) 01:11:00 that's even worse 01:11:03 that one's \otimes 01:11:15 and it looks reasonable that that one might have a name 01:11:18 or at least a short impl 01:11:24 @pl \f g h -> g . f . h 01:11:28 flip ((.) . (.)) . (.) 01:11:28 optimization suspended, use @pl-resume to continue. 01:11:44 is Haskell this bad at functors? 01:11:52 @pl \f g -> (.) (g . f) 01:11:56 ((.) .) . flip (.) 01:11:56 optimization suspended, use @pl-resume to continue. 01:11:57 isn't "\(f,g)->\(x,y)->(f(x),g(y))" a short implementation 01:12:41 :t uncurry (***) 01:12:42 Arrow a => (a b c, a b' c') -> a (b, b') (c, c') 01:12:57 uh i'm pretty sure arrow is the antichrist....>??? 01:13:03 OKAY 01:13:23 going off reliable esoteric information here 01:13:23 -!- prooftechnique has joined. 01:13:37 > uncurry (***) ((+1),(*2)) (5,10) 01:13:37 in the categorical semantics I'm working on 01:13:38 (6,20) 01:13:38 uncurry = id 01:14:01 oerjan: I like your impl 01:14:03 ais523\unfoog: anyway @pl doesn't know much about Arrows. 01:14:12 right 01:14:27 Arrow's a bit of an ironic name, though 01:14:46 given that arrow/morphism in categorical semantics = function in programming language 01:15:59 sometimes I feel that Haskell isn't as category-theoretical as its supporters make it out to be 01:17:16 i feel betrayed. 01:19:09 ais523\unfoog: well Arrow is a subclass of Category nowadays. 01:19:20 oerjan: that makes even less sense 01:19:21 in particular, (->) is only one instance. 01:19:26 category theory doesn't work like that 01:19:37 the name Arrow is not particularly well-chosen, though. 01:20:15 It seems better it should have: functor between categories, tensor categories, fanout category, etc, instead of being like such Arrow class 01:20:52 ais523\unfoog: the idea is that a value of type Arrow a x y is an arrow from x to y in a category given by a. 01:21:00 oerjan: right 01:21:18 although the class heaps on stuff that isn't just arrow requirements. 01:21:26 (Category alone has those.) 01:21:32 the only actual requirement is for a composition operator 01:21:33 I think it'd be fun to make a nethack clone with ed keybindings 01:21:49 prooftechnique: NetHack 4 (my fork) has rebindable keys 01:22:02 not sure if ed's keybindings even remotely correspond to the commands ed has, though 01:22:08 prooftechnique: Try it, what key would they command 01:22:18 ais523: What changes did you make though? 01:22:22 zzo38: loads 01:22:25 mostly to internals 01:23:03 Did you add rebindable keys or is that already in there? 01:23:52 NetHack 3.4.3 doesn't have rebindable keys 01:24:03 but the key rebinding code was copied from another NetHack fork (NitroHack) 01:24:12 O, well that is good you add it, then; it was missing then 01:24:18 however, the key /parsing/ code is new 01:28:14 OK 01:32:48 is there a bf clone that does non-deterministic computations in deterministic time? because if not, i want to make one 01:33:00 quintopia: interesting 01:33:04 a sort of quantum BF? 01:33:07 yes 01:33:08 q: what does that mean 01:33:12 everettian BF 01:33:13 actually, I think Brainfork can do that 01:33:19 * quintopia looks 01:33:21 so long as there's a way to kill threads, not sure if there is 01:33:44 who is the guy in the google logo 01:34:18 btw, Brainfork was invented before BF derivs got old 01:34:23 quintopia: The doodle today? 01:34:26 it was like the second or third semantics-changing deriv 01:34:48 prooftechnique: yes 01:34:52 i'm too lazy to click it 01:34:54 http://en.wikipedia.org/wiki/Shakuntala_Devi 01:35:02 Mental math world record or something 01:36:50 ais523\unfoog: yeah, it doesn't look like there's a way to kill forked threads in brainfork. but you can sort of do nondeterministic computations by doing some very fancy algorithming. but my idea would add no new commands and wouldn't require thread management. 01:37:04 -!- glogbackup has joined. 01:37:18 hi glogbackup 01:38:35 `welcome glogbackup 01:38:37 glogbackup: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: . (For the other kind of esoterica, try #esoteric on irc.dal.net.) 01:40:12 -!- Uguubee1111113 has joined. 01:40:52 ais523\unfoog: my idea is to have the semantics of . change to "append byte to waiting-to-output queue" and the semantics of program halt change to "if waiting-to-output queue matches the tail of the input list, output, otherwise, move all amplitude of this universe's wavefunction into the universe where the tail of the input list is equal to the contents of the current waiting-to-output queue at the moment of program instantiation" 01:41:17 obviously, , would not consume the input list, just advance an index into it 01:41:37 `run grep '^: Welcome to the' /var/logs/_esoteric/*.txt | cut -f1 -d':' | cut -f2 -d'>' | sort | uniq -c 01:41:38 grep: /var/logs/_esoteric/*.txt: No such file or directory 01:41:48 `run grep '^: Welcome to the' /var/irclogs/_esoteric/*.txt | cut -f1 -d':' | cut -f2 -d'>' | sort | uniq -c 01:42:03 wait, there's a stray ^ there 01:42:06 `run grep ': Welcome to the' /var/irclogs/_esoteric/*.txt | cut -f1 -d':' | cut -f2 -d'>' | sort | uniq -c 01:42:19 No output. 01:42:37 No output. 01:43:08 hmm 01:43:10 is it timing out? 01:43:17 `run grep ': Welcome to the' /var/irclogs/_esoteric/2013*.txt | cut -f1 -d':' | cut -f2 -d'>' | sort | uniq -c 01:43:25 or is there something wrong with the line? 01:43:31 18 /var/irclogs/_esoteric/2013-01-04-raw.txt \ 18 /var/irclogs/_esoteric/2013-01-04.txt \ 1 /var/irclogs/_esoteric/2013-01-05-raw.txt \ 1 /var/irclogs/_esoteric/2013-01-05.txt \ 1 /var/irclogs/_esoteric/2013-01-06-raw.txt \ 1 /var/irclogs/_esoteric/2013-01-06.txt \ 1 /var/irclogs/_esoteric/2013-01-07-raw.t 01:43:38 One thing is you should indicate whether you want raw or non-raw 01:43:39 oh, right 01:43:54 -!- Sgeo has joined. 01:44:13 `run grep -h ': Welcome to the' /var/irclogs/_esoteric/2013-??-??.txt | cut -f1 -d':' | cut -f2 -d'>' | sort | uniq -c 01:44:17 7 00 \ 16 01 \ 18 02 \ 6 03 \ 7 04 \ 11 05 \ 10 06 \ 11 07 \ 2 08 \ 2 09 \ 4 10 \ 10 11 \ 6 12 \ 7 13 \ 6 14 \ 11 15 \ 11 16 \ 14 17 \ 16 18 \ 15 19 \ 16 20 \ 19 21 \ 20 22 \ 18 23 01:44:36 aha, there are colons in the timestamp 01:44:40 `run grep -h ': Welcome to the' /var/irclogs/_esoteric/2013-??-??.txt | cut -f1 -d'4' | cut -f2 -d'>' | sort | uniq -c 01:44:43 7 0 \ 2 \ 2 00: \ 1 00:10:3 \ 1 00:3 \ 1 00:32: \ 7 01: \ 1 01:05:3 \ 1 01:08: \ 1 01:13: \ 1 01:31: \ 1 02: \ 1 02:09: \ 1 02:56: \ 1 03: \ 1 03:19:0 \ 1 03:31:2 \ 1 03:5 \ 1 {^Raven^}: Welcome to the international 01:44:48 `run grep -h ': Welcome to the' /var/irclogs/_esoteric/2013-??-??.txt | cut -f4 -d':' | cut -f2 -d'>' | sort | uniq -c 01:44:50 1 boily \ 1 lambdabot \ 1 Phantom_Hoover \ 1 {^Raven^} \ 1 intensity \ 1 kyyni \ 1 shikhin \ 1 matthiaskrgr \ 1 simmarine_ \ 1 snuffeluffegus \ 1 audioPhil 01:44:56 this is good 01:45:02 `run grep -h ': Welcome to the' /var/irclogs/_esoteric/2013-??-??.txt | cut -f4 -d':' | cut -f2 -d'>' | sort | uniq -c | sort -rn 01:45:05 13 Bike \ 8 oerjan \ 5 monqy \ 5 kmc \ 4 sivoais \ 4 `run grep -h ' \ 3 Taneb \ 3 SirCmpwn \ 3 Sgeo \ 3 Regis_ \ 3 Fiora \ 2 yhojeyisaac \ 2 windmill \ 2 Tritonio \ 2 ThatOtherPerson \ 2 surma \ 2 sploknee \ 2 Snowyowl \ 01:45:15 :o 01:45:21 `run grep -h ': Welcome to the' /var/irclogs/_esoteric/20??-??-??.txt | cut -f4 -d':' | cut -f2 -d'>' | sort | uniq -c | sort -rn 01:45:36 22 fdhsfdsgdsg \ 19 oerjan \ 16 elliott \ 15 Bike \ 13 monqy \ 9 shachaf \ 8 ais523 \ 6 HackEgo \ 5 Taneb \ 5 sivoais \ 5 `run grep -h ' \ 5 -!- oerjan changed the topic of #esoteric to \ 5 kmc \ 5 fungot \ 5 Fiora \ 4 zzo38 \ 4 shub 01:46:10 who was fdhsfdsgdsg and why did they get welcomed so much? 01:46:16 `pastlog fdhsfdsgdsg 01:46:24 2012-03-02.txt:12:29:26: ^ul (?so `ok fdhsfdsgdsg: Welcome to the international hub for esoteric programming language design and deployment! For more information, check out our wiki: http://esolangs.org/wiki/Main_Page)S 01:46:38 botloop, I guess 01:46:50 ais523\unfoog: sorry to be zzo38ish, but does my description above make sense? should i wiki it? 01:46:53 wait, i've been welcomed 15 times 01:46:54 ? 01:46:58 Bike: yes 01:47:11 quintopia: it needs to be explained in a bit more detail, I think 01:47:38 `run grep -h ': Welcome to the' /var/irclogs/_esoteric/20??-??-??.txt | cut -f4 -d':' | cut -f2 -d'>' | sort | uniq -c | sort -rn | grep -v '!' | xargs 01:47:43 xargs: unmatched single quote; by default quotes are special to xargs unless you use the -0 option \ 22 fdhsfdsgdsg 19 oerjan 16 elliott 15 Bike 13 monqy 9 shachaf 8 ais523 6 `run grep -h 01:47:56 bleh 01:48:09 ais523\unfoog: so it doesn't make sense? 01:48:14 `run grep -h ': Welcome to the' /var/irclogs/_esoteric/20??-??-??.txt | cut -f4 -d':' | cut -f2 -d'>' | sort | uniq -c | sort -rn | grep -v '!' | grep -v '`' | xargs 01:48:17 it almost makes sense 01:48:20 22 fdhsfdsgdsg 19 oerjan 16 elliott 15 Bike 13 monqy 9 shachaf 8 ais523 6 HackEgo 5 Taneb 5 sivoais 5 kmc 5 fungot 5 Fiora 4 zzo38 4 shubshub 4 Sgeo 4 Gracenotes 4 CHeReP 4 carado 4 andrew12 3 WeThePeople 3 tswett 3 ThePope 3 test 3 Snowyowl 3 sirdancealot 3 SirCmpwn 3 segorev 3 Regis_ 3 ogrom 3 mean 3 john_metcalf 3 Jafet 3 hi 3 GreyKnight 3 Grego 01:48:25 there, that's what I wanted 01:48:36 sorry about all the pings, everyone 01:49:02 ais523\unfoog: can you add background coloration to each number/name pair to make it easier to group them mentally? 01:49:15 quintopia: no, because my client doesn't support background colors 01:49:41 ais523\unfoog: colors are cool. lrn2bcolorful. 01:50:00 it does support foreground colors except I disabled them 01:50:15 which lead to this bug report: https://bugs.kde.org/show_bug.cgi?id=317034 01:50:32 you could be coolour with more colour. 01:51:50 -!- nisstyre has quit (Ping timeout: 264 seconds). 01:52:09 ais523\unfoog: do you know of a way to compare two positive numbers in unbounded-cell BF using less than 5 cells? 01:52:39 quintopia: does the interp support negative numbers? 01:52:47 but yeah, I think regardless, it's doable 01:52:47 sure 01:53:14 i can't think of a way to do it in less than 5 01:53:15 or, hmm 01:53:24 yeah, the way I was thinking of takes 5, if you include the two numbers themselves 01:53:37 what way are you thinking? 01:53:46 http://esolangs.org/wiki/Brainfuck_algorithms#z_.3D_x_.E2.89.A5_y 01:53:57 oh, hmm, that needs a wrapping impl 01:54:24 hadn't looked at that algo in a while 01:54:51 you can do it in 2 if you don't mind an infinite loop if one particular number is higher 01:54:56 haha 01:54:58 and don't mind destroying the original number 01:55:01 *numbers 01:55:23 let's call that "might compare" :P 01:55:58 I bet it's doable in 3 if they're all consecutive and you don't mind the algo destroying the original numbers and returning using the position of the tape pointer 01:56:20 show me 01:56:52 the method i'm thinking of requires both those conditions and still takes 5 01:59:10 -!- Sprocklem has joined. 02:00:09 actually no, that takes 4 02:00:14 put a zero either side of the two numbers you're comparing 02:00:32 then from the first number, do [->-[>]<<] 02:01:12 the two zeros never change, though 02:02:02 come to think of it, that looks like the fastest compare algo I've ever seen 02:02:34 it runs in O(a-b) if a is larger, O(a) if a is smaller, and the constant's pretty low too 02:03:11 if you had an if rather than a while, you could do it in 3 02:03:20 by changing the [>] to an if[>] 02:05:09 that's better than what i had, but the basic idea is the same. 02:05:38 but how do you do a conditional on the result? 02:05:44 without using an extra cell? 02:06:42 -!- nisstyre has joined. 02:07:26 go right one cell 02:07:29 then check for zero 02:07:50 that leaves you still uncertain where the pointer is if you find that it's zero 02:08:00 because it'll be in a different position for equal and for less than 02:08:06 but it does technically fulfil the conditions 02:09:35 no 02:09:46 one cell to the right will always be nonzero unless x=y 02:10:26 if x>y, then one cell to the right is x, and nonzero 02:10:35 if x=y then one cell to the right is x, and zero 02:10:46 if x err, no 02:10:51 it is nonzero :( 02:10:53 you're right 02:11:02 hmm 02:12:10 the problem is it leaves either x or y at zero 02:12:23 and the pointer to the left of the other one 02:12:36 i think you never test the cell to the left of x 02:12:42 so why not initialize it to 1 02:12:48 you do 02:12:53 it's tested in the loop termination 02:12:56 ah right 02:12:59 when x>y 02:13:19 what would happen if you initialized it to 1 02:13:46 you end up back there with it zero 02:13:47 you'd go round the loop again, it'd become 0, x would be tested for equality to 1 02:13:59 depending on the result you'd either end up there or to the left 02:14:09 if you know that x-y isn't 1 it would work 02:14:16 what would happen if you initialized it to -1 02:15:06 then it'd keep being decremented until x became 0, then you'd end up on the cell to the left 02:15:37 yep 02:15:39 -!- shikhin has joined. 02:16:21 which makes 5 cells :) 02:16:47 what if you replace the << with <[<]>? same algorithm basically? 02:17:08 oh no 02:17:09 nvm 02:17:30 that acts differently in the x=0 case 02:17:34 err, x=y 02:17:45 it also acts differently if y is less :P 02:18:07 ah i know 02:19:20 oh yeah, it just continues looping 02:20:21 [>->-[>]<<<->[<]+] or something like that. starting with the left cell at 1. 02:21:17 that may not terminate :P 02:21:26 idea 02:21:35 wait, no 02:21:50 ah yes 02:22:01 start the left cell at 1 and do [>-[>]<<-] 02:22:26 not sure if that helps though :P 02:22:47 huh? i'm not sure what that does... 02:22:58 did you start it at x or at 1 02:23:31 I started it at the x 02:23:34 ah 02:23:40 makes more sense 02:24:38 looks like it runs off the left edge printing -1s 02:24:49 no it doesn't 02:24:55 * quintopia looks again 02:25:07 if x is smaller or equal, it ends when x is set to 0 02:25:16 if x is larger, it moves to the 1, sets it to 0, and exits 02:25:40 ah i see 02:25:41 so you end up at (0) x-y 0 0, or at 1 (0) y-x 0 02:25:56 sadly those still can't be distinguished without extra cells 02:26:11 also there's an off by one in there somewhere I think 02:26:38 yeah 02:26:51 so it misses x=y 02:27:16 i'll just stick with my five cell solution then :) 02:28:38 it feels like it /should/ be possible :) 02:32:38 -!- shikhin has quit (Ping timeout: 264 seconds). 02:35:03 i think i might have a way. it's pretty damn slow, but i think it can be done in 4 02:35:21 it's long so i'll just describe 02:36:08 start by turning x,y,0,0 into 0,0,x-y,y-x in the obvious way 02:36:57 then, increment the second cellwait no 02:37:02 forget y-x 02:37:06 * Bike writes this down 02:37:12 you can't forget y-x 02:37:24 there's no test for negative/positive in nonwrapping BF without loads of temporaries 02:37:27 because you have to keep taking it past 0 02:37:29 wiggle clear style 02:37:35 err, past its original value 02:37:38 that's what i was going to do 02:37:41 wow 02:37:45 test the sign of x-y 02:37:54 by going down 1, up 2, down 3, up 4 etc. 02:38:00 well you have three temporaries to do that 02:38:03 but I'm not sure that's enough 02:38:11 using the two temps as indices, used alternately 02:40:27 so one round increments temp0 temp1 and test until temp0 (known to be negative) is zero, then tests the test value. then temp1 is incremented and the next round decrements temp0 temp1 and test until temp1 (known to be positive) is zero, then tests, then temp0 is decremented and repeat 02:40:33 uses only 3 cells 02:41:06 which round you get test=0 in (even or odd) gives you test's sign 02:42:38 -!- utopian has quit (Quit: Lost terminal). 02:44:07 (i'll still use the five-cell solution for being blindingly faster) 02:44:12 hmm 02:44:17 what's that in code 02:44:26 let's start with x-y temp0 temp1 02:44:38 x-y (temp0) temp1 02:44:39 I think you have problems leaving the loop 02:44:45 nah you have to start on x-y 02:44:54 to test for zero 02:45:17 the test is "which round am i on" so it doesn't matter where you start 02:45:34 I'm not going to believe this doesn't need a fourth cell until you write it out 02:45:47 okay give me a few 02:46:14 Is it wrong that I was expecting Prolog to be a bit saner than the List monad? 02:46:58 yes, turn yourself in 02:47:31 Prolog's more /interesting/ than the List monad :) 02:47:47 actually, Prolog has two main features 02:48:03 nondeterminism, which the List monad does in a different and arguably saner way 02:48:10 and unification, which the List monad doesn't do at all 02:49:06 Then you would need to make up the unification monad too 02:49:15 get on it 02:49:22 I want Prolog to be more like the Logic monad 02:49:33 Or, well, using the non-monady bits like >>- iirc 02:49:34 is that a real monad 02:49:45 http://hackage.haskell.org/package/logict-0.2.3/docs/Control-Monad-Logic.html 02:49:47 I think the Logic monad is equivalent to the list monad as far as I can tell? 02:50:20 I think the monad is equivalent but it also provides variations of bind and other things that make it... backtrack differently 02:50:44 I think 02:50:59 No, that is a class, which is provided to list monad too 02:51:38 But there is LogicT which may be useful in some way as a kind of list monad transformer that is correct 02:51:43 > do { x <- [1..]; y <- [2..]; guard (x == 2 && y == 2); return "Hi" } 02:51:50 mueval-core: Time limit exceeded 02:52:40 but the result is of the correct type! 02:56:00 > [1..] >>- \x -> [2..] >>- \y -> do guard (x == 2 && y == 2); return "Hi" 02:56:01 Not in scope: `>>-' 02:56:01 Perhaps you meant one of these: 02:56:01 `>>' (imported from... 02:56:07 eep 02:56:14 elliott!!! 02:56:35 oerjan: (>>-)? 02:56:40 Is that from semigroupoids? 02:56:55 shachaf: no from logict 02:57:24 oerjan: I don't think (Num a) => [a] is a (Num a) => Logic a, is it? Although I'd expect a different error 02:57:40 http://hackage.haskell.org/package/semigroupoids-4.0/docs/Data-Functor-Bind.html has it too 02:57:56 Sgeo: istr >>- worked just find with lists in previous versions of lambdabot 02:58:00 *fine 02:58:48 at least we were having fun with it 03:00:09 :t (Control.Monad.Logic.>>-) 03:00:12 Control.Monad.Logic.Class.MonadLogic m => m a -> (a -> m b) -> m b 03:00:15 oh. 03:00:45 > [1..] Control.Monad.Logic.>>- \x -> [2..] Control.Monad.Logic.>>- \y -> do guard (x == 2 && y == 2); return "Hi" 03:00:46 Not in scope: `Control.Monad.Logic.>>-' 03:00:46 Perhaps you meant one of these: 03:00:46 ... 03:00:52 wat 03:00:57 elliott!!!!!!!!!!!!!!!!!!!!!!! 03:01:08 :t (>>-) 03:01:09 Not in scope: `>>-' 03:01:09 Perhaps you meant one of these: 03:01:09 `>>' (imported from Control.Monad.Writer), 03:02:17 @let import Control.Monad.Logic 03:02:18 Defined. 03:02:51 huh 03:03:07 > [1..] >>- \x -> [2..] >>- \y -> do guard (x == 2 && y == 2); return "Hi" 03:03:09 *Exception: stack overflow 03:03:13 OKAY 03:03:17 AN IMPROBEMENT 03:04:21 > take 1 $ [1..] >>- \x -> [2..] >>- \y -> do guard (x == 2 && y == 2); return "Hi" 03:04:23 *Exception: stack overflow 03:04:28 fff 03:05:49 > take 1 $ [1..] >>- \x -> do guard (x == 2); return "Hi" 03:05:51 ["Hi"] 03:05:56 YAY 03:06:02 SO POWEFRYL 03:06:44 -!- shikhin has joined. 03:07:17 how illogical 03:08:05 > [1..] >>- \x -> [2..] >>- \y -> [(x,y)] 03:08:07 [(1,2),(2,2),(1,3),(3,2),(1,4),(2,3),(1,5),(4,2),(1,6),(2,4),(1,7),(3,3),(1... 03:08:24 oerjan: someone (Martin Escardo) at Birmingham University's looking into how you can search infinite spaces in finite time 03:08:26 > [1..] >>- \x -> [2..] >>- \y -> ["hi"] 03:08:28 ["hi","hi","hi","hi","hi","hi","hi","hi","hi","hi","hi","hi","hi","hi","hi"... 03:08:37 his research probably would allow you to filter a list of all integers to see which were 2 03:09:16 -!- shikhin_ has joined. 03:09:17 ais523\unfoog: i recall some haskell discussion about using bottom checking for that. 03:09:35 doesn't particularly work for integers, though. 03:10:00 How worthless is the one million dollar wedge? I do not expect it to be worth much. 03:10:14 i'm going to guess it's worth less than one million one dollars 03:11:08 -!- shikhin has quit (Ping timeout: 240 seconds). 03:11:22 Bike: let's not jump to conclusions here 03:11:47 > [1..] >>- \x -> [2..] >>- \y -> [] 03:11:49 *Exception: stack overflow 03:12:24 cool 03:12:39 It may be worth as much as a lottery card. 03:12:58 -!- lexande has joined. 03:15:01 -!- shikhin_ has quit (Ping timeout: 245 seconds). 03:30:42 Prolog strikes me as the sort of language that probably shouldn't have unrestricted side effects, like a lazy language probably shouldn't have unrestricted side effects 03:31:07 My AI professor looooooves Prolog 03:31:35 I've never seen someone gush so much about a language, and my PL prof used Ocaml 03:31:41