October 2008 - Haskell-Cafe

Restricted file reading monad
by oleg＠okmij.org 02 Oct '08

02 Oct '08

George Pollard wrote: > The structure of an ID3 tag goes something like this: > Header: > - total size of tag > - other header info > A series of frames, each with: > - total size of frame > - other header info > - frame data > > Since the ID3 tag as a whole has size information, I need to pass that > into the frame-reading functions to ensure that I never read past the > end of the total header. This exact same requirement is present within > the frame-reading function itself; it must read the frame data without > reading past the end of the frame. The structure of the chunk-encoded content is similar: first comes the size, then come data of that size. We have to read exactly that amount (and then check for the chunk trailer, CRLF). In a sense, we have a data stream embedded inside another stream. The Iteratee IO framework was specifically designed to process these sorts of arbitrarily nested and encoded data streams. In particular, the file http://okmij.org/ftp/Haskell/Iteratee/IterateeM.hs contains the function -- Read n elements from a stream and apply the given iteratee to the -- stream of the read elements. Unless the stream is terminated early, we -- read exactly n elements (even if the iteratee has accepted fewer). -- This procedure shows a different way of composing two iteratees: -- `vertical' rather than `horizontal' stake :: Monad m => Int -> IterateeG el m a -> IterateeGM el m (IterateeG el m a) The implementation is 11 lines long, counting the `where' line. I believe the code is declarative rather than imperative. The function stake (the analogue of List.take) is then used in enum_chunk_decoded :: Monad m => Iteratee m a -> IterateeM m a The end of the file IterateeM.hs has several complete tests.

1 0

Class Quantification
by oleg＠okmij.org 02 Oct '08

02 Oct '08

Bas van Dijk wrote: > ... it's possible to define 'foo' and 'bar' like so: > > > foo :: (Num c, Num d) => (forall b. Num b => a -> b) -> a -> (c, d) > > foo f x = (f x, f x) > > > bar :: (Read c, Read d) => (forall b. Read b => a -> b) -> a -> (c, d) > > bar f x = (f x, f x) > > Which allows us to write: > > > testFoo = foo fromInteger 1 :: (Int, Float) > > testBar = bar read "1" :: (Int, Float) > > Now I would like to generalise 'foo' and 'bar' to 'bla' so that I can write: > > testBla1 = bla fromInteger 1 :: (Int, Float) > testBla2 = bla read "1" :: (Int, Float) Quantification over classes is *easily* achievable in Haskell. Although functional dependencies are required, overlapping or undecidable instances are not. The technique is also far simpler than that in SYB3 and takes only a few lines to write. The complete code follows. One may note that Rank2 types are *not* needed. Therefore, we do not even have to give signature to the function bla as it can be inferred. That technique suggests that Rank2 types are already present in some form in Haskell (even Haskell98) already. In more detail, the technique is explained in Class-parameterized classes, and the type-level logarithm http://okmij.org/ftp/Haskell/types.html#peano-arithm Type-class overloaded functions: second-order typeclass programming with backtracking http://okmij.org/ftp/Haskell/types.html#poly2 Here is the code {-# LANGUAGE RankNTypes, MultiParamTypeClasses, FunctionalDependencies #-} {-# LANGUAGE FlexibleInstances, TypeSynonymInstances #-} -- the result type determines the argument type class RApply l a b | l b -> a where rapply :: l -> a -> b -- the expected instance instance RApply (a->b) a b where rapply = ($) data LRead = LRead instance Read b => RApply LRead String b where rapply _ = read data LFromInt = LFromInt instance Num b => RApply LFromInt Integer b where rapply _ = fromInteger bla x arg = (rapply x arg, rapply x arg) testBla1 = bla LFromInt 1 :: (Int, Float) testBla2 = bla LRead "1" :: (Int, Float)

1 0

Haskell Weekly News: Issue 87 - October 1, 2008
by Brent Yorgey 01 Oct '08

01 Oct '08

--------------------------------------------------------------------------- Haskell Weekly News http://sequence.complete.org/hwn/20081001 Issue 87 - October 01, 2008 --------------------------------------------------------------------------- Welcome to issue 87 of HWN, a newsletter covering developments in the [1]Haskell community. ICFP was held last week in Victoria, and by all accounts was a great success! This edition of the HWN includes much ICFP and Haskell Symposium-related content, including [2]videos of the Haskell symposium presentations, [3]programming contest results, some [4]notes on the future of Haskell, and slides from [5]a Haskell tutorial and [6]a talk about the Haskell Platform. But ICFP didn't seem to slow down the community all that much: you'll find the usual mix of newly released and updated packages, blog posts, mailing list discussions, and silly quotes as well. Announcements Haskell-Embedded System Design: ForSyDe 3.0 and Tutorial. Alfonso Acosta [7]announced the [8]3.0 release of [9]ForSyDe. The ForSyDe (Formal System Design) methodology has been developed with the objective to move system design (e.g. System on Chip, Hardware and Software systems) to a higher level of abstraction. ForSyDe is implemented as a Haskell-embedded behavioral DSL (Domain Specific Language). The 3.0 release includes a new deep-embedded DSL and embedded compiler, as well as a new user-friendly tutorial. Graphalyze-0.1. Ivan Miljenovic [10]announced the initial release of his graph-theoretic analysis library, [11]Graphalyze. This is a pre-release of the library he is writing for his mathematics honours thesis, "Graph-Theoretic Analysis of the Relationships in Discrete Data". Symposium videos. Malcolm Wallace [12]announced [13]guerrilla videos of the Haskell Symposium 2008 presentations. ICFP programming contest results. Malcolm Wallace [14]sent a link to [15]a video of the ICFP programming contest results presentation. Version 0.4.3 of happs-tutorial is a HAppS job board, done in HAppS.. Thomas Hartman [16]announced version 4 of the [17]self-demoing HAppS tutorial, a HAppS job board. TH code for deriving Binary and NFData instances. Tim Newsham [18]announced some [19]Template Haskell code for automatically deriving Data.Binary and Control.Parallel.Strategies.NFData instances. Notes on the future of Haskell from ICFP. Bryan O'Sullivan [20]posted a [21]writeup from the ICFP conference floor on the future of Haskell and functional programming. datapacker 1.0.1. John Goerzen [22]announced the [23]release of datapacker 1.0.1. A Functional Implementation of the Garsia-Wachs Algorithm. Nicolas Pouillard [24]announced a [25]Haskell implementation of an algorithm that builds a binary tree with minimum weighted path length from weighted leaf nodes given in symmetric order. This can be used to build optimum search tables, to balance a 'ropes' data structure in an optimal way. graphviz-2008.9.20. Ivan Miljenovic [26]announced a new version of [27]Matthew Sackman's Haskell bindings to Graphviz. See Ivan's original announcement for information on what new features are included, and what the difference is among the various graphviz-related packages on Hackage. darcs 2.1.0pre2. Eric Kow [28]announced the release of darcs 2.1.0pre2, formerly known as 2.0.3. See Eric's announcement for a list of new features and bug fixes in this release. protocol-buffers-0.2.9 for Haskell is ready. ChrisK [29]announced the release of the [30]protocol-buffers package, which generates Haskell data types that can be converted back and forth to lazy ByteStrings that interoperate with Google's generated code in C++/Java/python. panda blog engine. Jinjing Wang [31]announced the release of [32]panda, a simple blog engine written in Haskell. OpenSPARC project applicant chosen. Duncan Coutts [33]announced that Ben Lippmeier has been chosen for the [34]OpenSPARC project. Ben will spend three months hacking on GHC to make it perform well on the latest multi-core OpenSPARC chips. Hugs on the iPhone. Alberto Galdo [35]announced that he has gotten Hugs to run on the iPhone, and has made packages available for others who would like to install it as well. Discussion Shooting yourself in the foot in Haskell. John Van Enk [36]asked how to shoot yourself in the foot with Haskell, with humorous results. Total Functional Programming in Haskell. Jason Dagit started a [37]discussion on total functional programming, Haskell, abstraction boundaries and the IO monad, and related topics. Health effects. Andrew Coppin told a [38]story about a chocolate bar and recursion, which led to a discussion of optimization problems, Dedekind cuts, some meta-discussion of the discussion, and entirely too many puns. The container problem. Andrew Coppin [39]asked about the possibility if abstracting over various sorts of containers in Haskell, and why there isn't a widely used library that does this. A discussion of various container libraries and the language issues that arise followed. Red-Blue Stack. Matthew Eastman [40]asked how to implement a certain data structure (red-blue stacks) in Haskell. Several people responded with increasingly clever solutions, and a comparison of mutating vs. non-mutating algorithms. Climbing up the shootout.... Don Stewart began a long and ongoing [41]discussion about improving Haskell's performance on benchmarks in the [42]Shootout, now that there is a quad core machine for running benchmarks! Line noise. Andrew Coppin started an [43]interesting discussion about perceptions of Haskell syntax by programmers who aren't familiar with it. Jobs London FP job in asset management. Michael Bott [44]announced an opportunity for two functional programmers based in London, with a software house specialising in asset management. Blog noise [45]Haskell news from the [46]blogosphere. * Creighton Hogg: [47]Some first steps with Data.Reactive. Creighton gives some simple examples of using Conal Elliott's Reactive library. More to come! * Bryan O'Sullivan: [48]Unix hacking in Haskell: better pseudoterminal support. * Creighton Hogg: [49]One last thought on laziness. In Creighton's opinion, laziness is the single hardest thing to get used to in Haskell. If you're learning Haskell, don't despair, break out the pencil and paper! * Douglas M. Auclair (geophf): [50]Animal as RDR, part III. * Neil Mitchell: [51]General Updates. * Don Stewart (dons): [52]Newest Mersenne Prime. Haskell doesn't even break a sweat computing the largest known prime number. * Douglas M. Auclair (geophf): [53]Animal as RDR, part II. * Bryan O'Sullivan: [54]Using Bloom filters for large scale gene sequence analysis in Haskell. A paper that Bryan and Ketil Malde submitted to PADL 09. "The Cliff's Notes version: Bloom filters are almost unused in bioinformatics; they're tremendously useful; and our Haskell code is really fast. * >>> Zubin Wadia: [55]Simon Peyton Jones & Microsoft Research Cambridge. Zubin thinks quite highly of SPJ and MSR Cambridge. * Bryan O'Sullivan: [56]Slides from my DEFUN 2008 Haskell tutorial. * Mads Lindstroem: [57]Inheritance in Composites and Overlapping Instances. * >>> Micah Cowan: [58]Adventures in Haskell. Micah shares some thoughts on learning Haskell. * Bryan O'Sullivan: [59]Some notes on the future of Haskell and FP. * Well-Typed.Com: [60]Slides from the Haskell Platform talk. * Paul Johnson: [61]Why the banks collapsed, and how a paper on Haskell programming can help stop it happening next time. * >>> Nathan Sanders: [62]Two weeks of Haskell. Nathan shares some thoughts on his first two weeks learning Haskell. * Bryan O'Sullivan: [63]Twittering from ICFP / Haskell symposium / CUFP. * Real-World Haskell: [64]Slides from ACCU talk. * Eric Kow (kowey): [65]darcs weekly news #5. * John Goerzen (CosmicRay): [66]New version of datapacker. * >>> James Cowie: [67]Haskell, the verdict!. James is impressed with Haskell after using it for a few weeks. * >>> Alex Combas: [68]What's all this fuss about Haskell?. Alex is thinking of learning Haskell in his spare time. * Aaron Tomb: [69]Parsing the Linux kernel with Haskell: experience with Language.C. Aaron is impressed by the new Language.C libraries, which parses all 18 million pre-processed lines of Linux kernel source with no problems! Quotes of the Week * Fuse_: Oh, sorry for hijacking mathematical purity with dirty fiscal dynamical systems. :o * mauke: <mauke> data Mushroom badger = Mushroom badger badger badger badger badger badger badger badger badger <leimy> where's the snake <mauke> deriving Snake * ddarius: higher order of lambdabot deployment and management engineers or HOLDME * Botje: #haskell: parallellising your homework answers! * olsner: most everything gives nicer everything than perl * Botje: fuzzy feelings aren't always aerodynamic, unfortunately. * chrisdone: benchmarks only exist to make fun of ruby * Claus Reinke: [on breaking code up into smaller bits] Once your readers understand your code, you can add the one-liner and ask for applause. * Jake Mcarthur: A fold by any other name would smell as sweet. * lispy: Schroedinger's cat is really in a thunk not a box * Bulat: Haskell was developed with goal to hide implementation details from egg-headed scientists and this obviously should have some drawbacks About the Haskell Weekly News New editions are posted to [70]the Haskell mailing list as well as to [71]the Haskell Sequence and [72]Planet Haskell. [73]RSS is also available, and headlines appear on [74]haskell.org. To help create new editions of this newsletter, please see the information on [75]how to contribute. Send stories to byorgey at cis dot upenn dot edu. The darcs repository is available at darcs get [76]http://code.haskell.org/~byorgey/code/hwn/ . References 1. http://haskell.org/ 2. http://haskell.org/haskellwiki/Video_presentations/Haskell_Symposium_2008 3. http://video.google.com/videoplay?docid=-4697764813432201693 4. http://www.serpentine.com/blog/2008/09/26/some-notes-on-the-future-of-haske… 5. http://www.serpentine.com/blog/2008/09/27/slides-from-my-defun-2008-haskell… 6. http://blog.well-typed.com/2008/09/slides-from-the-haskell-platform-talk/ 7. http://article.gmane.org/gmane.comp.lang.haskell.general/16443 8. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/ForSyDe 9. http://www.ict.kth.se/org/ict/ecs/sam/projects/forsyde/www/ 10. http://article.gmane.org/gmane.comp.lang.haskell.general/16442 11. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/Graphalyze 12. http://article.gmane.org/gmane.comp.lang.haskell.general/16440 13. http://haskell.org/haskellwiki/Video_presentations/Haskell_Symposium_2008 14. http://article.gmane.org/gmane.comp.lang.haskell.general/16438 15. http://video.google.com/videoplay?docid=-4697764813432201693 16. http://article.gmane.org/gmane.comp.lang.haskell.cafe/45332 17. http://happstutorial.com:5001/ 18. http://article.gmane.org/gmane.comp.lang.haskell.cafe/45305 19. http://www.thenewsh.com/~newsham/store/DeriveBinary.hs 20. http://article.gmane.org/gmane.comp.lang.haskell.cafe/45272 21. http://www.serpentine.com/blog/2008/09/26/some-notes-on-the-future-of-haske… 22. http://article.gmane.org/gmane.comp.lang.haskell.cafe/45185 23. http://changelog.complete.org/posts/760-New-version-of-datapacker.html 24. http://article.gmane.org/gmane.comp.lang.haskell.cafe/45002 25. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/garsia-wachs 26. http://article.gmane.org/gmane.comp.lang.haskell.cafe/44863 27. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/graphviz 28. http://www.haskell.org//pipermail/haskell-cafe/2008-September/048094.html 29. http://article.gmane.org/gmane.comp.lang.haskell.cafe/44845 30. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/protocol-buffers 31. http://article.gmane.org/gmane.comp.lang.haskell.cafe/44840 32. http://jinjing.blog.easymic.com/ 33. http://article.gmane.org/gmane.comp.lang.haskell.cafe/44836 34. http://haskell.org/opensparc/ 35. http://www.haskell.org//pipermail/haskell-cafe/2008-September/047838.html 36. http://www.haskell.org//pipermail/haskell-cafe/2008-October/048506.html 37. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/45393 38. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/45370 39. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/45242 40. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/45122 41. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/44916 42. http://shootout.alioth.debian.org/u64q/benchmark.php?test=all&lang=all 43. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/44886 44. http://article.gmane.org/gmane.comp.lang.haskell.cafe/45506 45. http://planet.haskell.org/ 46. http://haskell.org/haskellwiki/Blog_articles 47. http://abstractabsurd.blogspot.com/2008/09/some-first-steps-with-datareacti… 48. http://www.serpentine.com/blog/2008/09/30/unix-hacking-in-haskell-better-ps… 49. http://abstractabsurd.blogspot.com/2008/09/one-last-thought-on-laziness.html 50. http://logicaltypes.blogspot.com/2008/09/animal-as-rdr-part-iii.html 51. http://neilmitchell.blogspot.com/2008/09/general-updates.html 52. http://www.cse.unsw.edu.au/~dons/blog/2008/09/30#primes 53. http://logicaltypes.blogspot.com/2008/09/animal-as-rdr-part-ii.html 54. http://www.serpentine.com/blog/2008/09/28/using-bloom-filters-for-large-sca… 55. http://zwadia.com/?p=58 56. http://www.serpentine.com/blog/2008/09/27/slides-from-my-defun-2008-haskell… 57. http://lindstroem.wordpress.com/2008/09/27/inheritance-in-composites-and-ov… 58. http://micah.cowan.name/2008/09/26/computers/software-development/adventure… 59. http://www.serpentine.com/blog/2008/09/26/some-notes-on-the-future-of-haske… 60. http://blog.well-typed.com/2008/09/slides-from-the-haskell-platform-talk/ 61. http://paulspontifications.blogspot.com/2008/09/why-banks-collapsed-and-how… 62. http://sandersn.com/blog/index.php?title=two_weeks_of_haskell 63. http://www.serpentine.com/blog/2008/09/25/twittering-from-icfp-haskell-symp… 64. http://www.realworldhaskell.org/blog/2008/09/25/slides-from-accu-talk/ 65. http://koweycode.blogspot.com/2008/09/darcs-weekly-news-5.html 66. http://changelog.complete.org/posts/760-New-version-of-datapacker.html 67. http://www.jcowie.co.uk/2008/09/haskell-the-verdict/ 68. http://combas3d.com/2008/09/whats-all-this-fuss-about-haskell/ 69. http://www.galois.com/blog/2008/09/17/parsing-the-linux-kernel-with-haskell… 70. http://www.haskell.org/mailman/listinfo/haskell 71. http://sequence.complete.org/ 72. http://planet.haskell.org/ 73. http://sequence.complete.org/node/feed 74. http://haskell.org/ 75. http://haskell.org/haskellwiki/HWN 76. http://code.haskell.org/~byorgey/code/hwn/

1 0

flipped IO sequence
by Cetin Sert 01 Oct '08

01 Oct '08

warn :: String → IO Int warn = return 1 << putStrLn -- causes an error -- = \msg → return 1 << putStrLn msg -- works just fine -- = \msg → putStrLn msg >> return 1 -- works just fine (<<) :: Monad m ⇒ m b → m a → m b b << a = a >>= \_ → b Why do I get this compile-time error?? How can one define << ? cetin@linux-d312:~/lab/test/qths/p> ghc -fglasgow-exts -O2 -o d64x --make demo2.hs system.hs [1 of 2] Compiling Netman.System ( system.hs, system.o ) system.hs:23:14: No instance for (Num (IO Int)) arising from the literal `1' at system.hs:23:14 Possible fix: add an instance declaration for (Num (IO Int)) In the first argument of `return', namely `1' In the first argument of `(<<)', namely `return 1' In the expression: return 1 << putStrLn

5 4

(A little humour)
by Andrew Coppin 01 Oct '08

01 Oct '08

Apparently C++ lets you overload the arithmetic operators, the assignment operator, the initialisation and destruction operators, the pointer dereference operator, the memory allocation operator, and even the function call operator. o_O But only _Haskell_ lets you overload the ";" operator! ;-) [I believe they call this "ha ha, only serious" humour...]

9 14

cabal upgrade
by Achim Schneider 01 Oct '08

01 Oct '08

% cabal update Downloading package list from server 'http://hackage.haskell.org/packages/archive' % cabal install pureMD5 All requested packages already installed. Nothing to do. % cabal install pureMD5-0.2.4 Downloading 'pureMD5-0.2.4'... [...] Well, I might be spoiled by portage but shouldn't there be a thing like cabal upgrade pureMD5 and cabal upgrade --all ? -- (c) this sig last receiving data processing entity. Inspect headers for copyright history. All rights reserved. Copying, hiring, renting, performance and/or broadcasting of this signature prohibited.

12 22

The Haskell Platform
by Galchin, Vasili 01 Oct '08

01 Oct '08

Hello, I probably missed some details for which I apologize. My feeling is that "periodically" the "haskell platform server" should attempt to rebuild the Haskell library. Any library that fails to rebuild then the "maintainer" of that library should be notified, e.g. email, pager(;^)), .... Kind regards, Vasili

2 1

Haskell, GHC, and ABI
by Sean Leather 01 Oct '08

01 Oct '08

> Dynamic linking doesn't solve all the problems, we still have the problem > that GHC does a lot of cross-module inlining, regardless of whether dynamic > linking is used. However, I really would like to have a way to have > complete control over what is exposed across a package boundary. We need > this not just for licensing reasons, but also for making a dynamic library > with a fixed ABI, so it can be upgraded later. > > Incedentally the lack of this feature is one reason I've not being rushing > to get shared libraries into GHC. They're just not that useful unless you > can upgrade a library independently of the things it depends on. > I completely agree. Having a fixed ABI that survives library changes and compiler upgrades is an important factor in achieving a useful system. Incidentally, I perused http://hackage.haskell.org/trac/ghc/wiki/DynamicLinking and didn't see much about this here. Is there some notes or a plan for reaching this goal with GHC (or any other Haskell compiler)? Sean

1 0

Re: [Haskell-cafe] Version 0.4.3 of happs-tutorial is a HAppS job board, done in HAppS.
by Thomas Hartman 01 Oct '08

01 Oct '08

Oh, the whole point of this was to put preview of the latest chapters I've been working on. The new stuff is at http://happstutorial.com:5002/ So far there's an introduction to macid, macid data safety and backups, and using utf8 data with happs. I'm particularly intereted in hearing what the happs devs and others have to say on the topic of keeping data safe: http://happstutorial.com:5002/tutorial/macid-data-safety Thomas. 2008/9/30 John Melesky <john(a)phaedrusdeinus.org>: > Thomas, > > The tutorial site seems to be unreachable or otherwise down. > > I know the modern programming landscape is fast-moving, HAppS is new and > constantly changing, so i probably should have grabbed the content when you > posted it, but perhaps keeping it up for more than a day and a half might be > good? :-) > > -john melesky > > > On Sep 28, 2008, at 6:36 PM, Thomas Hartman wrote: > >> Hello, world. >> >> In Version 4 of the ongoing self-demoing HAppS Tutorial, we implement >> a HAppS job board using HAppS. >> >> demo: http://happstutorial.com:5001 >> >> install: sudo cabal install happs-tutorial >> >> darcs head: darcs get http://code.haskell.org/happs-tutorial >> >> There aren't any new lessons compared to the last release, but the >> code is much cleaner, and... well... the web site actually does >> something. >> >> Coming soon, lessons on form processing and HAppS State. >> >> Enjoy! >> >> Thomas. >> _______________________________________________ >> Haskell-Cafe mailing list >> Haskell-Cafe(a)haskell.org >> http://www.haskell.org/mailman/listinfo/haskell-cafe >> > >

1 0

Re: [Haskell-cafe] Class Quantification
by Bas van Dijk 01 Oct '08

01 Oct '08

On Wed, Oct 1, 2008 at 3:01 AM, Reiner Pope <reiner.pope(a)gmail.com> wrote: > I believe there is no way to simply express this "abstraction over classes", > but the Scrap your boilerplate with class[1] paper discusses this same > problem and present a workaround by defining the class's dictionary of > methods as an explicit type. > > What follows is the code to implement their workaround for your example. > > First some fairly standard extensions: > >> {-# LANGUAGE Rank2Types, EmptyDataDecls, FlexibleInstances, KindSignatures >> #-} > > And some more controversial, but necessary ones: > >> {-# LANGUAGE UndecidableInstances, OverlappingInstances #-} >> >> module Cls where >> > > The working bla function. Unfortunately, the "pseudoclass" cls needs > to have its type explicitly named, hence the (uninhabited) Proxy type. > >> bla :: forall cls a c d. (Sat (cls c), Sat (cls d)) => Proxy cls -> >> (forall b. Sat (cls b) => a -> b) -> a -> (c,d) >> bla _ f x = (f x, f x) > > Again, testFoo and testBar unfortunately have to name which dictionary type > to use, via the Proxy. > >> testFoo = bla (undefined :: Proxy NumD) fromInteger 1 :: (Int,Float) >> testBar = bla (undefined :: Proxy ReadD) read "1" :: (Int,Float) > > The Sat class, straight from SYB: > >> class Sat a where >> dict :: a >> data Proxy (cxt :: * -> *) > > The explicit dictionary construction for the Read class: > >> data ReadD a = ReadD { readsPrecD :: Int -> ReadS a } >> instance (Read a) => Sat (ReadD a) where >> dict = ReadD readsPrec >> instance (Sat (ReadD a)) => Read a where >> readsPrec = readsPrecD dict > > The explicit dictionary construction for the Num class: > >> data NumD a = NumD { plusD :: a -> a -> a, >> timesD :: a -> a -> a, >> negateD :: a -> a, >> absD :: a -> a, >> signumD :: a -> a, >> fromIntegerD :: Integer -> a } >> instance (Num a) => Sat (NumD a) where >> dict = NumD (+) (*) negate abs signum fromInteger > > We define these fake Eq,Show instances just to make the > Num instance valid. It would be longer, but not more > difficult, to genuinely encode the (Eq,Show)=>Num hierarchy. > >> instance (Sat (NumD a)) => Show a where {} >> instance (Sat (NumD a)) => Eq a where {} >> instance (Sat (NumD a)) => Num a where >> (+) = plusD dict >> (*) = timesD dict >> negate = negateD dict >> abs = absD dict >> signum = signumD dict >> fromInteger = fromIntegerD dict > > [1] http://homepages.cwi.nl/~ralf/syb3/ Sections 3.2 and 4.1 > > On Wed, Oct 1, 2008 at 9:01 AM, Bas van Dijk <v.dijk.bas(a)gmail.com> wrote: >> >> On Tue, Sep 30, 2008 at 11:25 PM, Sean Leather <sean.leather(a)gmail.com> >> wrote: >> > But perhaps you're looking for potentially unknown classes? >> >> Yes indeed. >> >> Thanks, >> >> Bas >> _______________________________________________ >> Haskell-Cafe mailing list >> Haskell-Cafe(a)haskell.org >> http://www.haskell.org/mailman/listinfo/haskell-cafe > > Nice! The explicit passing of the proxy type is indeed unfortunate but it's nice that the type of your 'bla' is almost identical to mine: > bla :: forall cls a c d. (Sat (cls c), Sat (cls d)) => Proxy cls -> (forall b. Sat (cls b) => a -> b) -> a -> (c,d) > bla :: forall cls. (cls c, cls d) => (forall b. cls b => a -> b) -> a -> (c, d) Thanks, Bas

2 1