- Beginners - Haskell.org

State monad question
by Jordan Cooper 24 Jun '10

24 Jun '10

I have two pieces of state I'm managing: type FooState = State Foo Foo type BarState = State Bar Bar There is a bit of interplay between them, and I want a function that does something like: importantFunction :: Foo -> Bar -> (Foo, Bar) The problem is, >>= only works when I'm chaining all of the same type of state. So I can't do: modifiesFoo >>= \foo -> modifiesBar foo >>= \bar -> return (foo, bar) So how do I get the results of both modifications back? I hope that makes sense. Programming in Haskell is proving to be very difficult; I once again fear I'm too stupid.

4 5

searching for a file
by Michael Mossey 24 Jun '10

24 Jun '10

Can I get a hint how to do this: Given a directory D and a file extension E, find the newest file in D among all files with extension E. I wouldn't mind if you point me to the functions that are useful and leave the rest as an exercise. Or even just point me to the module(s) that is(are) useful. Thanks, Mike

2 2

howto reason infinite lists
by prad 24 Jun '10

24 Jun '10

i'm trying to figure out how to think out the circular definition for an infinite list and would appreciate suggestions. consider ones = 1 : ones this essentially says 1 : (1 : (1 : (1 : ones))) with ones being given by the original def. however, i had trouble with this one fib = 1 : 1 : [ a+b | (a,b) <- zip fib (tail fib) ] initially, i tried building it element by element and got stuck when i thought of what was fib tail fib and what was being zipped so i thought it out differently by assuming the infinite list was already created in the following fashion. fib is [1,1,a,b,c,d,e, ..] tail fib is [1,a,b,c,d,e,f, ..] to find a, we zip fib and tail fib getting (1,1) and add getting 2: fib is [1,1,2,b,c,d,e, ..] tail fib is [1,2,b,c,d,e,f, ..] get b from the next zip (1,2) and add giving 3: fib is [1,1,2,3,c,d,e, ..] tail fib is [1,2,3,c,d,e,f, ..] c comes from (2,3) which adds to 5: fib is [1,1,2,3,5,d,e, ..] tail fib is [1,2,3,5,d,e,f, ..] this way of thinking about it makes sense to me, but i'd be curious to know if it is proper thinking. it seems to me that if we consider the list to be already built, we can apply zip successively to get the next part of it, but if we think recursively we start at the beginning at each recursive level and go nowhere. -- In friendship, prad ... with you on your journey Towards Freedom http://www.towardsfreedom.com (website) Information, Inspiration, Imagination - truly a site for soaring I's

4 4

bad state monad instances
by Keith Sheppard 23 Jun '10

23 Jun '10

Hi, I'm working on understanding the state monad, and I got stumped pretty much right away. When I run the following script (with instances copied verbatim from http://www.haskell.org/all_about_monads/html/statemonad.html ) #!/usr/bin/env runhaskell \begin{code} {-# LANGUAGE MultiParamTypeClasses, FunctionalDependencies #-} import Control.Monad.State(Monad, MonadState(..)) newtype State s a = State { runState :: (s -> (a,s)) } instance Monad (State s) where return a = State $ \s -> (a,s) (State x) >>= f = State $ \s -> let (v,s') = x s in runState (f v) s' instance MonadState (State s) s where get = State $ \s -> (s,s) put s = State $ \_ -> ((),s) main :: IO () main = putStrLn "hello" \end{code} It fails with: statemonadtest.lhs:11:20: `State s' is not applied to enough type arguments Expected kind `*', but `State s' has kind `* -> *' In the instance declaration for `MonadState (State s) s' Can you see what I'm doing wrong? I must be making a really basic mistake but I'm not sure what it is. Thanks, Keith -- keithsheppard.name

3 3

for those who wondered where I was...
by Michael Mossey 22 Jun '10

22 Jun '10

I started to learn Haskell last year when I wanted a better language to implement a computer-assisted music composition system. As GUI-heavy, it proved too much of a challenge... learning a new paradigm (functional programming), trying to deal with GUI libraries and their documentation shortcomings, and dealing with my actual problem on top of that proved too much. I went back to Python. However, after a time I reassessed my reason for writing any of this code in the first place. I realized it made more sense to put my available time into ear-training in order to make the meat computer between my ears into a better composition system, rather than programming one. In the meantime, some smaller programming tasks came up (non-gui) and I started on them a few weeks ago, writing in Python. This involves translating a MusicXML document into a midi file, applying some custom algorithms I've written to provide greater expressive qualities. The first Python challenge was to parse the document. I am not aware of a MusicXML-specific Python module, so I used their more general XML code. It took me a few weeks to get a basic input system working. I.e. MusicXML documents translated into my own representation, and reasonable error checking. Then I thought, what the heck am I doing? This is a good task for Haskell! It was only the GUI challenge that turned me off to Haskell last year, but this program is smaller and is purely file-in file-out. Working with Haskell tonight, in about four hours I replicated several weeks of Python work. Yup, this is the right approach. Mike

3 2

XML
by Michael Mossey 22 Jun '10

22 Jun '10

What's the easiest way for a beginner such as myself to parse XML data, specifically MusicXML files? I want to load a partial representation of a MusicXML file into Haskell datatypes and manipulate it. I don't need to validate the file, and I can make some safe assumptions about the content, so this is really a pretty simple task. Thanks, Mike

5 5

suggested course of action
by prad 21 Jun '10

21 Jun '10

greetings! i am very intrigued with haskell having acquired some experience in pascal, elisp, c, perl, mostly python and some sql. i've been going through various tutorials and the online real world haskell. i've also been reading through the haskell98 report at appropriate points. are there any recommended approaches for learning and studying haskell? -- In friendship, prad ... with you on your journey Towards Freedom http://www.towardsfreedom.com (website) Information, Inspiration, Imagination - truly a site for soaring I's

3 3

Where does a real beginner begin? Scheme+SICP? Real World Haskell?
by Greg Morell 21 Jun '10

21 Jun '10

I'm not much of a programmer. I've only used PHP and Ruby for the last 10 years. But I've heard so many wonderful things about Haskell, I'd like to really spend the time to learn it. No particular purpose, except to broaden my mind and get to know (what I hear is) a completely different way of thinking about programming. But what's the best way to start from scratch? Should I start with "Structure and Interpretation of Computer Programs" book and Scheme as my intro? Then afterwards, get into the Real World Haskell book? Or just start with Haskell directly? Any advice appreciated.

5 4

Problem with typeclass
by Sebastien Geffroy 20 Jun '10

20 Jun '10

Hi all, Below is an example from the book "Real world haskell" data JValue = JString String | JBool Bool | JNumber Int deriving(Show) class JSON a where toJValue :: a -> JValue fromJValue :: JValue -> Either String a instance JSON JValue where toJValue = id fromJValue = Right instance JSON Bool where toJValue = JBool fromJValue (JBool a) = Right a fromJValue _ = Left "Not a JValue" I'm trying to test this in ghci : > fromJValue (JBool True) and i got <interactive>:1:0: Ambiguous type variable `a' in the constraint: `JSON a' arising from a use of `fromJValue' at <interactive>:1:0-22 Probable fix: add a type signature that fixes these type variable(s) I don't understand where i have to declare a type signature ? in the class definition for "JSON a" ? in the instance declaration for "JSON Bool" ? Thanks in advance. Sheb

2 1

Testing a data structure, etc
by matthew coolbeth 19 Jun '10

19 Jun '10

As an excercise, I have created a small module defining a data structure which is intended to be an efficient map from sequences (possibly very long ones) to objects of some other type. The source code is below for anyone interested. Three questions. Hopefully they aren't too naive: 1) Can anyone point me toward whatever standard modules would be most appropriate for timing this (or measuring other kinds of performance) against the naive Data.Map? 2) Any suggestions on how to improve the module itself are more than welcome. 3) In a language like this with meaningful white-space, how do people keep their lines of source code from getting too wide? There is one line below that stretched beyond the desired 80 characters and, in general, I seem to have a hard time avoiding this. Thanks for your help. -- mac -- A map from [a] to p module SequenceMap (SequenceMap, empty, insert, lookup, test) where import Prelude hiding (lookup) import qualified Data.Map import Data.Maybe data Ord a => SequenceMap a p = SequenceMap (Maybe p) (Data.Map.Map a (SequenceMap a p)) empty :: Ord a => SequenceMap a p empty = SequenceMap Nothing Data.Map.empty lookup :: Ord a => [a] -> SequenceMap a p -> Maybe p lookup [] (SequenceMap p m) = p lookup (a:as) (SequenceMap p m) = let msm = Data.Map.lookup a m in case msm of Nothing -> Nothing Just sm -> lookup as sm insert :: Ord a => [a] -> p -> SequenceMap a p -> SequenceMap a p insert [] p (SequenceMap oldP othrs) = SequenceMap (Just p) othrs insert (a:as) p (SequenceMap notP othrs) = let msm = Data.Map.lookup a othrs in case msm of Nothing -> SequenceMap notP (Data.Map.insert a (insert as p empty) Data.Map.empty) Just sm -> SequenceMap notP (Data.Map.insert a (insert as p sm) othrs) test :: Bool test = fromMaybe False (lookup "qwerty" (insert "qwerty" True empty))

4 3