00:00:10 but how do I specify an unsigned four-digit quater-imaginary number 00:00:51 I don't think cobol does quarter-imaginary numbers 00:01:16 clear deficiency. 00:02:32 Sgeo: is that your university doing cobol courses? 00:02:45 olsner, no 00:02:59 doubtful considering it's an irish tld 00:03:15 although ireland would be less rural than Sgeo's university 00:03:39 Phantom__Hoover, what, just because my school used to literally be an agricultural school? 00:04:01 yeah, they don't teach agriculture in cities 00:04:47 what's wrong with farming, eh? 00:04:54 Sgeo, it's called FARMINGDALE 00:05:12 for real? 00:05:16 haha. 00:05:22 if you asked me to come up with the most ridiculously agrarian name possible i would think that too blatant 00:17:38 'Formula 1 organisers in India say they are determined to ensure that the track near Delhi is completely "dog proof" ahead of Sunday's race' 00:19:32 "guard dogs have been deployed throughout the track" 00:19:51 I think I have another criticism of Clojure 00:20:18 * Phantom__Hoover sits ready in rapt fascination 00:20:41 There are several ways to find the value associated with a key in a hash-map. This, in and of itself, is not a problem. 00:21:12 You can use the hash-map as a function, or use the get or find functions, or use a keyword as a function if the key you want is a keyword 00:21:36 wait, like, you have :x as a key in the hash, so you can do (:x hashtable) to get the value? 00:21:40 Yes 00:22:10 weird 00:22:21 -!- augur_ has joined. 00:22:26 Now, the problem I have with this is, what if I want to make my own data structure that fakes being a hash-map. I have to implement both the structure as being funcallable, and the Associative protocol 00:22:39 -!- augur has quit (Read error: Connection reset by peer). 00:22:52 I think keywords as functions just use the Associative protocol. I might not even have the right name for it 00:24:03 -!- augur_ has quit (Read error: Connection reset by peer). 00:25:02 -!- augur has joined. 00:26:00 Bike, it seems to be more common than (hashtable :x) 00:41:43 -!- copumpkin has changed nick to Michelangelo. 00:41:52 -!- Michelangelo has changed nick to copumpki. 00:41:54 -!- copumpki has changed nick to copumpkin. 00:43:48 i also find that weird, but I'm already used to it. 00:46:41 And I don't know when (get hashtable :x) is ever used, maybe people who don't realize that the other forms also accept an optional what to return if it's not there parameter the way get does 00:50:13 shachaf: What is lens? 00:50:35 baby don't hurt me 00:50:37 don't hurt me, no more 00:50:40 elliott: A Norwegian word meaning "hi monqy". 00:51:06 -!- hagb4rd has quit (Quit: hagb4rd). 00:51:08 Where's oerjan, by the way? 00:52:32 Sgeo: When your key is not a keyword 00:52:32 or when your collection can be a number of things (all that support get) 00:53:03 I guess get could also be used with -> 00:53:32 ,({1 2 3 4} 1) 00:53:32 2 00:53:32 ,([1 2 3 4] 1 2) 00:53:32 # 00:53:32 ,(get [1 2 3 4] 1 2) 00:53:33 ,(get [1 2 3 4] 6 2) 00:53:35 2 00:53:37 2 00:55:24 -!- copumpkin has changed nick to OMARCOMINYO. 00:59:16 shachaf: eating, last I heard 00:59:42 but that was a few days ago, may have changed since 01:01:03 -!- OMARCOMINYO has changed nick to copumpkin. 01:05:04 Does lambdabot do some form of lisp? 01:08:14 yeah, it does haskell 01:08:41 I'm so glad we have Sgeo here to paste logs from Clojure channels in here. 01:08:45 lolololololololol 01:08:50 Otherwise we might forget that this is #clojure. 01:09:25 yeah in the mainstream programming world, haskell is 'some kind of lisp' and yet the differences between python and ruby are vast and important 01:09:37 .. 01:09:38 perspective is a funny thing 01:09:59 kmc 01:09:59 kmc 01:10:00 i think 01:10:01 I just wanted to be silly and make a long string using lots of conses 01:10:01 olsner 01:10:02 FreeFull: judging by @list, not really? 01:10:03 might have been joking 01:10:04 elliott 01:10:08 But the haskell form of that is much shorter =P 01:10:08 yes i know olsner was joking 01:10:37 kmc: I'm glad you know 01:10:56 i'm glad i know you're glad i know 01:10:57 (cons 'S' (cons 't' (cons 'r' (cons 'i' (cons 'n' 'g'))))) 01:11:00 kmc: Why was ε afraid of ζ? 01:11:07 why 01:11:18 Because ζηθ! 01:11:19 Because 7 8 9 01:11:20 man you can't use single quotes like that in lisp, you'll break something 01:11:43 'Bike' 01:11:49 D: 01:11:57 shachaf: idgi 01:12:09 kmc: Why was epsilon afraid of zeta? Because zeta eta theta! 01:12:14 I think int something = 'abcd'; is valid C 01:12:15 Zeta ate-a theta. 01:12:30 shachaf: Are they consecutive in the greek alphabet? 01:12:35 Yes. 01:12:57 i see 01:13:10 Hmph. Best try another joke. 01:13:15 kmc: Where does a general keep his armies? 01:13:19 where 01:13:23 In his sleevies! 01:13:32 hehehe 01:13:35 that's a popsicle stick joke 01:13:41 FreeFull: Strin . g 01:13:48 wow 01:13:48 What's that? 01:13:52 esolang just got an anonymous edit 01:13:53 what bothers me about that is that sleevies is (probably) just a made up word 01:13:55 from an IPv6 address 01:13:59 nice 01:14:03 it must be aliens from the future 01:14:36 ais523: wait, that works? 01:14:40 are strings lists in any common lisps? 01:14:40 well 01:14:43 good thing i set that up! 01:14:46 > 'S':'t':'r':'i':'n':'g':[] 01:14:48 "String" 01:15:00 i thought haskell was practically the only language where the "built-in" / "default" string type is a linked list 01:15:03 elliott: apparently it works 01:15:05 > 'S':'t':'r':'i':'n':'g':"ent" 01:15:05 well, common lisps other than haskell then 01:15:06 "Stringent" 01:15:10 olsner: clearly 01:15:17 wait, haskell strings are linked lists, really? 01:15:21 yes 01:15:26 type String = [Char] 01:15:35 why? 01:15:36 kmc: What, you didn't know lisp did strings as linked lists? 01:15:46 Bike: Because Haskell is all about List Processing. 01:15:51 snerk 01:15:51 Being a Lisp. 01:16:05 @quote list.processing 01:16:06 monochrom says: in retrospect, it seems lisp designers were more interested in list processing than functional programming 01:16:09 FreeFull: Common Lisp? Scheme? Clojure? Maclisp? what? 01:16:12 "Lisp" isn't a language 01:16:29 Lisp 1.5 is a language. 01:16:33 is that really "in retrospect" 01:16:44 in lisp 1.5 you could use lists as functions! 01:16:51 Bike: well, it avoids introducing new concepts to the language, if you have lists and chars already 01:16:55 processing lists with lists, the way of the future. 01:16:59 Clojure strings are not lists but you can convert them into.... a list-like thing 01:17:00 and it gives you lazy strings in a natural way 01:17:07 but it's horribly inefficient 01:17:13 yeah that's what I was going to say 01:17:14 Well I guess strings aren't lists in common lisp 01:17:18 if you want to efficiently store text in Haskell, you use Data.Text 01:17:20 is there not a type for vectors/arrays/whatever 01:17:22 since all existing lisps were designed in the past, anything we say about lisp is in retrospect 01:17:32 there are many many array/vector types in Haskell 01:17:33 too many 01:17:35 Bike: it's a historical accident 01:17:36 sweet. 01:17:39 it's just that they aren't imported by default 01:17:44 kmc: well there are like three 01:17:46 elliott: that strings are lists? 01:17:51 and one of them is a replacemen tfor one of the others 01:17:53 olsner: I think you'll find most languages were designed in the past, assuming there is a design 01:17:53 Bike: yes 01:17:54 Data.Text is the way to go for efficient text storage, afaik 01:17:59 unless that's changed again 01:17:59 Data.Text doesn't give you random access. :-( 01:18:00 Bike: it made sense in 1992, it doesn't make sense in 2012 01:18:00 it uses UTF 01:18:01 that's quite the accident 01:18:02 -16 internally 01:18:05 which is kinda gross 01:18:06 shachaf: random access is irrelevant 01:18:11 kmc: icu compatibility 01:18:16 random access to Unicode codepoints isn't too useful 01:18:16 Bike: it is actually worse than you think 01:18:21 Bike: since haskell has lots of pointers in its linked list 01:18:23 because non-strictness 01:18:23 elliott: You can't even mark a point in a string with an index and then index into that point later, though. 01:18:24 elliott: made sense in 1992? but lisps that existed in '92 used arrays for strings 01:18:24 because codepoints != characters != anything you care about 01:18:32 Bike: but they aren't lazy 01:18:44 and also Lisp had more pretentions to being real-world useful than Haskell circa 1992 01:18:45 it's like 16 bytes to the character or something iirc 01:18:49 strange, strange, very strange 01:18:50 In Haskell you can have an infinitely long String 01:18:52 oh I suppose that's true though. 01:19:00 actually String is sometimes faster than the alternatives amusingly 01:19:08 you have to remember, for the first 10+ years of its life, Haskell was explicitly a language for PL researchers 01:19:09 > fix ("Honk " ++) 01:19:11 "Honk Honk Honk Honk Honk Honk Honk Honk Honk Honk Honk Honk Honk Honk Honk... 01:19:12 > repeat 'a' 01:19:14 "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa... 01:19:16 how very strange to consider 01:20:02 kmc: arguably it still is :) 01:20:07 well yeah 01:20:22 but i think before 2000 there were very few people claiming otherwise 01:20:28 now there is a large IRC channel full of them 01:20:45 #haskell is pretty bad these days, by the way. 01:20:54 "surprised??" 01:20:56 i would like to see kmc visit #haskell sometime 01:21:05 shachaf: what now 01:21:06 what is wrong with #haskell? 01:21:13 kmc: Same old. 01:21:24 #haskell-blah alternates between bad and good, depending on time zone. 01:21:32 which timezones are the bad ones 01:21:44 :-( 01:21:47 Bike: newbie joins, newbie asks standard newbie question, everyone starts shouting out their favorite explanation, then everyone argues about which one is best, then irrelevant tangents 01:21:51 newbie is confused 01:21:57 nobody actually cares about figuring out the one best way to teach something 01:22:02 wait, are you telling me not every speech about programming is like that 01:22:06 everyone's too heavily invested in their pet explanation 01:22:08 kmc: Remember sorear? 01:22:14 i felt the same way 01:22:16 which is why i left 01:22:22 after making some halfassed attempt to be better 01:22:23 shachaf: no 01:22:27 Those were the days. 01:22:57 Bike: also they act like a research paper on a topic is equivalent to mature tested packaged software 01:23:14 groan. 01:23:23 if your problem can't generate an academic paper then it's already solved 01:23:29 tbh i don't really see that attitude in #haskell at all 01:23:30 it's just a trivial matter of engineering 01:23:35 the research paper thing that is 01:23:37 i've seen it a few times 01:23:49 somebody was like "lol, just write a JVM backend for GHC, problem solved" 01:23:53 as though this would be completely trivial 01:24:03 the #1 problem is idiots who don't know anything about haskell trying to answer an overly-broad question like "what are monads" that they don't themselves understand 01:24:04 imo 01:24:09 yeah 01:24:12 that is a bigger problem 01:24:15 Yes. 01:24:26 (and I say idiots only because they decide they are competent enough to educate others, not because they merely don't know) 01:24:32 also #haskell has a bunch of dumb inside jokes 01:24:43 if you mention monads everyone is like "lolololololololol burritos" 01:24:56 elliott, now I'm scared that I'm an idiot 01:24:58 completely drowning out any attempt at a serious answer (not that such attempts would generally be successful) 01:25:04 also lambdabot is a huge inside joke 01:25:18 because she implements this bizarre variant of haskell which seems to exist to confuse beginners and entertain experts 01:25:23 pretty trollish basically 01:25:28 kmc: You should post that post on what's wrong with #haskell 01:25:49 Sgeo: i don't know how to answer that 01:26:02 people are always like "lambdabot is a great resource for learning Haskell!!" and then when the learner actually uses it, they have to explain why it thinks functions are numbers or whatever 01:26:07 :t (.) 01:26:08 forall a b (f :: * -> *). (Functor f) => (a -> b) -> f a -> f b 01:26:08 it what. 01:26:10 :t 3 3 01:26:11 forall t t1. (Num (t1 -> t), Num t1) => t 01:26:14 and then everyone blames Cale for making lambdabot's haskell weird, and Cale vehemently denies having anything to do with it 01:26:21 kmc: btw you are way too cynical about lambdabot, cale obviously feels its variants are useful and helpful for its target audience 01:26:28 i seem to have opened up a terrifying line of information here 01:26:29 it is unfortunate that it causes problems but i don't think it is really malicious 01:26:35 > 3 5 01:26:36 3 01:26:37 just obviously lambdabot never get updated ever since have you seen its codebase 01:26:40 I don't think there's a whole lot of malice in #haskell 01:26:48 Bike, 01:26:49 (by "cale" here I mean "broadly-approximated cale") 01:26:51 Malice is easier to deal with, usually. 01:26:53 > (sin + cos) 2 01:26:54 0.4931505902785393 01:26:55 (which means "Whoever is responsible for them" or whatever) 01:26:58 *whoever 01:27:00 Bike: welcome to #esoteric 01:27:06 it's crap 01:27:07 hello #esoteric 01:27:08 > (sin + cos) 0 01:27:09 1.0 01:27:19 Sgeo: is the same as sin 0 + cos 0, or what 01:27:26 > (sin * sin + cos * cos) 12345 01:27:28 0.9999999999999999 01:27:29 Yes 01:27:38 that's kinda cool actually. 01:27:48 It's not a stanard Haskell thing 01:28:12 so I gathered 01:28:13 elliott: i don't think it's malicious, but the fact that it's gone on this long indicates a certain lack of regard for beginners 01:28:17 #esoteric is also a big inside joke. 01:28:17 Nothing here 01:28:28 yes I've gotten that impression, shachaf 01:28:31 which is incongruous with a community that is constantly patting itself on the back over how friendly it is 01:28:40 one of the main things i learned from #haskell is that friendly != helpful 01:29:09 kmc: well experts are misserved by lambdabot's other problems 01:29:13 like how it keeps crashing 01:29:17 and that stupid L.hs nondeterministic thing 01:29:28 it is just unmaintained, can't really single out any individual problem imo 01:29:38 Cale "maintained" flip into L.hs 01:29:41 @ty flip 01:29:42 forall a b c. (a -> b -> c) -> b -> a -> c 01:29:46 whoa, dudeb 01:29:48 What? 01:29:53 a lot of IRC channels are too unfriendly to be helpful, but #haskell is too friendly to be helpful 01:29:56 kmc: btw i think the patting themselves on the back thing is generally over 01:30:01 > flip 01:30:01 elliott: oh really 01:30:03 Overlapping instances for GHC.Show.Show 01:30:03 ((a ->... 01:30:06 kmc: well i don't see it 01:30:12 kmc: anyway looks like it's your birthday 'cuz flip looks normal to me 01:30:18 haha 01:30:22 but (.) is still weird 01:30:23 I asked Cale to change it back today. 01:30:26 heh 01:30:32 Maybe he listened. 01:30:34 flip used to be weird too? 01:30:45 olsner: yes, it was strong 01:30:53 Yep, f (b -> c) -> b -> f c 01:30:55 strong :: (Functor f) => f (a -> b) -> a -> f b 01:30:59 i think edwardk calls it strong 01:31:02 so i am calling it strong 01:31:03 to be cool 01:31:03 strength? 01:31:06 oh maybe 01:31:15 it's the proof term that every haskell Functor is a strong functor right 01:31:21 shachaf: I got my Stripe CTF shirt today! 01:31:26 kmc: Oh. 01:31:29 :-( 01:31:33 @djinn (Functor f) => f (a -> b) -> a -> f b 01:31:33 u no? 01:31:33 Error: Class not found: Functor 01:31:36 Nope. 01:31:43 djinn doesn't do classes, I guess 01:31:48 elliott: want another proof that every functor is strong?? 01:31:52 proof = undefined 01:32:13 @djinn-env 01:32:14 data () = () 01:32:14 data Either a b = Left a | Right b 01:32:14 data Maybe a = Nothing | Just a 01:32:14 data Bool = False | True 01:32:14 data Void 01:32:15 type Not x = x -> Void 01:32:17 class Monad m where return :: a -> m a; (>>=) :: m a -> (a -> m b) -> m b 01:32:20 class Eq a where (==) :: a -> a -> Bool 01:32:21 either :: (a -> c) -> (b -> c) -> Either a b -> c 01:32:25 olsner: you need to teach it about functor i think 01:32:28 idk how 01:32:33 idk either 01:32:39 @help djinn 01:32:39 djinn . 01:32:39 Generates Haskell code from a type. 01:32:39 http://darcs.augustsson.net/Darcs/Djinn 01:32:42 no 01:32:42 wrong 01:32:42 elliott: It won't help. 01:32:50 djinn is "stupid" about classes 01:32:50 it knows about monads but not functors? 01:32:52 Observe: 01:32:55 @djinn a -> m a 01:32:56 -- f cannot be realized. 01:33:01 Er. 01:33:04 Bike: so does the haskell 2010 standard :( 01:33:04 @djinn Monad m => a -> m a 01:33:05 f = return 01:33:08 @djinn Monad m => b -> m b 01:33:08 -- f cannot be realized. 01:33:19 shachaf: haha 01:33:19 good 01:33:26 i'm pretty dumb at haskell but I thought monads were functors? 01:33:39 Mathematically 01:33:39 Bike: yep. 01:33:39 *haskell, math, life 01:33:51 haskell only has a direct monad class 01:33:51 Functor? I 'ardly know 'er! 01:33:56 there's a functor class in the standard libraries 01:34:00 very strange 01:34:02 (and applicative, the "half-way" point in a sense) 01:34:10 but they're not superclasses of monad due to -- wait for it -- historical accident 01:34:12 kmc: When are you running the kmc code contest? 01:35:00 :t (.) 01:35:01 forall a b (f :: * -> *). (Functor f) => (a -> b) -> f a -> f b 01:35:04 Ah, good. 01:35:18 elliott, now I'm scared that I'm an idiot 01:35:28 @let switchify f = (.f) 01:35:30 Defined. 01:35:31 @ty switchify 01:35:33 forall (f :: * -> *) a b. (Functor f) => f a -> (a -> b) -> f b 01:35:36 zomg 01:35:55 you're not an idiot generally since you don't confidently spew bullshit 01:36:39 Phantom__Hoover: pssst, i heard Sgeo made 6 bf derivatives before breakfast today 01:36:55 then he's not an idiot, he's a dead man walking 01:37:33 shachaf: Doesn't the Functor (.) upset you? 01:37:37 It should use the Category (.). 01:37:59 obviously it just predates Category, and no-one's bothered to update it yet 01:38:03 elliott: Maybe it should use the -> . 01:38:16 shachaf: No. Prelude should use the Category (.). 01:38:35 -!- hagb4rd has joined. 01:40:24 :t (*) 01:40:25 forall a. (Num a) => a -> a -> a 01:41:07 shachaf: it shipped from colorado 01:41:27 elliott: is there a place in the uk called "colourado" 01:41:34 yes 01:41:40 "i knew it" 01:41:59 it's an australian hair salon apparently 01:49:30 -!- Arc_Koen has quit (Quit: Arc_Koen). 01:52:49 Phantom__Hoover, what if I make a meta-language to describe BF derivatives 01:53:03 you mean tr? 01:53:05 can the language describe itself 01:54:33 More-like sed 01:55:45 depends on how awful it is 01:55:49 kmc: Do you have any good train facts for me? 01:56:54 just... general train facts? 01:57:55 Specific ones are good too. 01:58:09 i mean, are you looking for a specific kind of train fact 01:58:11 choo choo 01:58:13 do you know about gauntlet track? 01:58:32 http://en.wikipedia.org/wiki/Gauntlet_track 01:58:37 Trains are actually electric carts. (in CJK) 01:58:46 it is a strange kind of arrangement which is useful in many ways 01:59:56 Why the fuck would the word "arrangement" make me think of Worlds? 02:00:09 kmc: interesting 02:01:46 kmc: Nope. 02:03:18 when a two-track line has to cross a narrow bridge, it's sometimes better to use two overlapping tracks on the bridge, rather than a true single track segment 02:03:25 you can still only have one train on the bridge at once 02:03:41