00:00:00 --- log: started haskell/02.01.21 00:52:33 --- join: shapr (~user@195.156.199.178) joined #haskell 02:22:12 --- join: wicky (~wicky@cc11277-a.ensch1.ov.nl.home.com) joined #haskell 02:22:57 --- part: wicky left #haskell 02:59:49 if YouSay = "Vb sucks" then youwrong = true 03:01:32 what's good about vb? 03:03:02 vb is easy to make GUIs with 03:03:13 the language is not conducive to good programming habits though 04:24:34 The type system scares me 04:25:54 * jewel wonders what a good project would be for a yellow belt 04:34:09 oh, good question 04:34:15 hey, we need to come up with a system for belt colors 04:34:18 or I need to ): 04:34:20 :) 04:34:24 oops, mis-smile 04:35:50 ok... 04:36:03 white belt is where you get any kind of program to work in Haskell, even helloworld 04:36:15 --- join: Vutr__ (~ss@port-212.169.142.151.flat4all.de) joined #haskell 04:36:20 yellow belt is where you write a small program of your own and it works 04:36:28 I guess my Mandelbrot.hs is sufficient 04:36:42 Well then you'll have to change the topic 04:36:46 yah, I agree 04:36:54 What's a good project for orange then? 04:37:10 --- topic: set to 'Functional Programming, that's the opposite of Dysfunctional Programming (like VB) | we be loggin' http://tunes.org/~nef/logs/haskell/ | welcome to the Haskell Dojo | jewel and julien are up to white belt! shapr is up to yellow belt!' by shapr 04:37:13 oh 04:37:14 hmm 04:37:19 are you up to yellow belt also? 04:37:54 btw 04:37:58 what are all the belts? 04:38:07 white, yellow, orange, ... black 04:38:16 isn't your Mandelbrot.hs fairly complicated, in the way it uses type classes and stuff? 04:38:22 er, hm. 04:38:25 yah, now it is 04:38:28 it wasn't before 04:38:50 I added type class support because Luke wanted to Show a MandelData 04:38:59 that might qualify as orange belt... maybe 04:40:04 I'd say creating your own datatypes and using typeclasses could be called orange belt 04:40:35 but in that case I'd say being proficient with one of several things could be sufficient 04:40:39 not just typeclasses 04:40:58 --- nick: Vutr__ -> Vutral 04:43:02 I guess Monad proficiency would be black belt :) 04:43:14 and input/output proficiency would be brown belt 04:43:20 or whatever is below black 04:43:52 is the haskell wiki down? 04:44:15 there's quite a bit more complicated stuff than Monads, but once you've got to that stage you're doing pretty well generally 04:44:30 like arrows? 04:44:44 monads as brown belt then? and arrows as black belt? 04:44:59 arrows are fairly specialised, they're not in common use yet 04:45:09 oh 04:45:23 what's more complicated than Monads and less than arrows? 04:45:29 yeah I think I'm yellow belt too 04:45:37 I reckon Monads for black belt makes sense; you just have to recognise that there's more stuff you can learn once you get to Black belt 04:45:47 but in a sense your "apprenticeship" is over 04:45:48 It depends on the system, but you usually have about 10 kyu 04:46:00 --- topic: set to 'Functional Programming, that's the opposite of Dysfunctional Programming (like VB) | we be loggin' http://tunes.org/~nef/logs/haskell/ | welcome to the Haskell Dojo | julien is up to white belt! jewel and shapr are up to yellow belt!' by shapr 04:46:28 white, yellow, orange, green, blue, brown, black is common (with purple and variations on each belt making up the rest) 04:46:42 cool, sounds like a good system to go with. 04:47:08 Are you going to make a wiki page for the gradings? 04:47:11 yup 04:47:15 if I can get to the haskell wiki 04:47:33 can you? 04:47:48 I can't seem to get to sf.net either 04:47:49 :( 04:48:08 I may need to restart gnome 04:49:12 me neither 04:50:12 working now 04:50:42 * shapr goes back to working also 04:54:23 --- quit: Vutra_ (Success) 05:07:56 http://haskell.org/wiki/wiki?HaskellIrcChannel 05:07:58 ok, it's up 05:22:50 you might want to have datatype and type aliases as Orange, and type classes as green 05:23:20 Heffalump: it's wiki, change it ;) 05:23:36 I opine that you know better than I do 05:24:44 what are datatype and type aliases, what would be an example of their use? 05:26:33 I just added some to Mandelbrot-0.2.hs 05:27:31 http://kungens.kemi.fi/~shae/Mandelbrot-0.2.hs 05:28:26 see the line of code that has "instance" in it 05:29:31 haskell type classes are much like Java Interfaces 05:30:15 some examples are Eq (for equality testing) Ord (ordered types) and Show (for getting a text represenation of a type) 05:33:49 last week Luke_ tried to play with Mandelbrot interactively, but the type alias MandelData was not an instance of Show 05:38:11 Ok, that's not hard 05:38:19 I thought your mandelbrot used a GUI 05:39:11 nah, I wish :) 05:40:56 I'd like to use gtk, but debian doesn't have it by default, and building ghc with gtk support is not simple 05:46:15 it can't be dynamically linked? 05:46:28 --- join: LoganH (~logan@up112.blacksburg.ntc-com.net) joined #haskell 05:46:45 hi LoganH 05:46:48 Hello. 05:46:49 jewel: I don't know. 05:46:55 I'd like to though. 05:46:56 I'm trying to learn haskell. 05:47:08 you've come to the right place. 05:47:12 this and #loty 05:47:13 How might I transform a string containing whitespace-delimited integers to a list of integers? 05:47:41 er, readInt ? 05:48:02 I think it's read("1") :: Int 05:48:07 or something really close to that 05:48:18 you'll probably want to split the string first 05:49:05 if you want to join a group of people who are learning Haskell: http://pragmaticprogrammer.com/loty 05:49:06 Yeah, the splitting is what I don't know how to do (easily). 05:49:08 check that website 05:49:27 one of the Prelude's List functions does that simply 05:49:31 luckily splitting has been well discussed :-) 05:49:38 yah, very well discussed 05:49:47 I wrote one, PragDave wrote one, 05:49:52 and there's one in the standard lib 05:50:12 I guess I chose a good year to learn Haskell. 05:50:14 http://haskell.org/onlinelibrary/ 05:50:17 I agree! 05:50:44 The documentation on haskell.org has little in the way of examples, and I can only glean so much about what a function does by the types it operates on and returns. 05:50:52 if you check the List module of the onlinelibrary, you'll see many useful things 05:50:58 yah, I agree with you there 05:51:14 there's also YetAnotherHaskellTutorial starting up at the same time 05:51:52 the #loty and YAHT people are going to work together, and hopefully end up with an excellent tutorial, and some programmers with excellent knowledge of Haskell 05:52:42 I thought a good way to dive in and learn it is to try some simple problems from ACM contests (where you're normally expected to solve it in C). 05:52:49 an unfoldr there 05:52:55 I find the solving part is sometimes very nice, but the I/O I can't do yet. 05:53:11 yah, IO is challenging because of referential transparency 05:54:02 hey jewel, I emailed the LL1 guy about Joy, and he said that FP is very powerful, but that he finds Haskell too difficult to "grok" 05:54:10 I replied and told him about #loty ;) 05:54:41 hehe 05:55:55 I can understand that it's more fun and easier to learn stuff with others, but I don't think anything is too complicated or too esoteric to learn. 05:57:11 It's purely a matter of changing old habits and thinking a bit differently 05:57:34 yah, I agree 05:58:33 he also said what you figured out about that thread, that they would like to give optimization information to the compiler in the source 05:59:40 I didn't say that 05:59:42 Where has this splitting "been well discussed"? 05:59:54 heh, just locally 06:00:02 No archive for me to look at? :P 06:00:08 yes, there is :) 06:00:16 the channel topic has the url for logging 06:00:33 Good point. 06:01:00 I don't know which day had the splitting discussion 06:02:27 * Heffalump reappears 06:02:35 * shapr gasps 06:02:41 Was it referred to as the splitting discussion? 06:02:43 and edits the page as suggested :-) 06:03:27 which channel was it? 06:04:03 I think part or all of it was here on #haskell 06:04:11 oh wait 06:04:16 maybe it happened on #loty 06:04:22 in which case, it wasn't logged :( 06:06:23 I found a little discussion. 06:06:26 wget + grep :P 06:06:55 good method 06:07:59 --- join: Vutra_ (~ss@port-212.169.142.239.flat4all.de) joined #haskell 06:08:21 Now I know to look in PreludeList, which has a lot of stuff I didn't notice before. 06:09:47 yah, it's handy 06:09:55 we have all kinds of good advice here. 06:11:23 I have the #loty from friday till now 06:11:48 oh, good 06:12:02 I should ask nef to get clog over there 06:13:47 LoganH: you want it? there's a lot of idle chatter in there too 06:14:19 * shapr is the primary guilty party in that case ;) 06:20:08 jewel: Sure. 06:20:25 How do I make a non-IO something into an IO something? 06:20:46 you should ask one of the black belts about that. 06:20:53 return 06:21:16 Like, I have an Int and an IO [Int], and I want to add that Int to the head of that IO [Int] 06:21:31 f x ioxs = do xs <- ioxs 06:21:39 So like return x : y ? 06:21:40 return (x:xs) 06:22:04 does that make sense? 06:22:08 Yeah. 06:22:28 Is there a more concice way of doing it? 06:22:32 umm 06:22:54 possibly, but do you want your code to be readable? 06:23:09 LoganH: are you a former Perl programmer? ;) 06:23:19 Sometimes... :P 06:23:23 aha! 06:23:58 This would be easier if I knew how to just temporarily hold a value in a variable. :P 06:24:10 that's what the "xs <- ioxs" does, roughly 06:24:22 liftM (x:) ioxs 06:24:29 is more concise, btw 06:24:47 and looks scary 06:24:48 but you'll need to "import Monad" 06:24:58 shapr: hence my comment about "more readable" above :-) 06:25:13 shapr: actually, if you read the definition of liftM it's just a more general form of my code above 06:25:27 yah, I've heard about lifting 06:25:31 so it's not really scary, but it does require a bit more knowledge of the library (I had to go and look it up first myself) 06:25:33 liftM takes an x and lifts it into a monad? :P 06:25:33 I'm still not sure why you need to... 06:25:54 LoganH: no, it takes a function from as to bs and lifts it into a function from monad as to monad bs. 06:26:01 if that makes sense 06:26:02 oh 06:26:07 that does make sense 06:26:18 it's a 2nd order ... whatever you call it 06:26:21 like map 06:26:23 yeah 06:26:27 exactly 06:26:32 * shapr tries to remember the name of those things 06:26:41 in fact, liftM on the list monad is precisely map. 06:26:52 2nd order function? 06:26:54 aha! 06:26:56 that's the word 06:26:58 though it's not, it's third order 06:27:06 yah, makes sense 06:27:12 well, depending on how you define order, but by the definition I generally use it is 06:27:14 --- quit: Vutral (Read error: 110 (Connection timed out)) 06:27:25 map is 2nd order, because it takes a function 06:27:25 values (6, "foo") etc are first-order 06:27:38 liftM takes a function and turns it into a Monadizer 06:27:42 functions that take only first-order things as parameters are second-order 06:27:57 functions that take only first or second-order things as parameters are third-order. Etc. 06:28:00 * shapr avoids the Monadic tribes of graduate students roaming the halls of the CS building 06:28:05 but other people's definitions of order might vary 06:28:12 er 06:28:20 so why is liftM third-order then? 06:28:29 cos (a -> b) is second-order 06:28:58 oh, so map is 3rd order? 06:29:00 yep. 06:29:04 --- nick: Vutra_ -> Vutral 06:29:04 by my definition 06:29:08 ok, I got it. 06:29:19 (which is common in the literature of the subject of my thesis, but might not be elsewhere) 06:29:24 treerec in Joy is fourth order then 06:29:40 if you say so :-) 06:29:46 according to your definition 06:29:59 well, I don't know what it is, so I can't agree or disagree 06:29:59 because it requires something like map as one of its args 06:30:03 right 06:36:46 Heh, I just realized that lazy evaluation is my friend. :P 06:36:57 as always :-) 06:38:47 --- quit: Vutral (Read error: 104 (Connection reset by peer)) 06:39:23 --- join: Vutral (~ss@port-212.169.142.239.flat4all.de) joined #haskell 06:41:58 So why loty, as opposed to lotm? 06:44:19 LoganH: loty is based on some advice given in 'the pragmatic programmer', namely that programmers should learn at least one new language a year 06:44:21 have you read "The Pragmatic Programmer" ? 06:47:01 * LoganH shakes his head. 06:47:14 it's very much worth reading 06:47:17 it's in my top two books. 06:47:37 I don't know whether I'd recommend it over SICP, but then, I'm not sure I'd recommend SICP over it... 06:47:44 who wrote it? 06:48:07 Dave Thomas and Andy ... someone 06:48:16 Hunt 06:48:18 right 06:48:33 it's a great book. 06:48:56 content-wise, 50% of it is just about managing any kind of projects 06:49:15 my fiancee has been a project manager in several european projects, and she likes the book too. 06:49:23 --- join: PragDave (~dave@pragdave210.august.net) joined #haskell 06:49:24 the other 50% don't interest her at all. 06:49:34 in fact, PragDave is the Dave Thomas we just spoke of. 06:50:02 ah 06:50:12 * PragDave 's ears were burning 06:50:14 PragDave: my fiancee wants me to tell you that about 50% of PP is excellent advice for anyone managing projects. 06:50:26 she's managed several european projects. 06:50:44 she says the other 50% is just geeky stuff 06:50:46 * shapr grins 06:51:00 shapr: there's the pragmatic half, and the programming half :-) 06:51:06 right, exactly. 06:51:14 * PragDave signs dblack as his agent 06:51:19 happily, this means that she and I can read the book together. 06:51:23 PragDave: think you can afford me? 06:51:28 I consider that a great benefit. 06:51:36 Beatrice doesn't really want to read Haskell books with me. 06:51:43 * PragDave fires dblack on consideration of the financial side 06:51:46 even if she did buy them for me as presents. 06:52:05 wow, that's life in the fast lane dblack 06:52:09 * dblack_ sues PragDave for 5 times what would have been his salary 06:52:22 --- join: chad (trilluser@65.15.137.250) joined #haskell 06:52:29 * PragDave pays 06:52:34 * shapr laughs 06:53:15 shapr: we're actually looking at putting together some kind of workshop or course for managers 06:53:24 I think this would be wise. 06:53:29 along the lines of "How To Herd Cats" 06:53:38 *exactly* !! 06:54:40 i think many managers fail to realize just how to help developers do their jobs 06:54:47 they view is as carrot and stick 06:54:59 while it's really opportunity and cost 06:55:07 yea, as if developers *don't want* to work. ;) 06:55:30 so we're doing somethig about setting up an environment where developers and projects can flourish 06:55:33 some negative reviews on amazon 06:55:44 count yourself lucky if management even wants to help you do your job... 06:57:10 ignore my three dots 06:57:14 ignore the negative reviews on amazon 06:57:34 jewel: of? 06:57:42 your book, on amazon.co.uk 06:58:20 oh - there's that really quite hurtful one at the top over there "Garden Variety Programming" or somesuch 06:58:56 yeah 07:00:07 some folks come to books with a particular expectation. If the bok doesn;t match, then it fails them. 07:00:47 jewel: and another thing.... 07:01:01 note that the people who liked it (of that small sample of reviews) are the ones who used it for something 07:01:15 the people who didn't are doing a kind of literary-history take on it, or something 07:01:50 yes, comparing it to their favourite book 07:03:37 jewel: if you're an experienced developer who stays on top of modern practices, then the book probably won't tell you much you don't know 07:03:54 You'll be surprise at what other titles you've authored: http://www.amazon.co.uk/exec/obidos/Author=Thomas%2C%20Davis/026-0863672-4490001 07:05:28 I can never get patterns to work the way I want... 07:05:41 I have: 07:05:42 output :: [Cell] -> [Cell] -> IO () 07:05:47 output [oc:orest] [c:rest] = do 07:05:48 putStrLn ("Cell data in " ++ (show oc) ++ " " ++ (getResult c)) 07:05:48 output orest rest 07:05:55 Is that not correct? 07:06:54 no 07:06:57 i'm new to this, but try it without the '[]'s 07:07:03 yeah, use ( ) instead of [ ] 07:07:15 : is a constructor that gives you a list 07:07:36 so you don't need the [ ] - that ends up giving you a list of lists 07:07:37 : is an infix operator? 07:07:41 yep. 07:07:42 Oh. 07:07:56 Much better now. (: 07:07:57 Thanks. 07:10:03 PragDave: I've worked in Birmingham, Alabama; Seattle, Washington; and Tornio, Finland. I think your book rocks, and I haven't met any developer who did *not* need to read it. 07:10:20 oh yes, that's what I was getting at, shapr should write a review 07:10:25 heh :) 07:10:46 yah, I will. 07:11:19 I think my nifty/sucks meter is pretty accurate, and I rate tPP neck-and-neck with SICP. 07:11:52 what is SICP? 07:11:58 (having failed to find it in scrollback) 07:12:15 Structure and Interpretation of Computer Programs by Ablesson, Sussman, and Sussman 07:12:23 ah 07:12:46 have you read it? 07:12:56 nope 07:13:01 you should 07:13:13 * jewel finishes his SQL Server service pack download and prepares to go over the top 07:13:13 I'm an academic, I don't go in for practical things like that :-p 07:13:29 SICP is both academic and practical. 07:13:33 though actually I'm currently thinking about whether to take a software development job 07:13:54 Heffalump: no! not the red pill! 07:14:00 er, blue pill? 07:14:06 crap I can't remember which was which. 07:14:23 red was what he took, IIRC 07:14:28 oh 07:14:34 then I said it correctly. 07:14:51 :-) 07:15:05 Keanu Reeves is a wizard programmer 07:15:13 * shapr blinks 07:15:17 er, really? 07:15:21 not! 07:15:26 * shapr grins 07:16:00 speaking of which, the only show/movie I've seen that has *good* geek graphics is DarkAngel 07:18:43 --- join: smklsmkl (~sami@glubimox.yok.utu.fi) joined #haskell 07:18:43 --- quit: smkl (Read error: 104 (Connection reset by peer)) 07:23:05 What's the simplest way to sort integers in descending order? 07:23:52 sort ? 07:24:07 just guessing from looking at the List module 07:24:11 sortBy probably 07:24:15 cos he wants reverse order 07:24:23 or reverse . sort , but that'll be less efficient 07:24:58 sortBy (flip compare) 07:25:00 should do the job 07:25:12 shapr: thank's very kind of you (sorry - I'm on a non net-connected box most of today) 07:25:18 Didn't know I had to import List. 07:25:45 That works perfectly, thanks. 07:26:14 jewel: so, did you charge the enemy successfully? 07:26:21 do you see why? (I don't like just providing canned answers without explaining why they work) 07:27:09 no, still rebooting the machine 07:27:24 I'm still trying to figure out the correct DER encoding of an IMPLICITly tagged [0] SET in ASN.1 07:27:28 Yes. 07:27:35 cool 07:27:58 (flip compare) is the comparison operator for the sort operation, the flip just makes the comparator give opposite answers. 07:28:23 Not that I knew any of that before I saw the example (shows the value of having lots of nice small examples around :P) 07:28:44 I saw a neat project, someone providing implementations for everything that's in the Perl Cookbook in various languages (Haskell being one of them)... 07:28:56 that would be cool 07:28:59 The only bad thing is that it's worthless, because all those problems are kind of perl specific, and all the solutions are very perlish. 07:29:13 So there's a big helper library for Haskell that contains all these functions that provide perl-like routines. 07:29:17 You can't really learn much from it. 07:29:28 I think a more basic cookbook would be really n ice. 07:30:23 yah, I agree. 07:30:47 I tried to create a Haskell User Submitted Libraries project on sourceforge, I haven't heard whether it's been approved or not. 07:31:10 when it's up, I'd like for it to be a central repository for useful libs 07:31:27 are there any regex libs for haskell? (or should i just throw that paradigm away?) 07:31:35 yah, there are 07:31:48 rumor has it they're very nice 07:31:51 I haven't tried them 07:32:16 if only I could persaude my boss to let me prototype in Haskell :) 07:33:52 Heh, it only took me 4 hours, but I got this simple ACM problem implemented, complete with proper I/O. :P 07:34:04 cool! 07:34:15 And I even did the algorithmic part functionally. :P 07:34:27 :-) 07:34:38 Though that was the easy part. 07:35:06 btw flip compare works by swapping the arguments to compare, not by taking the opposite value of the result (I'm not sure which of the two you meant above) 07:35:18 Ok, that makes more sense. :P 07:35:25 (that flip woudl be more general like that) 07:35:26 what was the problem? 07:35:44 Problem G from the '98 ACM ICPC... 07:35:55 Basically, you're given a sequence of spreadsheet commands... 07:36:14 Each command is either a list of rows to delete, a list of rows to insert (that is, positions to insert new rows), or two cells to exchange. 07:36:30 Then after that you're given a list of cells, and you're supposed to output their locations after all the commands are applied. 07:36:35 It really lends itself to a functional solution. :P 07:36:58 url? 07:37:48 http://icpc.baylor.edu/past/default.htm seems to have a list of past contests 07:38:07 http://icpc.baylor.edu/past/icpc98/Finals/Report/Problems/Problems98.pdf is the specific set of problems 07:38:07 Oops, my mistake, it was '97 07:38:08 http://icpc.baylor.edu/past/icpc97/Finals/ProblemsF97.pdf 07:38:19 ah :-) 07:39:46 I should try a problem that calls for some sort of data structure now. 07:39:51 Like A. 07:40:12 yeah, I was thinking A would be an ideal candidate 07:40:33 Yeah, basically a tree of records, and some tree modifying routines. 07:40:47 * jewel curses microsoft for shipping buggy software 07:41:06 DAG of records, rather. 07:41:09 The 50 MB service pack indeed fixed our obscure problem 07:42:00 wow 07:42:09 apt-get upgrade ;) 07:42:15 dist-upgrade 07:42:25 I love apt. 07:42:46 The whole notion of "shipping" or a "release" of an OS is foreign to me now. :P 07:43:17 well, dist-upgrade isn't a good idea unless you're really going from potato to sid or something 07:43:52 What in Haskell has similar functionality to associative arrays? 07:47:00 FiniteMap 07:47:02 shapr: it's quite often needed to cope with package changes in woody, IME 07:52:20 I use sid 07:53:00 I'd guess the same would apply when someone reorganises packages there 07:53:12 probably so 07:53:14 it did when woody was unstable too, IIRC 07:53:48 What provides FiniteMap? 07:53:50 but I haven't done a dist-upgrade since I went from potato to sid 07:54:09 LoganH: FiniteMap 07:54:42 shapr: have you had to install certain packages manually? 07:55:14 loganH: see Manuel Chakravarty's stuff 07:55:25 I think it's in GHC too, but I just use his version all over 07:56:24 No, I haven't. 07:56:28 I'm using GHC. 07:56:47 It's in ghc 07:58:04 Heffalump: yah, I've done that 07:59:24 shapr: ah, ok, I generally just do dist-upgrade to avoid that hassle :-) 08:00:10 Do I have to import something to get these FiniteMap functions? 08:01:26 have you found the module? just import it, if so 08:01:56 I can't tell what the name would be. 08:02:47 run ghc with -syslib data 08:02:52 then import FiniteMap should work 08:03:52 * Heffalump isn't quite sure how to work out the module interface without downloading the GHC source, though... 08:03:58 or at least hslibs 08:04:04 Ahhh. 08:04:23 if you look in /usr/lib/ghc-5.00.2/imports/ 08:04:32 (replace 5.00.2 with whatever you have installed) 08:04:37 That took care of it. 08:04:45 then you'll see a bunch of subdirectories each of which has some hi files 08:04:45 It added several packages I didn't have before. 08:05:11 Yeah, I saw FiniteMap.hi in imports/data. 08:05:30 -syslib is the way to get at those subdirectories 08:05:38 there is a packages mechanism too, but I haven't really looked at that 08:05:48 so I can't tell you how it works :-) 08:06:07 I do :set -package data 08:06:12 that gives me access to lots more stuff 08:06:30 I wonder if there's a file I can edit that does all this by default. 08:06:53 strace ghci and see what files it reads? :-) 08:07:04