00:00:02 I'm pretty sure that I can't use map there though 00:00:11 do? 00:00:13 what's that about 00:00:20 why is that there 00:00:22 monqy, if I find it unnecessary, I'd remove it 00:00:28 At the end of this writing 00:03:30 Need imports 00:03:53 Don't care about imports 00:03:58 Show? 00:04:17 I still have no idea if it's typed correctly! 00:04:27 main = getArgs >>= map (openFile ReadMode) >>= map hGetContents >>= sequence >>= putStr 00:04:31 Still working on checking 00:04:47 No, just realized it won't typecheck 00:04:51 Prelude System.Environment System.IO> :t getArgs >>= map (openFile ReadMode) >>= map hGetContents >>= sequence >>= putStr 00:04:52 :1:13: 00:04:52 Couldn't match expected type `IO a0' with actual type `[b0]' 00:04:52 Expected type: [String] -> IO a0 00:04:52 Actual type: [String] -> [b0] 00:04:53 In the return type of a call of `map' 00:04:55 In the second argument of `(>>=)', namely `map (openFile ReadMode)' 00:04:57 Need to concatenate the strings 00:05:07 Oh, blah 00:05:26 I could hack a fix with returns probably 00:05:45 Sgeo: You must print out each file as it's read, btw. You can rely on laziness to do this, but if one file takes half a second to read and another half an hour, then the half-second read should get printed before the half an hour starts. 00:05:50 (Assuming it's earlier on the command-line.) 00:06:04 That should be pretty easy. 00:06:53 Let me work on getting this to type first? 00:07:31 I'm just telling you how the final program needs to work. 00:08:10 Blargh 00:08:19 My return hack didn't wor 00:08:39 Your implementation of cat has hacks? 00:08:42 I'm going to use full do-notation first, I think, then work on getting that 00:08:55 so sgeo doesn't know haskell 00:09:44 He just hasn't finished yet, that's all 00:15:05 @hoogle [String] -> String 00:15:05 Prelude unlines :: [String] -> String 00:15:05 Prelude unwords :: [String] -> String 00:15:05 Data.List unlines :: [String] -> String 00:15:16 Sgeo: concat. 00:15:26 wouldn't unwords work? 00:15:27 @hoogle concat 00:15:28 Prelude concat :: [[a]] -> [a] 00:15:28 Data.ByteString concat :: [ByteString] -> ByteString 00:15:28 Data.Foldable concat :: Foldable t => t [a] -> [a] 00:15:55 Hmm, I probably should avoid straightforward concatenation at this point 00:16:23 Although I don't see how concat would... oh, derp 00:16:32 ??? 00:16:52 type String = [Char] 00:17:17 ...did I actually forget the syntax for comments in Haskell? 00:18:08 --, {-...-} 00:18:15 You should not need any comments. 00:19:22 @hoogle openFile 00:19:22 System.IO openFile :: FilePath -> IOMode -> IO Handle 00:19:28 Something's lying to me here 00:19:28 Oh 00:19:37 hm? 00:19:50 n/m, must have misread that as IOMode -> FilePath -> IO Handle 00:19:51 before 00:20:05 Easy enough to fix 00:21:07 Although it's ugly 00:22:36 @hoogle (a -> b) -> IO a -> IO b 00:22:36 Prelude fmap :: Functor f => (a -> b) -> f a -> f b 00:22:36 Control.Applicative (<$>) :: Functor f => (a -> b) -> f a -> f b 00:22:36 Control.Monad fmap :: Functor f => (a -> b) -> f a -> f b 00:23:47 WTF am I doing? 00:24:33 fmap. 00:24:39 Or (<$>). 00:25:10 I'm pretty sure I'm using it wrongly here 00:26:52 * Sgeo cries 00:27:52 Sgeo: You're having an awful lot of troubles writing cat. 00:28:06 Would be easier if it was one file 00:28:08 >.> 00:28:25 Maybe I should just make a function that deals with one file, and just... do it repeately 00:32:08 printArg :: String -> IO () 00:32:08 printArg file = openFile file ReadMode >>= hGetContents >>= putStr 00:32:10 That was easy 00:32:52 there are better ways 00:33:48 monqy: Hey, let him write his own pogram. 00:33:49 also are you ever closing those files? 00:33:52 :( 00:33:54 hGetContents closes files. 00:33:58 oh 00:34:59 http://hpaste.org/50411 00:35:17 Seems to work, not 100% certain if it follows the requirement for dealing with large files 00:35:23 But I .. think it should 00:35:32 Sgeo: It does. (It isn't about size, but latency.) 00:35:43 Sgeo: OK, can I ask you to make a simple modification to that program? 00:35:51 Ok... 00:35:52 Before the contents of each file, print the line "--- filename ---" (with no quotes) above. 00:35:53 elliott: without tidying it up first? :( 00:36:01 monqy: He can write it however he likes. 00:36:05 -!- PatashuDragonite has quit (Quit: MSN: Patashu@hotmail.com , Gmail: Patashu0@gmail.com , AIM: Patashu0 , YIM: patashu2 .). 00:38:24 hpaste gives recommendations... 00:38:34 They're just HLint recommendations. Feel free to ignore. 00:38:42 http://hpaste.org/50412 00:39:28 Okay. Here are my solutions (imports Control.Monad and System.Environment): 00:39:29 main = getArgs >>= mapM_ (readFile >=> putStr) 00:39:32 and 00:39:32 main = getArgs >>= mapM_ (\fn -> putBanner fn >> readFile fn >>= putStr) 00:39:33 where putBanner = putStrLn . (++ " ---") . ("--- " ++) 00:39:44 The first one took about thirty seconds, the latter took about the same time to modify. 00:40:28 @hoogle (>=>) 00:40:29 Control.Monad (>=>) :: Monad m => (a -> m b) -> (b -> m c) -> a -> m c 00:42:19 Gregor: not to ask you a question I've already asked you before, but what's a better programming language than Python? 00:42:38 Python has, like, first-class functions and duck typing and "eval". 00:42:39 tswett: Almost every language in existence? 00:42:41 (I am Gregor.) 00:42:50 tswett: Two bugs and a feature? 00:43:20 tswett: and I'm mostly asking about P-languages specifically. Haskell doesn't count. 00:43:36 P-languages? 00:43:47 Python doesn't have nice lambdas 00:44:03 Python has, like, first-class functions and duck typing and "eval". // One poorly-implemented feature, one nonfeature and one irrelevant feature. 00:44:20 Gregor: respectively? 00:44:23 Yes. 00:44:26 tswett: P-languages? 00:44:32 nonfeature? P-languages? help? 00:44:33 elliott: ask Gregor. 00:44:37 Can Ruby even be said to have first-class functions? 00:44:37 It has 1.5th-class functions, as its lambda syntax is a joke. 00:44:38 Gregor: P-languages? 00:44:51 Sgeo: It has lambda, so... yes? 00:44:52 elliott: Those languages that start with P: Perl, Python, PHP, Ruby. Yes, Ruby starts with a P. 00:45:12 tswett: So you are asking "What's a better language than Python in [list of bad languages]?" 00:45:16 Awesome question, man. 00:45:24 It has a lambda that isn't an object unless you put lambda or proc or something in front 00:45:25 It starts with a P, followed by a little backslash, followed by three more letters. 00:45:28 Which is just... ick 00:45:32 ruby lambda is like it makes an object that encapsulates a function doesn't it 00:45:35 Sgeo: lambda {|x| puts "Yep this is a lambda."} 00:45:45 (lambda (x) (display "Yep this is a lambda.") (newline)) 00:45:49 if you actually try using functions without putting them in weirdo objects it doesn't work 00:46:01 where by using I mean treating them like first-class citizens 00:46:12 \x -> putStrLn "Oi ay, lambda" 00:46:13 monqy: No wrapping is involved. 00:46:19 oh? 00:46:20 monqy: It makes the tradeoff that functions apply without needing parentheses. 00:46:27 So it's ambiguous with zero-argument applications. 00:46:37 So you use method(:foo) instead. 00:46:40 Gregor: perhaps my real question is whether or not Ruby is better than Python. 00:46:45 Or [ampersand]:foo in recent Rubies as sugar. 00:46:47 And, if so, whether Ruby is still shit-level. 00:46:50 I have insufficient experience with Ruby to judge. 00:46:55 anyway I dislike ruby 00:46:56 All languages are terrible. 00:46:56 I've used Ruby extensively. 00:47:01 It's worse than Python and shit-level. 00:47:06 I know Ruby has some inheritance from Perl, which is scary ... 00:47:06 Or it's slightly better. 00:47:08 It's hard to tell. 00:47:09 Worse than Python? 00:47:19 But basically it's on the same rung as Python so I'd rather die. 00:47:30 tswett: Out of curiosity, what are your reasons for excluding Haskell by name? 00:47:45 it doesn't start with p 00:47:57 monqy: It starts with P as much as Ruby does. 00:48:07 ruby starts with p okay 00:48:08 Or a letter that is almost like a P 00:48:10 elliott: it doesn't have runtime compiling. Or runtime setting, for that matter. 00:48:19 tswett: Sure it does. 00:48:24 http://hackage.haskell.org/package/hint 00:48:24 what's runhaskell help 00:48:31 tswett: Does this do what you need? 00:48:49 tswett, syntax-wise, Smalltalk is much nicer than Ruby imo 00:48:49 "P-languages" really means "that family of imperative dynamic languages that are all used for approximately the same purpose and came to exist for pragmatic reasons" 00:48:56 unless you want like eval and stuff in which case use one of those funky libraries yeah 00:49:10 Gregor: Haskell can be used as an imperative language trivially; as a dynamic language when required pretty easily. 00:49:20 Gregor: I'm saying this separately from my usual half-joking fanboyishness. 00:49:38 You really can just stuff your entire program in IO if that's convenient :-P 00:49:43 Sure, whatever you say, just face away from me while Haskelljackoffing :P 00:51:27 is there a good p-language 00:51:32 also why do you want a p-language 00:51:54 tswett: I'm not sure what exactly you mean by "dynamic setting", but hint definitely provides the first thing. 00:52:07 With much more useful infrastructure than most languages' "eval", I note. 00:56:30 Gregor: what's your FAVORITE programming language. 00:56:32 Period. 00:58:37 Ruby has a nicer equivalent to C# properties than either C# or Python, imo 00:58:40 Plof 4. 00:58:42 Gregor: amirite 00:58:58 Then again, I forgot what Python's are. 00:59:02 what are c# properties and why do you need them 00:59:27 tswett: So are you gonna answer me? :-P 00:59:47 Code that runs when you try to access something like myobj.name or set myobj.name if myobj.name is a property 01:00:13 elliott: no, it doesn't have this arbitrary feature that, despite being necessary to what I want to do, I haven't actually told you. 01:00:31 feature: not being haskell 01:00:43 monqy: remarkably close. 01:00:48 tswett: Then what's the feature? I would be very surprised if it were not achievable in Haskell, since I just provided eval. 01:00:55 http://msdn.microsoft.com/en-us/library/x9fsa0sw.aspx#Y86 01:01:01 The feature is being able to turn any value into a string of bytes. This is an antifeature of Haskell. 01:01:09 I.e. one of the features of Haskell is that it lacks this feature. 01:01:27 tswett: You can serialise any value with a Data instance. 01:01:42 tswett: And all the standard types have Data instances. 01:01:49 And you can give a Data instance to pretty much any type you want. 01:01:53 (It's the generics infrastructure.) 01:01:56 -!- zzo38 has joined. 01:02:18 * Gregor reappears. 01:02:26 elliott: so -> has a Data instance? 01:02:30 Vs the second example in v 01:02:31 http://www.rubyist.net/~slagell/ruby/accessors.html 01:02:31 Gregor: what's your FAVORITE programming language. Plof 4. // daaaaaaaaamn rite 01:02:32 And IO? 01:02:45 But more to the point, every language is so terrible that that's a difficult question to answer. 01:03:22 My favorite programming language is, of course, Jath. Which is, you know, that programming language I'm going to specify and implement one of these days. 01:03:34 tswett: Serialising (->) is a bit more difficult, but I think you could do it with the GHC API (if you run the code with the functions through the GHC API). 01:03:34 is it a bad language 01:03:36 It's going to be exactly like two of the following: Haskell, Lua, Smalltalk. 01:03:43 IO is just a function, so that'll give you the rest. 01:04:10 haskell is okay. I think I dislike lua. never bothered smalltalk oops. 01:04:13 Gregor: I should really finish off my Plof changes so I can play about with implementing Plof-like languages on Fythe :-P 01:05:04 In fact, all of the goals of Jath could theoretically be achieved by implementing Haskell in Lua and Lua in Smalltalk. But that's a pretty awful solution. 01:05:27 Why do you want it to be like Lua, Lua is awful. 01:05:29 So, what is this GHC API? It sounds like an API to GHC, or something. 01:05:30 Gregor: Though un-Haskell-like, it is entirely feasible to write what amounts to C code in Haskell without too much difficulty 01:05:44 Lua has precisely one useful feature, and that's the setfenv function. 01:06:03 pikhq_: Yes, if you want to be ostracized by both communities. 01:06:14 Your entire program will be in IO, and you'll have enough peek and poke to make anyone sweat, but it's not *hard*. 01:06:19 Gregor: Only one might ostracise you. 01:06:26 Gregor: Plenty of internal code in e.g. bytestring and the like looks like C. 01:06:36 Haskell is a multi-paradigm language. 01:07:05 setfenv looks kind of gross 01:07:16 -!- Vorpal has quit (Ping timeout: 240 seconds). 01:07:16 surely there's a better way 01:07:22 setfenv looks like a great way to break every security guarantee in the world. 01:08:38 setfenv is kind of a gross way to do security, yes. 01:08:56 Though it is... similar... to Unix's security model. 01:09:04 tswett, are you looking to make an online nomic like thing? 01:09:21 Sgeo: yep. 01:09:29 What's wrong with the MOO? 01:09:31 JathNomic. The *only* nomic. 01:09:39 Well... nothing, really. 01:10:15 Ruby... has some security stuff. But supposedly the built-in stuff isn't meant for defending against malicious code 01:10:16 What MOO? 01:10:25 LambdaMOO is pretty good at being a MOO, and I would have difficulty building something better. 01:10:36 elliott: nomic.zbasu.net, a supposedly-codenomic MOO. 01:10:55 tswett: Lemme know how much longer than Normish it lasts 01:10:56 It's not actually a nomic, because I haven't gotten around to writing the part that runs code that people vote in favor of. 01:11:01 Will do. 01:11:38 Also, although Ruby's security stuff could partition untrusted from trusted, it would put all untrusted in the same boat I think 01:17:18 I think there should be a function in Haskell of type overrideIO :: SystemInterface -> IO x -> IO x which allows you to have a SystemInterface type telling all function to override kind of I/O actions and then you can override it and it return the proper I/O action which can be main or whatever. 01:25:27 It's not a bad idea to use continuations in Ruby, is it? 01:25:51 ,addquote 01:27:21 `dadquote 01:27:25 /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: dadquote: not found 01:41:08 * Sgeo decides to try "Ruby Koans" 01:42:03 `querk 01:42:04 /home/hackbot/hackbot.hg/multibot_cmds/lib/limits: line 5: exec: querk: not found 01:43:24 ruby koans? 01:43:32 Sgeo: IM INTERNET FAMOUS BE JEALOUS 01:43:45 help 01:44:04 elliott, hmm 01:44:06 ? 01:44:14 http://contemplatecode.blogspot.com/2011/08/haskell-weekly-news-issue-195.html 01:44:16 FAAAAAAAMOUS 01:45:16 elliott c++ templates guru 01:45:18 elliott, lol 01:45:32 monqy: Yep 01:46:52 what's this trifecta thing 01:47:03 monqy: a good parsing library by edwardk 01:47:15 monqy: you know how clang has like precise error locations with carets and other things pointing to ranges 01:47:24 trifecta gives you that pretty much automagically 01:47:28 mmm 01:48:07 will give it a look next time I need to parse something 01:48:37 I should ask him how fast it is 01:48:55 monqy: oh and it does automatic syntax highlighting?? 01:48:59 so errors get highlighted and stuff... 02:06:35 -!- Nisstyre has changed nick to yourstruly. 02:36:13 http://stackoverflow.com/questions/3568222/array-slicing-in-ruby-looking-for-explanation-for-illogical-behaviour-taken-fro 02:36:17 That's entirely "sane" 02:42:50 -!- monqy has quit (Quit: hello). 02:54:52 -!- monqy has joined. 02:58:24 -!- pikhq has joined. 02:58:26 -!- pikhq_ has quit (Ping timeout: 252 seconds). 03:07:51 Trying to teach lambdas to someone who's being first introduced to them in Ruby 03:07:55 I'm going to cry 03:10:08 Please just never try and teach anyone anything. 03:36:29 -!- oerjan has joined. 03:49:38 lets not wander around the mulberry bush beating our heads 03:50:16 oops i may have misinterpreted the last few posts 03:51:03 `addquote lets not wander around the mulberry bush beating our heads 03:51:05 599) lets not wander around the mulberry bush beating our heads 03:51:10 just saving elliott work, here 03:52:10 i kinda did that one on purpose 03:52:24 scandalous! 03:52:43 so.... most aspects of mathematics is considered non-trivial 03:53:08 arithmetic for example involves some memorization of the core set of additions 03:53:24 4+5=9, 3+3=6, 7+5=12 03:53:33 um that one is pretty trivial, thinks a mathematician 03:54:09 depends... everyone learns it soem time or other 03:54:48 incidentally, homological algebra and probably category theory are known for having nearly all proofs be absolutely trivial routine 03:55:11 what i am heading towards is, 03:55:28 the theorems and definitions are important, but the proofs all go essentially the same way 03:55:28 if the tricky part of addition is memorizing rote single digit additions 03:55:44 what is the tricky part of lambda calculus? 03:56:08 itidus20: if you used binary there wouldn't be any tricky part left to addition. which is presumably why computers do that. 03:56:28 well that and two states being simple for electronics 03:56:50 why doesn't your grandma know lambda calculus? 03:57:10 maybe she does.. i am supposing a lot 03:57:13 in implementations, the tricky part of lambda calculus is frequently getting efficient alpha conversion, i think 03:57:37 why would i expect the local mayor not to know lambda calculus 03:57:40 you either have to make up lots of names on the fly, or use a clever nameless representation 03:58:13 what makes mathematicians privy to an understanding of this thing lambda calculus :D 03:58:37 is it like another language? 03:58:49 i am not fluent in italian for example 03:59:10 Being able to break if is not a good argument against a language 03:59:16 it's the concept of naming and definition distilled down to its smallest core, i think 04:00:36 i am simply trying to understand what lambda calculus is by defining why i don't know it 04:00:39 hehehe 04:01:34 well..... is it widely useful? 04:01:37 because it has never been declared a skill every citizen needs to know 04:01:43 when extended yes 04:01:43 ahh 04:02:06 haskell and scheme both derive from lambda calculus in some form 04:03:42 was scheme the first programming language which included lambda calculus properly? 04:03:55 original lisp bungled the scoping 04:03:56 well scheme is strict... 04:04:07 oh right, so maybe even scheme didn't 04:04:13 lisp had lex scoping before scheme i think. 04:04:26 albeit nondefault 04:04:29 hm 04:04:36 miranda? 04:04:52 maybe 04:04:56 i guess what i am wondering is what is the crucial bootstrapping concept to understand lambda calc 04:05:08 now this sounds odd to me cos i have a gigantic book on regular calc 04:05:13 itidus20: the lambda calculus. 04:05:19 well if we are going to disqualify scheme for being strict, we must disqualify miranda for being statically typed ;P 04:05:27 it is unrelated to calculus 04:05:51 oerjan: i am sure it can do lc = (lc -> lc)... 04:06:13 hm or maybe not 04:06:29 itidus20: church representations is crucial, i think, to understand how lambda calculus can embed all other (pure) programming concepts 04:06:33 *are 04:06:38 oerjan: http://en.wikipedia.org/wiki/SASL_(programming_language) 04:06:51 untyped, lazy 04:07:06 aha 04:07:22 oerjan: it just seems like it is one concept which no matter how long you look at it, it resists yielding 04:07:34 i am trying to see where i am going wrong 04:07:35 lc is easy 04:08:10 oerjan: http://en.wikipedia.org/wiki/Kent_Recursive_Calculator seems to be the big leap towards miranda/haskell 04:08:15 itidus20: it is just a bit obfuscated? 04:08:27 implemented in bcpl... 04:08:52 is it the case that you have to revise dozens of other concepts to make room for it? :D 04:08:54 -!- pumpkin has joined. 04:08:58 itidus20: maybe it's the core of higher-order functionalness 04:09:12 itidus20: no 04:09:21 how hard can 1 concept be, ya know :D 04:09:33 lc isn't hard 04:09:39 i wonder how new scientist would explain it 04:09:57 itidus20: because unlike arithmetic, lambda calculus does not talk about anything but itself, so you cannot ground it in anything but itself, without translating to it 04:10:42 ahhh 04:10:48 itidus20: maybe that is it. 04:11:00 -!- copumpkin has quit (Ping timeout: 258 seconds). 04:11:02 (don't worry i ask the buddhists similar questions about nirvana) 04:11:05 oerjan: /msg :P 04:11:13 elliott: i noticed. 04:11:23 how would new scientist expain nirvana 04:11:38 the best answer i was turned towards was the vacchagotta sutta 04:12:14 you guys might like it 04:12:19 itidus20: of course you can _add_ the ability to talk about other things than just functions, like haskell does 04:12:26 the specific part of it i have in mind anyway 04:12:33 but it's not part of the core 04:13:03 oerjan: ok so.. I intuitively want to link it to other ideas 04:13:12 to exploit things like analogy and metaphor 04:13:22 you can't. 04:13:30 and that would be what is hard 04:13:50 no. this applies to everything nontrivial. 04:13:59 analogies are not knowledge. 04:14:14 and knowledge not finite 04:15:05 lol siriously i came here cause i thought this was an occult chat 04:15:13 programmers in a different sence 04:15:18 it is most definitely not 04:15:23 sut-heb: hahahah hahahah 04:15:38 freenode is mainly about programming.] 04:15:39 sut-heb: well you're not the first. although it may be rare for them to stay this long... 04:15:53 cause i find the way you guys speak interesting 04:16:15 where not to different 04:16:26 anyway knowledge is finite 04:16:38 i am a newbie here anyway 04:16:39 brain has finite capacity, qed 04:16:47 conciousness is finite, knowledge is material 04:16:48 -!- MDude has changed nick to MSleep. 04:16:58 sigh 04:17:06 i wont go there 04:17:11 you guys can get back to your chat 04:17:19 sorry for the interruption 04:17:35 sut-heb: in reality esoteric programming languages seems to focus on mathematically interesting ones 04:17:42 i am not a mathematician though 04:17:54 i hate math myself 04:18:13 it may be that i have got it all wrong 04:18:28 and that mathematicians just happen to be the people who are most interested in esoteric languages 04:19:09 well without math, attempts to make esoteric languages tend to end up as just syntactic obfuscation 04:19:27 i.e. bad 04:19:35 In the beginning was the Word, and the Word was with God, and the Word was God 04:19:49 but we speak of number and symbolism behinde each letter 04:20:11 sut-heb: no. 04:20:27 * oerjan should get some popcorn. if he had any. 04:20:35 each creating there own different manifestation when combiuned with one another, making words manifesdt into physical form 04:20:50 technobabble. 04:21:25 anywys you guys do what you do, 04:21:46 ok 04:22:44 hi im back from not paying attention 04:22:53 did i miss exciting envigourating chat 04:25:06 monqy: nah everyone backed off :P 04:25:15 :'( 04:27:41 What are comonads for? 04:27:53 things :) 04:27:53 comonadic stuff 04:28:23 I think one time I almost used comonads but then I didn't 04:30:11 so i have this old ikea catalogue. 04:31:23 Are extract and duplicate like the opposite of return and join? 04:31:30 yep 04:37:29 so the night before last I had a dream and I remember two interesting enough parts 04:38:26 yay 04:38:40 there's a human heart on a table and my other is poking it and I tell her not to poke it because she doesn't know what it is for and bad things could happen but she keeps poking it and peeling at it and this frustrated me 04:38:44