October 2009 - Haskell-Cafe

more improvable randomness
by Michael Mossey 15 Oct '09

15 Oct '09

I wrote some code to model the keystroke-to-keystroke delay in a person typing, with pseudorandomness. There are two kinds of delays.. one is a very small delay as the person reaches for a new key (call this 'reach' delays), the other is a larger delay that represents a pause to think or to take a break (call this 'break' delays). My thought is that I could create an infinite series of reach delays and an infinite series of break delays and zipWith (+). The inifinite series of break delays would look like: - The overall pattern is a large number of zeros, in the range (N,M) followed by a single positive number, chosen from a distribution D... and then repeat. Here's some code. I used the Gen monad to gain access to the 'frequency' function, which is useful, but I had trouble figuring out how to put the whole algorithm into Gen. I also looked at Rand. The problem is that if I want to create the list by 'sequence'ing a list of monads, they need to have state that remembers how many zeros are left to go. breakSeries :: Int -> Int -> StdGen -> [Float] breakSeries lowerB upperB gen = let (n,gen1) = randomR (lowerB,upperB) gen (gen2,gen3) = split gen1 delay = generate 1 gen2 breakM in replicate n 0 ++ [delay] ++ breakSeries lowerB upperB gen3 breakM :: Gen Float breakM = frequency [ (10, choose( 1::Float , 2)) , (10, choose( 4::Float , 6)) ] test = (print . take 100 . breakSeries 2 4 ) =<< newStdGen

2 1

Re: [Haskell-cafe] Need help with ghc static compile for cgi using mysql
by Austin King 15 Oct '09

15 Oct '09

For now, I've given up on cheap hosting (via statically compiled CGI). I've created a GHC 6.10.4, cabal-install, Ubuntu 9.04 ec2 instance + MySQL and it works well. Running this plus EBS and Elastic IP will run a little over $60 a month. Ouch. The work going into Haskell Platform is amazing, but... Haskell needs a better commodity hosting story, to allow for casual, fun web hacking. For the record, I tried nearlyfreespeech.net and dreamhost. I wasn't able to install ghc and cabal due to: * FreeBSD GHC support (old version) * FreeBSD library support (no HDBC) * Linux - Missing 'dev' libraries like zlib and not having root access

4 3

CBN, CBV, Lazy in the same final tagless framework
by oleg＠okmij.org 15 Oct '09

15 Oct '09

Actually it is possible to implement all three evaluation orders within the same final tagless framework, using the same interpretation of types and reusing most of the code save the semantics of lam. That is where the three orders differ, by their own definition. In call-by-name, we have lam f = S . return $ (unS . f . S) In call-by-value, we have lam f = S . return $ (\x -> x >>= unS . f . S . return) In call-by-need, we have lam f = S . return $ (\x -> share x >>= unS . f . S) In CBV, the function first evaluates its argument, whether the value will be needed or not. In call-by-need, we arrange for the sharing of the result should the computation be evaluated. Here is an illustrative test t2 :: HOAS exp => exp IntT t2 = (lam $ \z -> lam $ \x -> let_ (x `add` x) $ \y -> y `add` y) `app` (int 100 `sub` int 10) `app` (int 5 `add` int 5) t2SN = runName t2 >>= print {- *CB> t2SN Adding Adding Adding Adding Adding Adding Adding 40 -} In CBN, the result of subtraction was not needed, and so it was not performed OTH, (int 5 `add` int 5) was computed four times. t2SV = runValue t2 >>= print {- *CB> t2SV Subtracting Adding Adding Adding 40 -} In CBV, although the result of subtraction was not needed, it was still performed. OTH, (int 5 `add` int 5) was computed only once. t2SL = runLazy t2 >>= print {- *CB> t2SL Adding Adding Adding 40 -} Now, Lazy is better than both CBN and CBV: subtraction was not needed, and it was not performed. All other expressions were needed, and evaluated once. The complete code is at http://okmij.org/ftp/tagless-final/CB.hs

5 6

Re: [Haskell-cafe] Finally tagless - stuck with implementation of "lam"
by Jacques Carette 15 Oct '09

15 Oct '09

(sorry for the slow reply on this topic...) Robert Atkey and Oleg presented some very interesting code in response to your query. But some of you might (and should!) be asking "why on earth did Jacques use unsafePerformIO?", especially when neither Robert nor Oleg did. Simply put: I answered your question exactly as asked, while they answered the question you *should* have asked. At the exact 'type' involved in your question, there is no good answer; if you want an instance of lam at a monadic type 'directly', you need to 'extract' from the monad. But the point is that that isn't really the right question. Rather than requiring complete parametric polymorphism in the answer-type, if you allow a little non-uniformity through a simple type family, the shift in types is sufficient to no longer require a monadic 'extract' at all. But even once you've gotten to that point, that's not enough, because at that point you still have the issue of what calling convention (by value, name or need) to use. And figuring that out is rather fun, so you got detailed answer from Robert and Oleg explaining that part in detail. Robert carefully used IntT and :-> to label the different types, and Oleg's code (http://okmij.org/ftp/tagless-final/CB.hs) showed how these were really just 'labels', which are useful mnemonics for humans, but not essential. If you dig into our JFP paper describing "finally tagless" in all its gory details, you'll see that we use a poor man's version of type families there too (see section 4.3). Jacques

2 1

Relational Algebra
by Günther Schmidt 15 Oct '09

15 Oct '09

Hi all, has anybody already developed an EDSL for relational algebra? HaskellDB does that, but its current implementation only allows for generating SQL, are there other implementations? Günther

2 2

ANNOUNCE: hs-ffmpeg 0.3.2 - Bindings to FFMpeg library
by Vasyl Pasternak 15 Oct '09

15 Oct '09

Hello, I am glad to announce next release of hs-ffmpeg library. Now you could download it from the Hackage along with the ffmpeg-tutorials, which show capabilities of this library. The installation process is a bit tricky now, so welcome to my blog post http://progandprog.blogspot.com/2009/10/video-processing-on-haskell-easy.ht…, where I am describing how to build and install library. Best regards, Vasyl Pasternak

1 0

(no subject)
by Stavenga, G.C. 15 Oct '09

15 Oct '09

Hi, I'm just started to learn Haskell. Coming from a programming contest background (where it is important to be able to solve problems in a small amount of code) I'm wondering what the best way is for simple IO. A typical input file (in a programming contest) is just a bunch of numbers which you want to read one by one (sometimes interspersed with strings). In C/C++ this is easily done with either scanf or cin which reads data separated by spaces. In Haskell I have not found an equally satisfactionary method. The methods I know of 1) Stay in the IO monad and write your own readInt readString functions. A lot of code for something easy. 2) Use interact together with words and put the list of lexemes in a State monad and define getInt where at least you can use read. 3) Use ByteString.Char8 which has readInt (but I couldn't find a readString). But one has to put it also in a State monad. I think that there must be standard function that can do this. What do experienced Haskellers use? Thanks in advance

10 11

Monotype error
by Martijn van Steenbergen 15 Oct '09

15 Oct '09

Dear café, > {-# LANGUAGE Rank2Types #-} > {-# LANGUAGE ImpredicativeTypes #-} > > type Void = forall a. a > > newtype Mono a = Mono { runMono :: [Void] } > > beep :: Mono a -> Mono a > beep (Mono vs) = Mono (map undefined vs) Compiling this with GHC results in: > Monotype.hs:9:28: > Cannot match a monotype with `Void' > Expected type: Void > Inferred type: a What does this error mean and why does the code not compile? Thanks! Martijn.

4 3

newtype deriving Alternative
by Martijn van Steenbergen 15 Oct '09

15 Oct '09

Hello café, I've never written an Alternative instance for a newtype yet that doesn't look like this: > instance Alternative T where > empty = T empty > T x <|> T y = T (x <|> y) Why does newtype deriving not work for Alternative? (It works fine for Monoid.) Thanks, Martijn.

3 4

Fuzzy Logic / Linguistic Variables
by Neal Alexander 14 Oct '09

14 Oct '09

So i was reading "Programming Game AI by Example" by Mat Buckland (http://www.ai-junkie.com/books/toc_pgaibe.html) and decided to rewrite his chapter on Fuzzy logic in haskell (from C++). My initial impression: its one of those scenarios where OOP grossly over complicates things Heres an example taken from the book: An agent needs to decide what weapon to use based on his distance from the target and the amount of ammo each has. The result is in the desirability domain (0-100). http://code.haskell.org/~hexpuem/fuzzyLogic/AI/Logic/Fuzzy/Example.hs An excerpt: ------------------------------------ weapons = [ Weapon "RocketLauncher" 9 rocketLauncher, Weapon "ShotGun" 13 shotgun, Weapon "AssaultRifle" 120 assaultRifle, Weapon "SniperRifle" 7 sniperRifle, Weapon "Knife" 1 knife ] chooseWeapon dist = maximum ratings where ratings = [ (f dist ammo, n) | Weapon n ammo f <- weapons ] ------------------------------------ shotgun :: Double -> Double -> Double shotgun dist ammo = unfuzz desireDom $ rules (fuzz distDom dist) (fuzz ammoDom ammo) where rules :: Fuzzy Distances -> Fuzzy Ammo -> FL Desirability rules distance ammo = do distance `is` SniperSuited =>> Pointless distance `is` Far =>> Undesirable distance `is` Medium =>> Undesirable distance `is` Melee =>> Undesirable (distance `fairly` Near) & (ammo `fairly` High) =>> VeryDesirable (distance `fairly` Near) & (ammo `fairly` Good) =>> Desirable (distance `fairly` Near) & (ammo `is` Low) =>> Undesirable ------------------------------------ Full code at http://code.haskell.org/~hexpuem/fuzzyLogic/. Suggestions welcome - maybe it'd be useful to upload to hackage at some point. It only supports triangle, shoulder, and singleton memberships at the moment.

2 1