- Haskell-Cafe - Haskell.org

What to say about Haskell?
by Patai Gergely 14 Jul '09

14 Jul '09

Hello all, I was asked to give a one-hour 'introductory' seminar on Haskell. The audience is a bunch of very clever people with a wider than usual perspective on programming and mathematics, and my talk should be rather informational than evangelistic. Which topics do you think I should touch by all means given the short time? Gergely -- http://www.fastmail.fm - Same, same, but different...

10 15

Re: [Haskell-cafe] Program with profiling runs faster than without
by Grigory Sarnitskiy 14 Jul '09

14 Jul '09

14.07.09, 15:16, "Malcolm Wallace" <malcolm.wallace(a)cs.york.ac.uk>: > This is only a guess, but maybe there is a context-qualified CAF-like > value that is being re-evaluated multiple times in the non-profiling > case, but is appropriately monomorphised in the profiling case, so it > is only evaluated once? > Regards, > Malcolm And I guess you are quite right. I knew there is a poorly optimized place in my program, but I didn't touch it for some ideological reasons. But this morning I get how to optimize it in an ideological friendly manner, so now without profiling the program ~ 2 times faster. The speed of profiled version hasn't changed, hence I've manually done some work the compiler does during profiling.

1 0

Program with profiling runs faster than without
by Grigory Sarnitskiy 14 Jul '09

14 Jul '09

Hello! I can't understand why the following happens. ghc --make -fforce-recomp -O2 -fexcess-precision -fvia-C -optc-O2 Run.lhs -o Run -prof -auto-all and time ./Run TestSim seems to be much faster (I got ~4 times faster indeed) than just ghc --make -fforce-recomp -O2 -fexcess-precision -fvia-C -optc-O2 Run.lhs -o Run and time ./Run TestSim with ghc 6.8.2 on pentium 4 2.4 GHz under Ubuntu I was on my way to optimize the program when I've discovered this. The program is rather complicated and badly optimized but still I think it is quite strange to the profiled version to run faster. All necessary files I send with this letter (Tann.7z).

2 2

Re: [Haskell-cafe] Are GADTs what I need?
by Chris Eidhof 14 Jul '09

14 Jul '09

Then you could add a specific constructor for String. The main point is: the case construct only works for values, not for types. There is no typecase construct. If you want to have certain restrictions on the 'a', such as the Show class, you could also do something like this: > data Value where > VWrapper :: (Show a) => a -> Value If you could elaborate a bit on what you're trying to accomplish (from a higher viewpoint) then maybe we can help you some more. -chris On 13 jul 2009, at 17:42, Kev Mahoney wrote: > Thanks, that helps. > > I was hoping to not have to parametrize Value as there is a fair bit > of code to change, and it cascades down through the data structures > (maybe a forall a . Value a will help here?) > > I will have a go at this approach. In case anyone is interested the > code is at http://github.com/KMahoney > > > 2009/7/13 Chris Eidhof <chris(a)eidhof.nl>: >> Hey Kev, >> >> The types are "thrown away" during compile time. Therefore, if you >> have a >> constructor "VWrapper :: a -> Value" nothing is known about that >> "a" when >> you scrutinize it. >> >> What you could do, however, is something like this: >> >>> data Value a where >>> VInt :: Integer -> Value Integer >>> ... >>> VWrapper :: a -> Value a >> >> And then you can write a function doSomething: >> >>> doSomething :: Value String -> String >>> doSomething (VWrapper s) = s >> >> HTH, >> >> -chris >> >> On 13 jul 2009, at 12:41, Kev Mahoney wrote: >> >>> Hi there, >>> >>> I'm currently writing an interpreter that I would like to be able to >>> use with other haskell programs. I would like to be able to pass >>> along >>> arbitrary types though the interpreter. I've seen hints that GADTs >>> can >>> do this, but I am having trouble understanding them. >>> >>> So far, I've learnt you can do this: >>> >>> data Value where >>> VInt :: Integer -> Value >>> ... >>> VWrapper :: a -> Value >>> >>> which can let you encode arbitrary 'dynamic' types into Value. I was >>> hoping to be able to pattern match to get the value out again e.g. >>> >>> doSomething :: Value -> .... >>> doSomething (VWrapper String s) = ..... >>> >>> Also, anything that can help me out with GADTs in general will be >>> much >>> appreciated. >>> >>> Thanks, >>> Kevin. >>> _______________________________________________ >>> Haskell-Cafe mailing list >>> Haskell-Cafe(a)haskell.org >>> http://www.haskell.org/mailman/listinfo/haskell-cafe >> >>

6 6

Removing polymorphism from type classes (viz. Functor)
by George Pollard 14 Jul '09

14 Jul '09

Ok, so I have a small idea I'm trying to work on; call it a Prelude-rewrite if you want. For this I want to be able to have the hierarchy Functor → Applicative → Monad. For Functor, I would like to be able to implement it for a wider variety of types, as there are types which have aren't polymorphic which would also benefit from having an instance. My running example for this set of types is ByteString; the module contains the method: map ∷ (Word8 → Word8) → ByteString → ByteString However, we cannot use this for Functor because ByteString isn't polymorphic. To get around this, I devised the following: Introduce a type family which represents ‘points’ inside the type: type family Point f ∷ ★ For ByteString we have: type instance Point ByteString = Word8 For a polymorphic example (lists) we have: type instance Point [a] = a Now Functor becomes: class SimpleFunctor f where fmap ∷ (Point f → Point f) → (f → f) However, this doesn't allow for the existence of functions with the type (a → b). I need to introduce another type into the class: class Functor f g where fmap ∷ (Point f → Point g) → (f → g) But having two types isn't very nice (for one thing we can't introduce a fundep because for lists as it fails one of the coverage conditions), so introduce another type family to represent types which can be produced by giving a free variable: type Subst f a ∷ ★ type Subst [a] b = [b] type Subst ByteString b = ByteString class Functor f where fmap ∷ (Point f → Point (Subst f a)) → (f → Subst f a) I'm not sure how much of a hack this is, or if there is a better way. It seems to be OK... Now I want to implement Applicative. It would make sense to have ‘return’ be split out into a separate class, because this can be restricted in a similar way to Functor: class Pointed f where return ∷ Point f → f instance Pointed [a] where return x = [x] instance Pointed ByteString where return = BS.singleton Now, I want to be able to restrict Applicative to things which have [Pointed f, and forall a b. Point f ~ (a → b)]. At the moment I can't figure this out because I believe it would require something like the ‘quantified contexts’ proposal: class (Pointed f, ∀ a b. Point f ~ (a → b)) ⇒ Applicative f where ... I could have something like: class (Pointed f, Point f ~ (a → b)) ⇒ Applicative f a b where apply ∷ f → Subst f a → Subst f b This is still not very nice, because it requires two more type variables in the class, and the non-type-families version is far more straightforward... in fact, it makes sense for the Applicative class to have a polymorphic type because it must be able to have ‘return’ applied to arbitrary functions (remember [fmap f xs ≡ return f `apply` xs]). So back to: class Applicative f where apply ∷ f (a → b) → f a → f b But then ‘return’ cannot be added via a superclass restriction to Pointed! I seem to have painted myself into a corner. Does anyone see a better way to go about this? Thanks, - George

3 3

Questions about haskell CPP macros
by Matthew Elder 13 Jul '09

13 Jul '09

Hello Cafe, I am trying to improve the error reporting in my sendfile library, and I know I can find out the current file name and line number with something like this: {-# LANGUAGE CPP #-} main = putStrLn (__FILE__ ++ ":" ++ show __LINE__) This outputs: test.hs:2 Unfortunately, if your file is in a hierarchy of folders, this flat file name doesn't give much context. Is there a macro to find out the current module? IE if I had a module Foo.Bar.Car.MyModule, I would like to be able to output something like this on error: Foo.Bar.Car.MyModule:2 Any help is appreciated! Thanks, Matt -- Need somewhere to put your code? http://patch-tag.com Want to build a webapp? http://happstack.com

4 4

Are GADTs what I need?
by Kev Mahoney 13 Jul '09

13 Jul '09

Hi there, I'm currently writing an interpreter that I would like to be able to use with other haskell programs. I would like to be able to pass along arbitrary types though the interpreter. I've seen hints that GADTs can do this, but I am having trouble understanding them. So far, I've learnt you can do this: data Value where VInt :: Integer -> Value ... VWrapper :: a -> Value which can let you encode arbitrary 'dynamic' types into Value. I was hoping to be able to pattern match to get the value out again e.g. doSomething :: Value -> .... doSomething (VWrapper String s) = ..... Also, anything that can help me out with GADTs in general will be much appreciated. Thanks, Kevin.

4 4

Re: [Haskell-cafe] Leaner Haskell.org frontpage
by Don Stewart 13 Jul '09

13 Jul '09

ttencate: > On Thu, Jul 9, 2009 at 18:33, Don Stewart<dons(a)galois.com> wrote: > > ttencate: > >> Are there any kind of hard statistics and analytics that we can base > >> this discussion upon? There is always room for improvement, but > >> stumbling around in the dark making blind guesses may not be the best > >> way to go. Although I personally feel that Lenny's proposed page is an > >> improvement, statistics could tell us what actual people actually use > >> the site for. > > Thanks Don, I should have thought of that. It's a start! No matter what we decide, I'd like to advocate for maintaining the RSS feed of Hackage uploads on the front page -- this is what makes the page seem alive and active, and was introduced in response to John Hughes complaining the page was always out of date. (Similar to the Arch Linux package updates: http://www.archlinux.org/) -- Don

11 15

Type families and polymorphism
by Jeremy Yallop 13 Jul '09

13 Jul '09

Why does compiling the following program give an error? > {-# LANGUAGE TypeFamilies, RankNTypes #-} > > type family TF a > > identity :: (forall a. TF a) -> (forall a. TF a) > identity x = x GHC 6.10.3 gives me: Couldn't match expected type `TF a1' against inferred type `TF a' In the expression: x In the definition of `identity': identity x = x -- The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336.

5 5

Hylomorphisms (was: [Haskell-cafe] excercise - a completely lazy sorting algorithm)
by Raynor Vliegendhart 13 Jul '09

13 Jul '09

On 7/12/09, Heinrich Apfelmus <apfelmus(a)quantentunnel.de> wrote: > Raynor Vliegendhart wrote: > > On 7/9/09, Heinrich Apfelmus <apfelmus(a)quantentunnel.de> wrote: > >> Of course, some part of algorithm has to be recursive, but this can be > >> outsourced to a general recursion scheme, like the hylomorphism > >> > >> hylo :: Functor f => (a -> f a) -> (f b -> b) -> (a -> b) > >> hylo f g = g . fmap (hylo f g) . f > >> > > > > Is that definition of hylo actually usable? A few on IRC tried to use > > that definition for a few examples, but the examples failed to > > terminate or blew up the stack. > > The implementation of quicksort with hylo works fine for me, given > medium sized inputs like for example quicksort (reverse [1..1000]) . > > What were the examples you tried? > One of the examples I tried was: hylo (unfoldr (\a -> Just (a,a))) head $ 42 This expression fails to determinate. Here are two examples copumpkin tried on IRC: <copumpkin> > let hylo f g = g . fmap (hylo f g) . f in hylo (flip replicate 2) length 5 <lambdabot> 5 <copumpkin> > let hylo f g = g . fmap (hylo f g) . f in hylo (flip replicate 2) sum 5 <lambdabot> * Exception: stack overflow

3 2