December 2011 - Beginners

Pattern matching over functions
by Giacomo Tesio 13 Dec '11

13 Dec '11

Hi Haskellers, I'm wondering why, given that functions and referential transparency are first class citizens in haskell, it isn't possible to write a mapping function this way: f1 :: a -> b f2 :: a -> b f3 :: c -> a -> b map :: (a -> b) -> T a -> T b map f1 = anEquivalentOfF1InTCategory map f2 = anEquivalentOfF2InTCategory map f3 $ c = anEquivalentOfF3withCInTCategory map unknown = aGenericMapInTCategory Is it "just" the implementation complexity of the feature in ghc that prevents this? Or is it "conceptually" wrong? At a first look, I thought that most of complexity here should be related to function's equality, but than I thought that the function full name should uniquely map from the category of meanings in the programmer mind to the category of implementations available to the compiler's. Giacomo

9 17

In Search Of a clue... (Defining and making use of a type)
by Allen S. Rout 13 Dec '11

13 Dec '11

So, I'm approaching a problem which I think I understand pretty well. I'm a novice to Haskell, though, and I'm having difficulty even getting started with e.g. data types and building more complex structures. Here's what I'm trying to do; this is an xmonad customization/extension. My end goal is thus: Innermost: a 'Screen Tuple' , which I'm calling a 'ScrUple'. This represents a statement like 'screen 0 is displaying workspace "mail"'. Next: A variable-length list of these, I'm calling 'ScrConfig'. It means something like 'display this workspace on screen 0, that one on 1, etc'. Next: [ haven't gotten here ] a hash, or something: pairs of 'label : ScrConfig'. I don't know if the most haskelly way to do that is to build another type, and then another aggregate of that type.. In PERL, it'd be something vaguely like: $configs = { 'initial' => { \(0,"mail"), \(1,"web"), \(2,"jabber") } 'project' => { \(4,"editor"), \(1,"compile"), \(2,"jabber") } }; Here's what I'm doing so far, ---------------- data ScrUple = ScrUple { xineramascreen :: Integer , workspace :: String } deriving (Show) data ScrConfig = ScrConfig [ ScrUple ] deriving (Show) s1 = ScrUple 0 "mail" s2 = ScrUple 1 "web" ScrConfig sc1 =ScrConfig( [s2 s1] ) ; main = putStrLn $ show sc1[1] ------------------ and I get errors like: play.hs:17:27: Couldn't match expected type `ScrUple -> ScrUple' against inferred type `ScrUple' In the expression: s2 s1 In the first argument of `ScrConfig', namely `([s2 s1])' In the expression: ScrConfig ([s2 s1]) play.hs:21:19: Couldn't match expected type `[t] -> String' against inferred type `String' In the second argument of `($)', namely `show sc1 [1]' In the expression: putStrLn $ show sc1 [1] In the definition of `main': main = putStrLn $ show sc1 [1] ------------------- ... I feel like I'm thinking about the problem wrong, because this kind of "Here's how you build up a data structure" doesn't seem to be in the tutorials. I've been working through LYAH and Gentle Introduction, but so far haven't found things that feel related. I'd be delighted with pointers to the right parts of the Fine Manual, and similarly pleased with discursion on how to think about this data storage problem from a haskelly point of view. - Allen S. Rout

5 4

Re: [Haskell-beginners] Fwd: Averaging a string of numbers
by Ben Kolera 12 Dec '11

12 Dec '11

There is some magic here that I'm not quite groking. Sorry for my slowness; but I seem to be missing a step: This is how I'd expect liftA2 to work ( and is why I didn't use lift in my initial response ): *Main Control.Applicative Data.Monoid> liftA2 max Nothing (Just 1) Nothing I expected all the magic to be the applicative class instance that was generated for Maximum by the GeneralizedNewtypeDeriving extension, but why do these not work? *Main Control.Applicative Data.Monoid> liftA2 max (Maximum Nothing) (Maximum (Just 1)) Maximum {getMaximum = Nothing} *Main Control.Applicative Data.Monoid> mempty `mappend` (Maximum (Just 1)) `mappend` (Maximum (Just 2) ) Maximum {getMaximum = Nothing} When this obviously works just fine? *Main Control.Applicative Data.Monoid> main Stats {ct = Sum {getSum = 0}, sm = Sum {getSum = 0.0}, mn = Minimum {getMinimum = Nothing}, mx = Maximum {getMaximum = Nothing}} Stats {ct = Sum {getSum = 1}, sm = Sum {getSum = 1.0}, mn = Minimum {getMinimum = Just 1.0}, mx = Maximum {getMaximum = Just 1.0}} Stats {ct = Sum {getSum = 1}, sm = Sum {getSum = 2.0}, mn = Minimum {getMinimum = Just 2.0}, mx = Maximum {getMaximum = Just 2.0}} Stats {ct = Sum {getSum = 2}, sm = Sum {getSum = 3.0}, mn = Minimum {getMinimum = Just 1.0}, mx = Maximum {getMaximum = Just 2.0}} Sorry if I am missing something obvious and this question is really silly! On Mon, Dec 12, 2011 at 5:18 PM, Dean Herington & Elizabeth Lacey <heringtonlacey(a)mindspring.com> wrote: > At 8:21 AM +1000 12/12/11, Ben Kolera wrote: >> >> That is just because you are calling min and max against the Maybe >> rather than on the values inside of your maybes. Max is working >> because there is an instance of Ord for Maybe and >> >> Nothing > Just n > Just ( n + 1 ) > > > You have the right idea, but replace `>` above by `<`. > > >> >> This is certainly not the most elegant solution ( I am a beginner, too >> ) but here is what I would do: >> >> instance Monoid Stats where >> mempty = Stats 0 Nothing Nothing 0 >> mappend (Stats sm1 mn1 mx1 len1) (Stats sm2 mn2 mx2 len2) = >> Stats >> (sm1 + sm2) >> (chooseMaybe min mn1 mn2) >> (chooseMaybe max mx1 mx2) >> (len1 + len2) >> >> chooseMaybe _ Nothing Nothing = Nothing >> chooseMaybe _ (Just a) Nothing = Just a >> chooseMaybe _ Nothing (Just b) = Just b >> chooseMaybe f (Just a) (Just b) = Just $ f a b >> >> >> Hopefully this quick answer can get you on your way to solving your >> problem and we can both learn a better way of doing it when someone >> optimises my solution. ;) > > > You've got the principle just right. Here's a way to cast it that makes it > apparent that `Stats` is a monoid in a "componentwise" fashion. > > > {-# LANGUAGE GeneralizedNewtypeDeriving #-} > > import Data.Monoid > import Control.Applicative > > > -- | Monoid under minimum. > newtype Minimum a = Minimum { getMinimum :: Maybe a } > deriving (Eq, Ord, Functor, Applicative, Read, Show) > > instance Ord a => Monoid (Minimum a) where > mempty = Minimum Nothing > mappend = liftA2 min > > -- | Monoid under maximum. > newtype Maximum a = Maximum { getMaximum :: Maybe a } > deriving (Eq, Ord, Functor, Applicative, Read, Show) > > instance Ord a => Monoid (Maximum a) where > mempty = Maximum Nothing > mappend = liftA2 max > > data Stats = Stats { > ct :: Sum Int, > sm :: Sum Double, > mn :: Minimum Double, > mx :: Maximum Double } > deriving (Eq, Show, Read) > > instance Monoid Stats where > mempty = Stats mempty mempty mempty mempty > mappend (Stats ct1 sm1 mn1 mx1) (Stats ct2 sm2 mn2 mx2) = > Stats (ct1 `mappend` ct2) > (sm1 `mappend` sm2) > (mn1 `mappend` mn2) > (mx1 `mappend` mx2) > > > mkStats v = Stats (Sum 1) (Sum v) (Minimum (Just v)) (Maximum (Just v)) > > st0, st1, st2, st3 :: Stats > > st0 = mempty > st1 = mkStats 1 > st2 = mkStats 2 > st3 = st1 `mappend` st2 > > main = mapM_ print [st0, st1, st2, st3]

4 3

IO ( stuff )
by Paul Monday 12 Dec '11

12 Dec '11

I've hammered through the bulk of my issues as I try to build a matrix that consists of random numbers, but the final hurdle is mixing pure and impure functions. Does "impurity" from something like a random number generator or file I/O have to move it's way all the way through my code? Here is the specific example. I have a function that makes a row in a matrix, it looks like this: makerow :: Int -> (Float, Float) -> IO (U.Vector Float) The IO (U.Vector Float) is the result of the random number generation that is an impure call, making this an impure result. I want to compose this into a list makerows :: Int -> Int -> (Float, Float) -> [U.Vector Float] makerows 0 _ _ = [] makerows r n range = makerow n range : makerows r' n range where r' = r - 1 But, of course, I can't mix the IO (U.Vector Float) with a U.Vector Float The compilation result in: Couldn't match expected type `U.Vector Float' with actual type `IO (U.Vector Float)' In the return type of a call of `makerow' So, at some point, I have to lift I believe … is there a simple lifting solution? It initially seemed that Monads were the solution … but liftIO resulted in the same thing … just with Monad (U.Vector Float) … There is just some "simple" Haskellism I'm missing here, but after an ginormous amount of reading and googling, it is still eluding me :( Any thoughts (besides the one page I found that helpfully … basically … said "Go back to Java" ;-) Paul Monday Parallel Scientific, LLC. paul.monday(a)parsci.com

6 7

Fwd: Averaging a string of numbers
by Ben Kolera 12 Dec '11

12 Dec '11

That is just because you are calling min and max against the Maybe rather than on the values inside of your maybes. Max is working because there is an instance of Ord for Maybe and Nothing > Just n > Just ( n + 1 ) This is certainly not the most elegant solution ( I am a beginner, too ) but here is what I would do: instance Monoid Stats where mempty = Stats 0 Nothing Nothing 0 mappend (Stats sm1 mn1 mx1 len1) (Stats sm2 mn2 mx2 len2) = Stats (sm1 + sm2) (chooseMaybe min mn1 mn2) (chooseMaybe max mx1 mx2) (len1 + len2) chooseMaybe _ Nothing Nothing = Nothing chooseMaybe _ (Just a) Nothing = Just a chooseMaybe _ Nothing (Just b) = Just b chooseMaybe f (Just a) (Just b) = Just $ f a b Hopefully this quick answer can get you on your way to solving your problem and we can both learn a better way of doing it when someone optimises my solution. ;) On Mon, Dec 12, 2011 at 7:37 AM, goodman.m.w(a)gmail.com <goodman.m.w(a)gmail.com> wrote: > Thanks for the reply, Brent! > > I have a question about the "later" I referred to: > > On Sun, Dec 11, 2011 at 7:19 AM, Brent Yorgey <byorgey(a)seas.upenn.edu> wrote: >> On Sat, Dec 10, 2011 at 07:21:58PM -0800, goodman.m.w(a)gmail.com wrote: >>> ... later I will add support for other statistics, like >>> max, min, sum, prod, etc. > > Specifically, how max and min work with the Monoid implementation. > >> Let's add a Monoid instance for Stats, which specifies how two Stats >> objects should be combined: >> >> instance Monoid Stats where >> mempty = Stats 0 0 >> mappend (Stats sm1 len1) (Stats sm2 len2) = Stats (sm1 + sm2) (len1 + len2) > > Monoids are new to me, but I just read a bit about them in Real World > Haskell, and what you have makes sense. mempty is the identity > function of Stats, and mappend is the associative binary operator. And > thus far these properties hold true for the definition of Stats. If I > add entries for min and max, it seems like it could still be valid. Of > course, for an empty list of numbers, whereas sum can be 0.0 and len > can be 0, there is no corresponding value for max and min, so I make > them Maybe Double with an initial value of Nothing, and modify your > code as follows: > > data Stats = Stats { sm :: Double, mn, mx :: Maybe Double, len :: Int } > deriving Show > > instance Monoid Stats where > mempty = Stats 0.0 Nothing Nothing 0 > mappend (Stats sm1 mn1 mx1 len1) (Stats sm2 mn2 mx2 len2) > = Stats (sm1 + sm2) (min mn1 mn2) (max mx1 mx2) (len1 + len2) > > mkStats x = Stats x (Just x) (Just x) 1 > > This compiles and works for max, but min does not work. I tested in ghci, and: > > Prelude> min (Just 0.1) (Nothing) > Nothing > Prelude> max (Just 0.1) (Nothing) > Just 0.1 > > Which explains the problem. Any idea for the solution? > > -- > -Michael Wayne Goodman > > _______________________________________________ > Beginners mailing list > Beginners(a)haskell.org > http://www.haskell.org/mailman/listinfo/beginners

2 1

Averaging a string of numbers
by goodman.m.w＠gmail.com 11 Dec '11

11 Dec '11

Hi beginners list, I want to take a string of numbers and return the average. It is a string because I expect the input to be from STDIN or a file. I would like it to handle infinite lists. Further, I create the Stats data structure because later I will add support for other statistics, like max, min, sum, prod, etc. The following code works, but I'm not sure it's the proper way to do things in Haskell. For instance, maybe I could use a fold method instead of the explicit recursion in getStats? Could you please let me know if anything looks clunky or could be improved (efficiency, aesthetics, etc)? Thank you -- avg.hs -- sm is sum, len is number of elements data Stats = Stats { sm :: Double, len :: Int } deriving Show getAverage :: String -> Double getAverage str = (sm s / fromIntegral (len s)) where s = (getStats . readNumbers . lines) str getStats :: [Double] -> Stats getStats [] = Stats 0.0 0 getStats [x] = Stats x 1 getStats (x:xs) = Stats (x + (sm s)) ((len s) + 1) where s = getStats xs readNumbers :: [String] -> [Double] readNumbers = map read -- end avg.hs And here's the results in ghci: *Main> :load avg.hs [1 of 1] Compiling Main ( avg.hs, interpreted ) Ok, modules loaded: Main. *Main> getAverage "0" 0.0 *Main> getAverage "1.4\n0.5\n85.2\n30.2\n-10" 21.46 *Main> getAverage "" NaN -- -Michael Wayne Goodman

2 2

correction: hGetLine, not getLine
by MJ Williams 11 Dec '11

11 Dec '11

Apologies guys, I meant, hGetLine, not getLine. Sorry about that. Regards Matthew

1 0

checking file type (without getLine)
by MJ Williams 11 Dec '11

11 Dec '11

Hi guys, When the hSetEncoding is set to utf8, getLine will attempt to read each line as belonging to a UTF-8 and when it encounters an error in the encoding it throws an exception. I wonder if there's a way of checking for errors without getLine. I didn't find any function or functions in System.IO that might do the job. is there another library that I should be looking at? Regards Matthew

1 0

let in GHCI
by Henry Lockyer 11 Dec '11

11 Dec '11

Hello Haskellers Why is it that in GHCI I can do, for example, > replicate 6 5 [5,5,5,5,5,5] > let my6 = replicate 6 > my6 5 [5,5,5,5,5,5] but if I do > sort "bav" "abv" this is ok, but > let mySort = sort > mySort "bav" <interactive>:1:8: Couldn't match expected type `()' with actual type `Char' Expected type: [()] Actual type: [Char] In the first argument of `mySort', namely `"bav"' In the expression: mySort "bav" and/or > mySort [6,5,9] <interactive>:1:13: No instance for (Num ()) arising from the literal `9' Possible fix: add an instance declaration for (Num ()) In the expression: 9 In the first argument of `mySort', namely `[6, 5, 9]' In the expression: mySort [6, 5, 9] This is eluding me at the moment..! ;-) / Henry

3 4

Installing regex-pcre (revisited)
by MJ Williams 11 Dec '11

11 Dec '11

Hi guys, I understand there's a UNIX-like environment that one can install on Windows and it may already include clibpcre that GHC requires for installing regex-pcre. for the life of me I can't remember its name. Does anyone know what I'm referring to? Regards Matthew

3 4