00:00:01 actually I've had a recurring nightmare since I was little and it's never stopped (as opposed to the more classic nightmares involving monsters) 00:00:09 and there's an automaton in it, or something 00:00:20 and the universe always end up implosing and I CAN'T STOP IT 00:00:25 fiora: Cool article. 00:00:34 reproduction, metabolism and 00:00:39 and whenever I wake up after that recurring automaton nightmare I try to make sense out of it but I've never succeeded 00:01:03 dont remember the last conditions 00:01:27 hagb4rd: oh, you mean the definition of real world life? nah that's a hilarious flamewar 00:01:31 simultor metaphysics? 00:01:50 kmc: the real world is running in a simulator 00:02:00 why would reproduction be necessary to define life? 00:02:07 It’s simulators all the way down. 00:02:10 the hashlife idea happens to be applicable to basically any kind of physics where you have a speed of life, so 00:02:14 speed of light* 00:02:28 hashlife is really cool and bike can explain it 100 times better than I can 00:02:52 i vaguely understand it but would appreciate a good explanation 00:03:36 what is a virus? vira don' t have metabolism.. is virus not a life form? i asked myself recently 00:03:39 ok, well, it's pretty easy. what you need is a space you want to simulate, and the physics have to have a speed of light. 00:03:49 -!- nooga_ has quit (Ping timeout: 248 seconds). 00:03:51 but they have dna 00:04:39 you divide space into areas (in Life, squares of width 2^n for some natural n). 00:04:41 hagb4rd, it depends on what you mean by 'life'. There is no 'reality' saying "This is life, and this isn't". The term 'life' is what we define it to be 00:04:59 yea i mean from the point of the biologists 00:05:00 then for each space, you can find what the center of that space will be in some time, due to the speed of light. 00:05:51 i.e., if you have a 2^8 block in Life, the center 2^4 block can be simulated 2^3 steps ahead, without caring about any other 2^8 block, because the speed of light is 1 square per step. 00:06:27 neat 00:07:01 for representing a block of space, rather than tracking the actual Life bitmap or whatever, you just have each block be five pointers (in Life anyway), each to a block half the size of the block. 00:07:08 four quadrants + a cached future center 00:07:16 Hashlife adds to this a hash map of subspaces. 00:07:30 (as a cache, of course) 00:07:52 Koen___: http://xkcd.com/505/ hth 00:08:00 Yeah, the idea is, if you need a new space for whatever reason, you make it out of its linked subspaces, and if those subspaces have been combined before, you can just use that same previous combination. 00:08:04 which might have its future cached. 00:08:09 It turns out that Life tends to have many common subspaces, so doing this lets you skip a lot of combination via memoization. 00:08:14 Erm, computation. 00:08:39 what if i'm simulating life with SKI pikhq, what then 00:08:55 Then you are memoizing combinations as well. :P 00:10:28 aw..you're talking about the game of life 00:10:29 oerjan: yeah that was a really good one 00:11:03 well, if you want real life, virology is pretty interesting 00:11:19 there are viruses that have more protein construction mechanisms than some bacteria 00:11:20 i won't insist but yea 00:11:22 :P 00:11:38 -!- Jafet has quit (Read error: Connection reset by peer). 00:11:41 "mamavirus" because what is even your deal, virologists 00:11:43 -!- nooga_ has joined. 00:12:27 -!- Jafet has joined. 00:12:51 also http://rybicki.wordpress.com/2008/03/19/from-what-did-viruses-evolve-or-how-did-they-initially-arise/ 00:13:01 thx 00:13:21 the idea that viruses have come about originally several times is a pretty neat one, since it's not the case for life in general 00:14:07 i understand it's still a controversial topic 00:14:51 well nobody really understands what the fuck happened with abiogenesis, so there's that. 00:18:34 Gotta love things where we just go "well, it clearly happened" 00:20:20 there's RNA world and clay world and all as hypotheses that are half-plausible but they're still pretty out there 00:20:41 panspermia 00:20:47 for sure 00:20:51 hail the mighty penis in the sky 00:20:55 main that ain't a real solution >: 00:21:09 main ain't usually a function 00:21:26 man* 00:22:01 yeah it isn't 00:22:19 it would be nice if we could at least conclude it, though. 00:23:47 http://en.wikipedia.org/wiki/Category:Borderline_life how useful 00:24:06 neat 00:26:17 -!- nooga_ has quit (Ping timeout: 246 seconds). 00:26:40 borderline useful 00:26:53 viroids are cool and that's that man 00:43:09 Fiyora 00:43:11 Hike 00:43:18 hiii 00:43:20 hatchetaf 00:43:44 um...................the ch is a fricative........................................................... 00:43:57 That sounds like nerd stuff. 00:44:32 so do you hth 00:44:38 :< 00:47:27 -!- Koen___ has quit (Quit: Koen___). 00:47:55 hey Bike write me a limerick 00:48:17 there once was a man from nantucket, whose cock was so long he could suck it, 00:48:52 he said write, not quote, Bike 00:48:55 a good limerick plz thx 00:55:52 Fiora: Do you know about de bruijn sequences? 00:58:35 um... they're the things you can use to do leading-zero count with bit hackiness, right? 01:00:23 Trailing, I thought? Possibly both. Anyway that's one thing you can do with them. 01:00:26 oh that thing that's in sanskrit for some reason. 01:00:26 Also lots of other things. De Bruijn was the best. 01:02:09 other things? I remember looking at the wikipedia article and being a bit confused 01:02:45 Like magic! 01:02:55 ? 01:03:18 i think i am slowly learning how to pronounce 'shachaf' 01:03:37 And, uh, let's say you have a keypad-combination-lock with no termination digit. You can use a de bruijn sequence to figure out the most efficient way to brute-force it! 01:03:47 (Maybe it's not called brute-force at that point...) 01:04:09 I went to a talk which mentioned them. 01:04:25 it's hebrew isn't it? and if i've learned anything from Rugrats it's that i can't pronounce hebrew "ch" 01:04:33 we had a lot of locks like that in college, and people did bruteforce them 01:04:37 שחף hth 01:04:44 except I think they weren't precisely with-no-termination-digit 01:04:59 シャチャフ? 01:05:08 i fucking love that first kana 01:05:24 It looks like a sideways smiley XD 01:05:40 the combination lock thing is neat though 01:05:47 Fiora: I doubt it. The "ch" is as in the recording on https://en.wikipedia.org/wiki/Voiceless_uvular_fricative 01:06:20 oh. darn. I can't kana that 01:06:34 we clearly need hebrew with a japanese alphabet 01:07:15 The effect was something along these lines: He handed out a deck of cards, told a few people to cut it, and then had 5 people take the top card and pass the deck to the person on their right. 01:07:30 I have a lot of trouble with the phonemes that look like "x" in ipa 01:07:37 Then he said "I have a strong sense of red... Stand up if you have a red card". 01:07:42 Then he told everyon what their card was. 01:07:47 Nice. 01:07:56 If you know about de bruijn sequences you can figure out how it works! 01:08:01 I think the Japanese approximation whould be "shahhafu". 01:08:13 シャハフ 01:08:21 (I still like the シ) 01:08:22 There'd be a sokuon there. 01:08:37 it's just so damn happy to be a character 01:08:40 makes sense... doubled-h is kind of a weird thing in japanese 01:09:07 it is, but that's how it's done. 01:09:07 doesn't japanese have "doubled consonants" as a usual thing? 01:09:14 Bach is バッハ 01:09:23 Bike: That particular consonant being doubled is unusual. 01:09:29 oh. 01:09:51 Though not *impossible*. 01:09:57 Admittedly only a thing in loan words. 01:10:37 What are other good applications of de bruijn sequences? 01:11:08 brute forcing code locks? 01:11:23 Maybe if you want to be able to sample all possible sequences of a certain length easily? 01:11:53 olsner: I gave that one. 01:12:22 apparently it has applications in sanskrit prosody 01:12:38 wondering if you could use it in DNA somehow, except that proteins are variably length coded to say the least 01:13:03 I guess de bruijn sequences are related to shift registers somehow? 01:14:57 -!- carado has quit (Ping timeout: 256 seconds). 01:15:24 -!- oerjan has quit (Quit: Ribbit). 01:15:41 -!- mnoqy has joined. 01:26:39 how do you generate de bruijn sequences 01:27:34 http://en.wikipedia.org/wiki/De_Bruijn_sequence#Algorithm something like this I guess? 01:28:31 I wonder if that's really the best way 01:29:03 the problem with that page is that i try to find a euler path in the pictured graph >_> 01:29:26 euler path? 01:29:44 Going to work on Rebol deadfish 01:30:07 Fiora: hit every edge exactly once 01:30:57 well it's supposed to be a cycle also. 01:30:57 And I can't even figure out how to read just one character from the console 01:31:02 Erm, from stdin 01:31:17 Bike: oh, so it's like that puzzle about crossing each bridge once 01:31:26 Fiora: exactly the same puzzle, even 01:31:27 yep 01:31:29 I guess I could try reading the whole thing in, and hoping that it doesn't actually store the whole thing in memory 01:31:31 they're easy to find, unlikel hamiltonian paths 01:31:36 http://en.wikipedia.org/wiki/Seven_Bridges_of_K%C3%B6nigsberg 01:31:37 "de Bruijn sequences can be generated by feedback shift registers" according to http://mathworld.wolfram.com/deBruijnSequence.html 01:31:38 (which visit every vertex once) 01:31:50 i knew it hth 01:32:15 kmc, which makes more sense, a Racket version of Qoppa, or a Rebol version of Qoppa? Racket is more Lispy, but there's less of an impedence mismatch between Rebol and Qoppa than Racket and Qoppa 01:32:45 i don't know 01:33:33 i do not understand the use of "impedence mismatch" here 01:35:19 Racket is a rather static language. Modules are expected to know what values they can export, for example 01:35:55 At ... a point in static analysis... hmm, suddenly I'm not sure whether that might be the case for various proposed module systems for Rebol 01:37:30 At any rate, right now I should be writing some documentation for work 01:37:35 Told my boss I'd do that when I get home 01:39:51 I think Qoppa is closer to Rebol than Racket conceptually 01:40:52 you could also implement Kernel instead of my weird bastardized version of Kernel 01:41:04 but cyclic lists are /so annoying/ kmc 01:41:13 a lot of the changes are just for the sake of the self-hosting trick 01:41:22 in a sense the Qoppa interpreter I posted is a polyglot program 01:41:39 kmc, but didn't you unbastardize the bizarre distinction between functions and operatives? 01:41:43 kinda 01:41:52 but I guess there are some unsavory consequences too 01:42:12 Hmm, such as? 01:44:49 i still don't really get why you all think that distinction is weird 01:45:40 Sgeo: i don't remember, it's in the blog post 01:46:21 Bike: it's just weird because fexprs let you remove so many cases from the evaluator, but then you end up with two cases that are almost the same 01:47:31 what two cases? 01:47:40 for operatives and applicatives 01:47:52 combine applicative combinand env = combine (unwrap applicative) (map (eval env) combinand) env 01:47:52 in Kernel they are distinct fundamental data types 01:47:59 or do i not understand what cases you're talking about 01:48:22 that's a case yes 01:48:22 applicatives aren't really fundamental, they're just a wrapper around operatives 01:48:37 in Kernel they are presented as fundamental 01:48:41 that's what Sgeo and I think is weird 01:49:42 "an applicative is nothing more than a wrapper to induce operand evaluation" doesn't seem that fundamental 01:49:47 and i don't get how the cases are the same? 01:50:06 Bike, the fact that it's physically different than just putting a wrapper around an operand 01:50:16 "physically different"? 01:50:16 -!- Phantom__Hoover has quit (Ping timeout: 246 seconds). 01:50:33 As in, there's an actual distinction, at the language level, saying "these are applicatives", and "these are operatives" 01:50:45 okay, they're not really the same because you can implement one in terms of the other, but it's one more case and one more fundamental data type than I would like to have. 01:50:55 Rather than "everything is an operatives, and some operatives happen to have a wrapper around them that evaluates their arguments when called" 01:51:19 it's just weird that "things which happen to evaluate all their args first" are seen as fundamentally different from "things which may or may not evaluate all their args first" 01:51:32 hm, maybe i'm thinking of it differently since i've been working maru style. 01:51:37 maru style? 01:52:09 you could have generic functions, for example, and use "wrap" on them to get them to evaluate their arguments, even as generic functions have a totally different combination mechanism from operatives 01:52:35 and the applicative path of the evaluator doesn't have to have any idea that the thing it's unwrapping isn't a regular operative 02:00:21 -!- madbr has joined. 02:12:52 @pl \x y -> all ((flip elem) y) x 02:12:53 flip (all . flip elem) 02:13:12 how's @pl work? 02:14:44 it uses Haskell 02:14:45 ) 13 : 'y + y' 02:14:46 Sgeo: + 02:14:59 ) 13 : 'x + x + y + y' 02:15:00 Sgeo: [ + [ + ] + ] 02:16:14 ) 13 : 'x + y + x + y' 02:16:14 Sgeo: [ + ] + + 02:16:20 thx olsner 02:19:44 Phantom_Hoovlogreader: o.O she's named Zhaan? I thought it was Xan! 02:21:23 "What's the most complex thing you do in your kitchen?" "Worry about death." 02:22:01 deep 02:22:21 haha 02:25:09 -!- Bike has quit (Ping timeout: 248 seconds). 02:27:09 -!- Bike has joined. 02:32:43 @showIntAtBase 2 intToDigit 23 "" 02:32:43 Unknown command, try @list 02:33:35 > showIntAtBase 2 (intToDigit 23 "") 02:33:36 The function `GHC.Show.intToDigit' is applied to two arguments, 02:33:37 but its ty... 02:33:48 w/e 02:34:07 > showIntAtBase 2 intToDigit 23 02:34:11 *Exception: show: No overloading for function 02:34:23 @showIntAtBase 2 intToDigit 23 "" 02:34:24 Unknown command, try @list 02:34:28 > showIntAtBase 2 intToDigit 23 02:34:30 *Exception: show: No overloading for function 02:34:33 @more 02:34:46 @lore 02:34:55 @diediedie 02:34:55 Unknown command, try @list 02:35:31 > showIntAtBase 2 intToDigit 23 "" 02:35:34 "10111" 02:36:31 > showIntAtBase 2 intToDigit 23 "bla" 02:36:32 "10111bla" 02:36:40 i'm told that false morels (Gyromitra esculenta) are sold in Finland and eaten after careful preparation 02:37:14 weird brain-looking mushroom. http://en.wikipedia.org/wiki/File:Fr%C3%BChjahrslorchel.JPG 02:37:37 i really don't understand why there is that 2nd argument..does that make any sense? 02:38:07 That mushroom looks like a funnel cake 02:38:46 > intToDigit 15 02:38:48 'f' 02:38:52 > intToDigit 35 02:38:54 *Exception: Char.intToDigit: not a digit 35 02:38:58 > intToDigit 16 02:39:00 *Exception: Char.intToDigit: not a digit 16 02:39:10 anyway it lets you use whatever alphabet you want 02:39:30 > showIntAtBase ("-."!!) intToDigit 23 "" 02:39:31 No instance for (GHC.Real.Integral 02:39:31 (GHC.Types.Int -> GH... 02:39:42 er 02:39:42 > showIntAtBase 2 ("-."!!) 23 "" 02:39:44 ".-..." 02:39:58 > showIntAtBase 2 intToDigit ff 02:40:00 Not in scope: `ff' 02:40:01 Perhaps you meant one of these: 02:40:01 `f' (imported from D... 02:40:30 nah 02:42:00 i'd like to somehow implement a little tool throwing 'i-ching' hexagrams 02:42:23 ~yi 02:42:31 dammit. 02:42:37 well, ~yi does that, if metasepia's around. 02:42:53 oh cool.. 02:43:09 Starting to see some flaws in the Rebol model 02:43:15 http://en.wikipedia.org/wiki/Category:Fungi_of_Finland 02:43:29 For instance, it's difficult to write a thing that puts all setting into a new scope, but leaves others untouched 02:44:03 If you don't look inside nested code, then you can't do sets inside control structures. But if you do, you could get odd behavior if you try to use the same function inside that thing 02:44:28 fwiw, what Rebol does with its context function is not look inside nested blocks 02:44:50 thanks for that hint byte 02:45:08 bike 02:45:13 :P 02:45:50 Maybe an example 02:46:37 >> a: 5 context [a: 6 print a] print a 02:46:41 That prints 6 then 5 02:46:59 a: 5 context [print a a: 6] print a 02:47:02 prints none then 5 02:47:08 Nothing too bizarre so far 02:48:03 a: 5 context [if 0 < 1 [a: 6] print a] print a 02:48:06 Prints two 6s 02:56:07 There's a hack that can be done to do varargs in Rebol 3. Apparently it's slated for removal 03:05:20 And I found a hack that doesn't need that hack 03:12:10 -!- mnoqy has quit (Quit: hello). 03:47:09 -!- mnoqy has joined. 05:15:49 -!- madbr has quit (Ping timeout: 248 seconds). 05:27:52 -!- Nisstyre has quit (Quit: Leaving). 05:56:23 -!- mnoqy has quit (Quit: hello). 06:22:05 -!- zamlierza has joined. 06:22:08 I'm currently working on my first compiler, for a lisp-like language, written in Common Lisp (sbcl). Recently I've been playing with Racket. Is anyone else here using one of the lisps? 06:22:33 esoteric lisp? 06:24:05 Bike: Sure 06:24:37 i'm just wondering why you're asking here 06:30:34 `WeLcOmE zamlierza 06:30:40 ZaMlIeRzA: 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. (FoR ThE OtHeR KiNd oF EsOtErIcA, tRy #EsOtErIc oN IrC.DaL.NeT.) 06:31:36 -!- Nisstyre-laptop has joined. 06:36:06 -!- FreeFull has quit. 06:45:07 -!- Frooxius has quit (Read error: Connection reset by peer). 06:45:15 -!- Frooxius_ has joined. 06:48:18 kmc: lens has a good intatbase thing 06:48:45 > "hello" ^? base 23 06:48:47 Nothing 06:48:48 > "hello" ^? base 28 06:48:49 Just 10773556 06:49:21 > base 36 # 12344321 06:49:23 "7ckxt" 06:49:27 :t (^?) 06:49:29 s -> Getting (First a) s a -> Maybe a 06:49:38 :t base 06:49:40 (Integral a, Applicative f, Choice p) => Int -> p a (f a) -> p String (f String) 06:49:54 hi Bike want to learn about lens 06:50:11 what is lens, really? we just don't know 06:50:13 :t (#) 06:50:15 parse error on input `)' 06:50:42 @ty ( # ) -- interferes with a GHC extension for (# ... #) tuple syntax 06:50:43 AReview s t a b -> b -> t 06:50:55 lens is a library for a bunch of things, including lenses. 06:51:11 And prisms, which are colenses. (base n) is a prism. 06:51:32 lens is a getter and a setter, you said. what's a colens. 06:52:20 Well, a colens is to a sum what a lens is to a product. 06:52:26 Do you know about sums and products? 06:52:35 not really. 06:52:54 (a,b) is a product, (Either a b) is a sum. 06:53:05 And all things which behave like tuples or Eithers. 06:53:33 If A has n inhabitants and B has m inhabitants, then (A,B) has n*m and (Either A B) has (n+m) 06:53:43 right 06:53:55 So you see how a lens is for a product? 06:54:03 data Foo = Foo { one :: Int, two :: Char } 06:54:25 You can have a "one" lens which lets you do two things: Read the "one" value, and change the "one" value to some new Int. 06:54:46 Similarly: 06:54:51 data Bar = One Int | Two Char 06:55:01 You can have a "One" prism which also lets you do two things: 06:55:13 The things are Int -> Bar and Bar -> Maybe Int 06:55:17 I'm still back at lenses being products. 06:55:26 Oh, lenses are *for* products. 06:55:32 "for"? 06:55:33 Since Foo is a product type (of Int and Char). 06:55:40 oh. 06:55:44 ok, that makes sense. 06:55:54 Bar is a sum type of Int and Char. 06:55:59 I'm working on a genetic program for the creation and optimisation of ciphers, whilst the optimisations still in process here're some early results: http://pastebin.com/GKwfKENX http://pastebin.com/FQdMNqFZ 06:56:12 data Bool = False | True; data Color = Red | Green | Blue 06:56:38 Actually, never mind. 06:56:46 Anyway, which part is confusing right now? 06:56:56 zamlierza: the second one is admirably hard to read 06:57:10 Bike: That is the fun of this channel! 06:57:11 shachaf: nothing in particular, that made sense. i'm just sleepy. 06:57:23 Good! 06:57:37 A traversal is like a lens onto more than one value. 06:57:46 Well, onto zero or more values. 06:57:47 So what's base operating on? or ok something else 06:58:19 Oh, base is on a "conceptual" sum, not an actual sum type. 06:58:29 well yeah. 06:58:34 but what's the concept. 06:58:58 Well, base is a prism. 06:59:08 The two prism functions are called review and preview 06:59:33 A string is sort of like a sum: Either it has a valid base-N value or an invalid value. 07:00:03 is Either in the Choice class? 07:00:08 But given a base-N value you can always turn it into a string. 07:00:21 I'm thinking of having it automatically encrypt an input of all Falses, deducting abs(sum(map(lambda x: 1 if >= 0 else -1,output))) this will ensure precedence to even distribution, seems to be the simplest part. For the main challenge; avoiding recurrent patterns, I'll be using something along the lines of. http://pastebin.com/5LvxG01V 07:00:21 No, Choice is something more complicated. 07:00:28 It has to do with the representation of prisms lens uses. 07:00:30 right, of course. 07:00:42 obviously accompanied by a similar function for handling lists of recurring True's or False's 07:00:43 :t base 36 07:00:44 (Integral a, Applicative f, Choice p) => p a (f a) -> p String (f String) 07:00:54 I'll upload what I've got so far if anyones got any idea's, otherwise I'll wait until it's at a functioning point. 07:00:56 I can explain that representation but that should be a separate thing. 07:01:06 yeah, don't let me derail you 07:01:24 What's derailing me is multisecond latency from my computer to my IRC client. 07:01:42 what's derailing me is multisecond latency from my brain to my blinking 07:01:46 obviously accompanied by a similar function for handling lists of recurring True's or False's 07:02:05 http://pastebin.com/J8Aer7Rx I'm going with that for now. 07:02:12 :-( 07:02:21 It's not genetic yet but I'll build from this: http://pastebin.com/E4dGw97L 07:02:26 So close, zamlierza. Socloszamlierza. 07:02:30 man, paste is not doing that indentation good 07:03:21 Bike: Am I going on and talking about the representation of lenses? 07:03:28 Thoughts? :D 07:03:38 is it supposed to have a huge amount of copy-pasted code? 07:04:20 shachaf: if you go on i'll listen. you had explained how a string can be interpreted wrt numerical bases as a prism, and were about to explain review and preview maybe. 07:04:34 > review _Left "hello" 07:04:36 Left "hello" 07:04:44 > preview _Left (Left "hi") 07:04:45 Just "hi" 07:04:47 > preview _Left (Right "hi") 07:04:48 Nothing 07:04:57 That's all there is to them. They take the two components of a prism. 07:05:23 i see. 07:05:26 :t review 07:05:27 MonadReader b m => AReview s t a b -> m t 07:05:32 no dont.......... 07:05:37 :t _Left 07:05:38 (Applicative f, Choice p) => p a (f b) -> p (Either a c) (f (Either b c)) 07:05:55 :t preview 07:05:56 MonadReader s m => Getting (First a) s a -> m (Maybe a) 07:05:57 _Left :: Prism (Either a c) (Either b c) a b 07:06:05 i remember y'all complaining about whoever's bizarre type signatures in lens 07:06:17 Um, the Prism ones are kind of my fault. 07:06:30 (To be fair it was even worse back in the day.) 07:06:37 you bastard 07:07:11 do you know the following things: mapM; Applicative; Traversable 07:08:12 Uh, the first two. 07:08:19 OK. 07:08:35 Traversable is just a generalization of mapM with Applicative. 07:08:49 You know how mapM could get away with just requiring Applicative instead of Monad? 07:09:15 yeah. 07:09:42 OK, so let's pretend mapM uses Applicative instead of Monad from now on. 07:09:43 Do you know Functor? 07:09:59 traverse generalizes mapM the same way fmap generalizes map 07:10:43 ok. 07:11:34 OK. 07:11:49 So let's define some functions. 07:12:08 one :: (a -> b) -> (a,e) -> (b,e) 07:12:18 two :: (a -> b) -> (e,a) -> (e,b) 07:12:25 both :: (a -> b) -> (a,a) -> (b,b) 07:12:31 Good functions, right? 07:12:36 p. good 07:13:07 These are sort of like fmap, except instead of using a type class to decide what to map over, we just name a function explicitly. 07:14:20 So these are "map" functions -- they let us modify a value/values -- but not "mapM" functions -- we can't have effects. 07:14:43 But we can generalize them in a pretty simple way to be "mapM" functions. 07:14:50 Can you generalize them for me? 07:15:48 one :: Monad m => (a -> m b) -> (a, e) -> m (b, e) maybe 07:15:50 probably not 07:16:12 Yep. 07:16:27 Except you don't need Monad. You can manage with Applicative, or even just Functor. 07:16:30 yep not, or yep yep 07:16:34 ok. 07:16:35 Yep yep. 07:16:48 the thing where monads aren't functors seems sillier and sillier. 07:16:54 Yep yep yep. 07:17:32