- Haskell-Cafe - Haskell.org

Reasons for Super-Linear Speedup
by burak ekici 05 Mar '12

05 Mar '12

Dear List; I have parallelized RSA decryption and encryption schemes by using second generation strategies of Haskell p.l. In that case, what I got in the sense of speed up was nearly 10 times of better performances (on a quad-core CPU with 8M cache) in the parallel evaluation of 125K long plain text with 180-bit of an encryption key, in comparison with the serial evaluation, abnormally. All my thoughts are in the direction of getting wrong serial-time performances. My question here is that is there any difference between serial and parallel evaluation of any arbitrary computation in terms of "cache usage methodologies" in Haskell compiler design. If no; what could be other possible reasons? I would be much more than appreciated, if somebody brings an acceptable pinpoint for the issue. Many thanks in advance. Kind regards Burak.

2 1

[haskell-cafe] Question about 64bit target on Windows platform
by C K Kashyap 05 Mar '12

05 Mar '12

Hi All, Can someone please let me know if there is a 64bit target on Windows on the horizon for GHC? I am trying to push for changing the current implementation in my organization in C++ to Haskell - Our primary targets are Windows and Mac. Not being able to generate 64bit DLL's on Windows would be a big bottleneck. Regards, Kashyap

3 4

Looking for an heap-like data structure allowing efficient update
by Arnaud Bailly 05 Mar '12

05 Mar '12

Hello Cafe, I am looking for a data structure that would have the following informal properties: - allow efficient retrieval of minimum element for some ordering relation on a computed property of the elements - allow efficient update of any element wrt to this property I have first tried to use a PQueue from fingertree package but it does not allow deletion/update of arbitrary elements. I then started to try combining this with a map but ran into the same issue. I skimmed through Okasaki's book (oldies but goldies...) and the basic finger tree structure but these does not seem to fit my needs. Thanks for your help Regards, Arnaud Bailly

2 2

Re: [Haskell-cafe] FreeSect -- generalised sections syntax extension
by Ras Far 05 Mar '12

05 Mar '12

So we can have ... chiastic freesects? This sounds like too much fun! I find chiasmus is a term from linguistics? From [1]: "The elements of simple chiasmus are often labelled in the form ABBA, where the letters correspond to grammar, words, or meaning." So we've got love, too (Aramaic word “Abba” = love of God or so...) Also coincidental, since a lot of the literary instances are apparently found in sacred texts (although Coleridge comes up a lot too, cough). So many coincidences around this. I'd written up the idea for the Haskell extension in 2003 (calling them "arbitrary sections"), but only decided to implement them about a week ago. Had just gotten started about an hour earlier, when eyebloom on #haskell asks about the same thing, right down to the _ syntax. (I offered him collaboration but he didn't get back to me, hope you're not sore eyebloom, anyhow I don't mind you were maybe too busy.) Then I get hacking Language.Haskell.Exts and -- the most recent extension already added is "TupleSections"! And there's no other extension so related to this idea. Now the serendipity with Wren's linguistic research. And even the freegeek domain. Fantastic. Wren, your slides look great, thanks for the link, I think I'll read them tonight. Do you have a paper version of the slides at all? Slides can be a bit terse on their own. Yours Chiastically, Andrew [1] http://en.wikipedia.org/wiki/Chiastic_structure

3 3

Impact of "try" on Parsec performance
by Omari Norman 04 Mar '12

04 Mar '12

The Parsec documentation says that Parsec performs best on predictive grammars, and Parsec does not backtrack by default to improve performance (though this also improves error messages). On the other hand, I notice that attoparsec and polyparse backtrack by default, and attoparsec claims to be faster than Parsec (I can't remember if polyparse makes this claim). The Parsec documentation says there is a penalty for using try and suggests left-factoring grammars when possible. Real World Haskell says that excessive use of try can degrade Parsec performance quite significantly. On the other hand, some of the combinators provided with Parsec, such as notFollowedBy, use try. So my question is: what is the practical impact of using try? I ask because as a novice I have a simple grammar that could, perhaps, be factored to avoid use of try, but it's quite a brain puzzle for me and I wonder if it would ever be worth it. Does it matter how many characters "try" would have to store, or how often it would run?

7 7

IO() and other datatypes
by Kevin Clees 04 Mar '12

04 Mar '12

Dear Haskell programmers, I'm very confused, because I really don't know how to handle with IO's and other datatypes, such as Int or String. If I want to build a haskell program can I only use IO() method outputs ? How can I "give" a Int result from a different method back to the main? Would this mean that I have to create only IO() outputs? Is this correct? For example: -- A User has to choose something, so I need a IO() datatype main :: IO() main = do [...] let listtournementTime = [8,20,10,15] -- This is wrong: Couldn't match expected type `IO t0' with actual type `Int' -- The results of the method dauer is a Int. Do I have to transform the method to an IO() Output a <- Dauer listtournementTime [...] -- the methods dauer:: [Int] -> Int dauer (x:xs) | laenge(x:xs) == 1 = 0 | mod (laenge (x:xs)) 2 == 0 = (tmp x xs) + dauer xs | mod (laenge (x:xs)) 2 /= 0 = dauer xs | otherwise = 999999 -- failure tmp:: Int -> [Int] -> Int tmp y (x:xs) = x-y laenge :: [a] -> Integer laenge [] = 0 laenge (x:xs) = 1 + laenge xs The last three methods are working correct, if I directly put some data into the methods, like: dauer [10,15] ===Result===> 5 Thank you for any help Best greetings from Namibia

2 1

ANNOUNCE: Haddock 2.10.0
by David Waern 04 Mar '12

04 Mar '12

-------------------------------------------- -- Haddock 2.10.0 -------------------------------------------- Version 2.10.0, which was released with GHC 7.4.1, is now uploaded on Hackage. Here are the release notes: -------------------------------------------- -- Changes in 2.10.0 -------------------------------------------- * Require GHC >= 7.4 * Safe Haskell indications on module pages * Type declarations on identifiers no longer necessary * Add flag --interface-version * Warn when comment refers to in-scope identifier without documentation * Bug fix: links to out-of-scope things (#78) * Bug fix: module references to other packages work again -------------------------------------------- -- Links -------------------------------------------- Hackage page: http://hackage.haskell.org/package/haddock-2.10.0 Bugtracker and wiki: http://trac.haskell.org/haddock Mailing list: haddock(a)projects.haskell.org Code repository: http://darcs.haskell.org/haddock.git -------------------------------------------- -- Contributors -------------------------------------------- The following people contributed patches to this release: Ian Lynagh Jose Pedro Magalhaes Simon Peyton-Jones Ryan Newton David Terei David Waern Also thanks to Simon Hengel for spotting a serious bug before the release. -------------------------------------------- -- Get Involved -------------------------------------------- We would be very happy to get more contributors. To get involved, start by grabbing the code: http://darcs.haskell.org/haddock.git Then take a look at the bug and feature tracker for things to work on: http://trac.haskell.org/haddock

2 4

ANN: 11th GhentFPG Meeting Tuesday, 20th of March: Final Program
by Jeroen Janssen 04 Mar '12

04 Mar '12

(apologies if you receive multiple postings) Dear all, The program and *new location* for the 11th GhentFPG meeting has been determined. The meeting will take place on Tuesday, the 20th of March at 19:30 in auditorium V1 of building S9 at Campus De Sterre (Krijgslaan 281, 9000 Gent). We will post signs at building S9 to guide you towards the auditorium. The program is as follows: 1. Jasper Van der Jeugt: digestive-functors 0.3 Formlets are a way of building HTML forms that are type-safe, handle errors, abstract and are easy to combine into bigger forms. I present the digestive-functors library, a formlets implementation which solves a number of issues with the original formlets library, such as annotation of errors, labelling, and separation of the validation model from the actual HTML output. 2. Jeroen Janssen: the typeclassopedia. This is part of a new type of talks we are going to try with GhentFPG. Each meeting we would like to introduce you to some of the important papers in the world of FP. This meeting I am going to introduce you to the brilliant "Typeclassopedia" paper. This paper, which originally appeared in the monad reader issue 13, introduces many of the important typeclasses in the Haskell standard library and explains how they are interrelated, both in a formal fashion as in a practical fashion. Hope to see you all there! The GhentFPG Organizing Committee.

1 0

Problems installing Data.Encoding package
by Pēteris Paikens 04 Mar '12

04 Mar '12

Hello, I'm stuck at trying to get Data.Encoding package functional. Cabal gives the following error message: cabal install encoding Resolving dependencies... /var/folders/xc/jnh69_cx1n32n1qwkgfrlc100000gq/T/encoding-0.6.620435/encoding-0.6.6/dist/setup/setup: /var/folders/xc/jnh69_cx1n32n1qwkgfrlc100000gq/T/encoding-0.6.620435/encoding-0.6.6/dist/setup/setup: cannot execute binary file cabal: Error: some packages failed to install: encoding-0.6.6 failed during the configure step. The exception was: ExitFailure 126 What could be done here? I'm on MacOS Lion, ghc v. 7.0.4 Regards, Peteris

3 2

anyone know what happened to nlpwp?
by serialhex 04 Mar '12

04 Mar '12

i was going through nlpwp.org learning haskell & natural language processing at the same time, and a few days ago it just disappeared! does anybody know what happened to it? i can still access the google cache, and learn from there, but that'd be a pain for anybody new coming along who can't access this site... :( hex -- * my blog is cooler than yours: http://serialhex.github.com * The wise man said: "Never argue with an idiot. They bring you down to their level and beat you with experience." * As a programmer, it is your job to put yourself out of business. What you do today can be automated tomorrow. ~Doug McIlroy No snowflake in an avalanche ever feels responsible. --- CFO: “What happens if we train people and they leave?” CTO: “What if we don’t and they stay?”

2 1