- Beginners - Haskell.org

Equivalent of IO Monad in other functional languages?
by Simon Kitching 17 Mar '15

17 Mar '15

[Note: learning haskell, not familiar with other functional languages] I have a question about handling impure operations in functional or semi-functional languages other than Haskell (eg OCaml, F#, Scala, ..). Haskell uses the IO monad to (a) cleanly separate pure and impure code, and (b) prevent function-calls from being reordered when they have side-effects that must happen in the right order. How do other languages handle this? AIUI, whether a Haskell function is 'pure' or not can be seen from its method signature : if it returns IO then it is impure. All standard library functions that have side-effects (read/write/etc) return IO, and any function that uses an IO value must return type IO. Do other functional languages (1) use a similar approach, or (2) not provide a way to be certain that a function is 'pure' (side-effect-free)? And AIUI the Haskell compiler/runtime can postpone evaluation of any function (laziness), or reorder function calls whever it thinks this good. However using the monad design pattern (deliberately) prevents this (each operation passes its value as a parameter to the subsequent operation, thus forcing an order of evaluation). Do other languages (a) not support compiler/runtime reordering of functions, or (b) have some other way of preventing functions with side-effects from being reordered? Thanks.. Simon

8 15

Re: [Haskell-beginners] Saving intermediate calculations
by Henk-Jan van Tuyl 17 Mar '15

17 Mar '15

On Thu, 12 Mar 2015 16:34:19 +0100, Abhinav Kalawatia <kalawatia_abhinav(a)yahoo.com> wrote: > Hi Henk, > I hope you are well :) > Thanks for your help with the solution for saving the intermediate > calculations. Henk, could you please help me understand the following > code snippet? > ewma1 a (x:xs) = reverse $ foldl' f [x] xs > where f m@(x':xs') n = ((a * n) + ((1 - a) * x')):m You can do mathematical substitution: ewma1 a (x:xs) = reverse $ foldl' f [x] xs where f m@(x':xs') n = ((a * n) + ((1 - a) * x')):m + (definition of foldl, which performs the same calculation as foldl', though not strict) foldl f z [] = z foldl f z (x:xs) = foldl f (f z x) xs => (substitute definition of foldl in ewma1) ewma1 a (x1:x2:xs) = reverse $ foldl' f (f [x1] x2) xs where f m@(x':xs') n = ((a * n) + ((1 - a) * x')):m => (substitute f) ewma1 a (x1:x2:xs) = reverse $ foldl' f (((a * x2) + ((1 - a) * x1)):[x1]) xs where f m@(x':xs') n = ((a * n) + ((1 - a) * x')):m Repeat the last two steps until the list is processed. There is a tool to display this automatically: Hat, but I could not get it to run properly on my Windows PC. The homepage is at: http://projects.haskell.org/hat/ Regards, Henk-Jan -- Folding@home What if you could share your unused computer power to help find a cure? In just 5 minutes you can join the world's biggest networked computer and get us closer sooner. Watch the video. http://folding.stanford.edu/

1 0

testing IO code
by Maurizio Vitale 17 Mar '15

17 Mar '15

suppose I have a restricted IO monad, RIO that only exposes readFile. and then I have a monad SIO that will eventually provide a virtual file system from a map path->content, also with a readFile function returning SIO(String). What is the way to write a function parseFile that can operate in both monads so that I can use SIO for testing? should I define a third monad CompileMonad that has instances for both RIO and SIO and then having parseFile :: CompileMonad ast? Thanks, Maurizio

2 2

Re: [Haskell-beginners] [Haskell-cafe] Mathematical functions with multiple arguments
by Sumit Sahrawat, Maths & Computing, IIT (BHU) 15 Mar '15

15 Mar '15

The fixed-vector package uses a similar technique. The only trouble I'm having is with converting Vec v (Double, Double) to [(Double, Double)] for further use. I don't want to change all the code, but only the part where the user provides me with arguments. I'll keep looking into it. Thanks for the help. On 12 March 2015 at 04:44, David Feuer <david.feuer(a)gmail.com> wrote: > There are a lot of ways to do this sort of thing, and which one you choose > will depend on exactly what you're trying to do. For example, you can write > something vaguely like > > data Nat = Z | S Nat > data SL (n :: Nat) a where > Nil :: SL Z > Cons :: a -> SL n a -> SL (S n) a > > plot :: forall (n::Nat) . (SL n Double -> Double) -> > SL n (Double, Double) -> IO () > On Mar 11, 2015 5:45 PM, "Sumit Sahrawat, Maths & Computing, IIT (BHU)" < > sumit.sahrawat.apm13(a)iitbhu.ac.in> wrote: > >> Hi everybody, >> >> I have a function of type >> >> plot :: ([Double] -> Double) -- A function to plot >> -> [(Double, Double)] -- Range for all arguments >> -> IO () >> >> I want to enforce the fact that ranges for all arguments should be >> provided. >> Is there a way to make the type system enforce it? >> >> -- >> Regards >> >> Sumit Sahrawat >> >> _______________________________________________ >> Haskell-Cafe mailing list >> Haskell-Cafe(a)haskell.org >> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe >> >> -- Regards Sumit Sahrawat

2 3

Problems installing lambdabot
by Joel Neely 13 Mar '15

13 Mar '15

I am using a recent install of Haskell platform (GHC/GHCi version 7.8.3) in OS X 10.9.5, and have been working through Real World Haskell and last year's CIS194 lectures and exercises. I attempted to follow the instructions for installing lambdabot (as given by https://wiki.haskell.org/Lambdabot ) via cabal install lambdabot. After 24 1/2 minutes, the process ended with a little over a dozen error messages apparently tracing to two occurrences of ExitFailure 1: MonadRandom-0.3.0.1 failed during the building phase. The exception was: ExitFailure 1 regex-pcre-0.94.4 failed during the building phase. The exception was: ExitFailure 1 The remaining messages were from components dependent on one or the other of those two. The logged failure for regex-pcre-0.94.4 included the following: Preprocessing library regex-pcre-0.94.4... Wrap.hsc:148:10: fatal error: 'pcre.h' file not found #include <pcre.h> ^ 1 error generated. compiling dist/build/Text/Regex/PCRE/Wrap_hsc_make.c failed (exit code 1) and for MonadRandom-0.3.0.1 included: Preprocessing library MonadRandom-0.3.0.1... [1 of 2] Compiling Control.Monad.Random.Class ( Control/Monad/Random/Class.hs, dist/build/Control/Monad/Random/Class.o ) [2 of 2] Compiling Control.Monad.Random ( Control/Monad/Random.hs, dist/build/Control/Monad/Random.o ) [1 of 2] Compiling Control.Monad.Random.Class ( Control/Monad/Random/Class.hs, dist/build/Control/Monad/Random/Class.p_o ) [2 of 2] Compiling Control.Monad.Random ( Control/Monad/Random.hs, dist/build/Control/Monad/Random.p_o ) /Users/joel/.ghc/x86_64-darwin-7.8.3/package.conf.d/package.cache: openBinaryFile: does not exist (No such file or directory) I am at a loss for corrective actions. I would assume that I need to locate the pcre.h header file and put it in the "right place", but was unable to find the location of Wrap_hsc_make.c. And I don't know how to address the non-existence of openBinaryFile in package.cache. Thanks in advance for any advice on correcting the above problems! -jn- -- Beauty of style and harmony and grace and good rhythm depend on simplicity. - Plato

2 1

Dipping Toes Into Haskell
by Timothy Washington 13 Mar '15

13 Mar '15

To get started, I'm trying to implement a simple *tictactoe* game. And I would like to be able to represent a Piece on the board, as either the string "X" or "O". This is what I have so far. module Main where data Piece = X | O type Row = [Piece] type Board = [Row] -- put an X or O in a position move :: Board -> Piece -> Board move board piece = board -- check win vertically -- check win horizontally -- check win diagonally main :: IO () main = putStrLn "Hello World" *A)* Now, I'd like to be able to *load code interactively*, preferably within emacs. However I don't have access to my types with *ghci* or *ghc-mod (Interactive-Haskell)*. In either case, this call fails with the below error. let p = Piece X <interactive>:20:9-13: Not in scope: data constructor `Piece' *B)* And how do I make a *custom datatype* that's one of two strings (enumeration of either "X" or "O"). Cabal builds and runs the abouve code, so I know it can compile. But I'm confused as to where X or O is defined, and how I would supply it as an input. *C)* Finally, how do we update nested lists in Haskell. I want the move function to take a Board, Piece, and Position, and return a Board. I see some results from Hoogle <https://www.haskell.org/hoogle/?hoogle=update>. Is this where Lenses <https://www.haskell.org/hoogle/?hoogle=lens> or Zippers <https://www.haskell.org/hoogle/?hoogle=zipper> come into play? Thanks Tim Washington Interruptsoftware.com <http://interruptsoftware.com>

4 5

Re: [Haskell-beginners] [Haskell-cafe] Mathematical functions with multiple arguments
by Sumit Sahrawat, Maths & Computing, IIT (BHU) 12 Mar '15

12 Mar '15

It's done. fixed-vector provides many implementations for Vec. The one I was using (Primitive) requires primitive constraint on data types. I shifted to Unboxed ones, and now it's working. This means that I am able to use toList with Vec v (Double, Double), which was previously complaining of missing constraints. Also, I'll slowly convert the whole thing. I wanted to just see if it works initially. Thanks, once again. On 12 March 2015 at 04:58, David Feuer <david.feuer(a)gmail.com> wrote: > I'd urge you to *try* to maintain that safety throughout, but you can try > something like this if necessary (I'm not at my computer, so I can't test > it): > > {-#LANGUAGE FlexibleContexts, FlexibleInstances #-} > instance Foldable (SL Z) where > foldr _ n Nil = n > > instance Foldable (SL n) => Foldable (SL (S n)) where > foldr c n (Cons x xs) = c x (foldr c n xs) > > In fact, I would expect whatever package you're looking at to offer this. > Then you can just use Data.Foldable.toList. > On Mar 11, 2015 7:16 PM, "Sumit Sahrawat, Maths & Computing, IIT (BHU)" < > sumit.sahrawat.apm13(a)iitbhu.ac.in> wrote: > >> The fixed-vector package uses a similar technique. The only trouble I'm >> having is with converting Vec v (Double, Double) to [(Double, Double)] for >> further use. I don't want to change all the code, but only the part where >> the user provides me with arguments. >> >> I'll keep looking into it. Thanks for the help. >> >> On 12 March 2015 at 04:44, David Feuer <david.feuer(a)gmail.com> wrote: >> >>> There are a lot of ways to do this sort of thing, and which one you >>> choose will depend on exactly what you're trying to do. For example, you >>> can write something vaguely like >>> >>> data Nat = Z | S Nat >>> data SL (n :: Nat) a where >>> Nil :: SL Z >>> Cons :: a -> SL n a -> SL (S n) a >>> >>> plot :: forall (n::Nat) . (SL n Double -> Double) -> >>> SL n (Double, Double) -> IO () >>> On Mar 11, 2015 5:45 PM, "Sumit Sahrawat, Maths & Computing, IIT (BHU)" < >>> sumit.sahrawat.apm13(a)iitbhu.ac.in> wrote: >>> >>>> Hi everybody, >>>> >>>> I have a function of type >>>> >>>> plot :: ([Double] -> Double) -- A function to plot >>>> -> [(Double, Double)] -- Range for all arguments >>>> -> IO () >>>> >>>> I want to enforce the fact that ranges for all arguments should be >>>> provided. >>>> Is there a way to make the type system enforce it? >>>> >>>> -- >>>> Regards >>>> >>>> Sumit Sahrawat >>>> >>>> _______________________________________________ >>>> Haskell-Cafe mailing list >>>> Haskell-Cafe(a)haskell.org >>>> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe >>>> >>>> >> >> >> -- >> Regards >> >> Sumit Sahrawat >> > -- Regards Sumit Sahrawat

1 0

Re: [Haskell-beginners] [Haskell-cafe] Mathematical functions with multiple arguments
by Sumit Sahrawat, Maths & Computing, IIT (BHU) 12 Mar '15

12 Mar '15

I was wondering whether it was possible using fixed-vector, ut dependent types seem to be a new thing I can experiment with. Thanks for the quick reply. On 12 March 2015 at 03:26, Adam Bergmark <adam(a)bergmark.nl> wrote: > If I understand you correctly you seem to want dependent types, this > article uses the same example you need, promoting the length of a > vector/list to the type level: > https://www.fpcomplete.com/user/konn/prove-your-haskell-for-great-safety/de… > > You'd end up with `plot :: (Vector n Double -> Double) -> Vector n > (Double, Double) -> IO ()' where `n' is the length of the vector. > > HTH, > Adam > > > On Wed, Mar 11, 2015 at 10:45 PM, Sumit Sahrawat, Maths & Computing, IIT > (BHU) <sumit.sahrawat.apm13(a)iitbhu.ac.in> wrote: > >> Hi everybody, >> >> I have a function of type >> >> plot :: ([Double] -> Double) -- A function to plot >> -> [(Double, Double)] -- Range for all arguments >> -> IO () >> >> I want to enforce the fact that ranges for all arguments should be >> provided. >> Is there a way to make the type system enforce it? >> >> -- >> Regards >> >> Sumit Sahrawat >> >> _______________________________________________ >> Haskell-Cafe mailing list >> Haskell-Cafe(a)haskell.org >> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe >> >> > -- Regards Sumit Sahrawat

1 0

Mathematical functions with multiple arguments
by Sumit Sahrawat, Maths & Computing, IIT (BHU) 12 Mar '15

12 Mar '15

Hi everybody, I have a function of type plot :: ([Double] -> Double) -- A function to plot -> [(Double, Double)] -- Range for all arguments -> IO () I want to enforce the fact that ranges for all arguments should be provided. Is there a way to make the type system enforce it? -- Regards Sumit Sahrawat

1 0

What's the type of function "abs"?
by Zhang Hengruo 11 Mar '15

11 Mar '15

I'm a newbie, knowing a little about Haskell, and hope my question isn't very silly... ===================================================================== My absolute value function looks like this: abs2 :: Num a => a -> a abs2 n = if n >= 0 then n else 0 - n GHCi tells me that I should add Ord type class to its definition. Well, it's true. It has used relational operators and Num isn't a subclass of Ord. However, when I input ":t abs" in GHCi, the interpreter shows "abs :: Num a => a -> a". I read GHC/Num.lhs and find that abs is defined as "abs :: a -> a" and has no concrete content. So I think the abs function is written in C as a module to implement directly and the type of abs just follows its class GHC.Num. Is it right? Or there are any other reasons? Thanks, Hengruo

3 3