- Haskell-Cafe - Haskell.org

Re: [Haskell-cafe] implementing python-style dictionary in Haskell
by Tillmann Rendel 18 Nov '08

18 Nov '08

Hello Alberto, I cc this to haskell-cafe again. Alberto G. Corona wrote: > Not so much memory, because data is referentially transparent, the new Map > can point to whole subtrees of the old map that stay the same. is similar > than when a new list is created by prefixing a new element from a old list > ys= x:xs. ys is not at all a fresh copy, but x plus a pointer to the head > of xs. this is the only new data that is needed to create ys. You could just as well compare with appending a new element to the end of the list, which needs a complete copy of the list structure to be made. One has to look more closely to see which case it is here. More specifically, I do not see how this sharing of substructures could be employed in the implementation of hash tables, which rely on O(1) random access into arrays. Tillmann

2 1

Re: Statically dimension-checked hmatrix
by Patrick Perry 18 Nov '08

18 Nov '08

> What is the situation regarding statically dimension-checked linear > algebra libraries? It seems that Frederik Eaton was working on one in > 2006 (see the paper "Statically typed linear algebra in Haskell"), and > he produced the Vectro library from this, based on GSLHaskell. > > Are there any more recent efforts into this, particularly using the > new TFs? If not, I might have a go at it, as a thin wrapper for > hmatrix. The BLAS bindings I wrote use phantom types to make sure the dimensions are consistent. See, for example the functions to get row and column views of a matrix: http://hackage.haskell.org/packages/archive/blas/0.6/doc/html/Data-Matrix-D… Also, the multiplication routines: http://hackage.haskell.org/packages/archive/blas/0.6/doc/html/BLAS-Matrix-I… I've found that phantom types are a good trade off: they provide enough type safety to catch common mistakes without being too much of a hassle to use. Another benefit (which I didn't anticipate) is that they help a *lot* with documentation. I'm skeptical of value that comes from stronger typing. At some point the types become too much of a hassle to be worth using. Even phantom types get in the way sometimes and require use of either GADTs, unsafeCoerce, or both, which is annoying. I'd be interested to hear about any developments you make in this area. Patrick p.s. If you're interested in trying it out, the version of blas on hackage doesn't compile with ghc-6.10.1, but the version in the darcs repository at http://stat.stanford.edu/~patperry/code/blas does.

1 0

simple generic / data view library?
by Conal Elliott 18 Nov '08

18 Nov '08

Is there a simple existing library that provides views of data types in terms of unit, product and sum? Here's what I threw together for my own use. I used associated types, though functional dependencies would work as well. class HasView t where type View t view :: t -> View t unview :: View t -> t -- View instances instance HasView (Maybe a) where type View (Maybe a) = Either () a view Nothing = (Left ()) view (Just a) = (Right a) unview (Left ()) = Nothing unview (Right a) = (Just a) instance HasView [a] where type View [a] = Either () (a,[a]) view [] = (Left ()) view (a:as) = (Right (a,as)) unview (Left ()) = [] unview (Right (a,as)) = (a:as) onView2 :: (HasView a, HasView b, HasView c) => (View a -> View b -> View c) -> (a -> b -> c) onView2 op a b = unview (view a `op` view b) Thanks, - Conal

2 1

Statically dimension-checked hmatrix
by Reiner Pope 18 Nov '08

18 Nov '08

Hi, What is the situation regarding statically dimension-checked linear algebra libraries? It seems that Frederik Eaton was working on one in 2006 (see the paper "Statically typed linear algebra in Haskell"), and he produced the Vectro library from this, based on GSLHaskell. Are there any more recent efforts into this, particularly using the new TFs? If not, I might have a go at it, as a thin wrapper for hmatrix. Cheers, Reiner

1 0

Find unused exports
by Jason Dagit 18 Nov '08

18 Nov '08

Hello, Has anyone already made a tool to check if exported functions, data constructors, types, etc are unused within a set of modules? For my usage it would probably suffice if the tool only compared import lists against export lists as we compile with -Wall and 99% of our imports say exactly which names are imported (and that last 1% could easily be fixed to say as well). If the tool doesn't already exist is there an easy way to approximate it? Thanks, Jason

5 7

How to make interface IO() to IO a in a command-driven tool
by Kyongho Min 17 Nov '08

17 Nov '08

Dear Haskell users, I am working on a command-driven tool (::IO ()). If the system invokes itself again, and the system's returned data type would be 'IO a', say (IO Exp) (Expression). tool:: a -> IO() The nested tool invocation returns (IO Exp) to the previous tool. I am using dynamic error exception handling like (throwDyn exp). If I use 'forkIO', then it works for this situation (e.g. IO Exp) but it is not working for 'quit' command (quit the tool:: thread blocked indefinitely) - quit the subtool and return to the previous tool. Any suggestion would be helpful? tool = {tool1, tool2, tool3, ...} tool1:: (IO ()) invoked tool2:: (IO a) may invoked tool3:: (IO a) tool2 is working basically on IO() and if I got the Exp (proved), then return it to the previous tool (say tool1). Thanks in advance, min

1 0

Haskell Weekly News: Issue 93 - November 15, 2008
by Brent Yorgey 17 Nov '08

17 Nov '08

--------------------------------------------------------------------------- Haskell Weekly News http://sequence.complete.org/hwn/20081115 Issue 93 - November 15, 2008 --------------------------------------------------------------------------- Welcome to issue 93 of HWN, a newsletter covering developments in the [1]Haskell community. Community News Congratulations to Ganesh (aka Heffalump) and Amanda on the [2]birth of Alexander Suresh Kerr Sittampalam! Announcements bustle-0.1. Will Thompson [3]announced the [4]release of [5]Bustle, a tool to show diagrams of D-Bus traffic for profiling purposes. It consists of a small C executable to log traffic, and a Gtk+ application which draws diagrams using Cairo. haskell-src-exts 0.4.1. Niklas Broberg [6]announced a new major release of the [7]haskell-src-exts package, an extension of the standard haskell-src package which handles most common syntactic extensions to Haskell. The new release features a cleaned up AST and names without ugly disambiguation prefixes. darcs 2.1.1rc2. Eric Kow (kowey) [8]announced the release of [9]darcs 2.1.1rc2, which adds support for GHC 6.10.1. It also includes a Windows bug fix. If you're using GHC 6.10.1 or Windows, give it a try and let the darcs development team know how it works. hpapi 0.1 release. Michael D. Adams [10]announced the first release of [11]hpapi, Performance API (PAPI) bindings for Haskell. PAPI provides access to various CPU counters such as cache-miss, instruction and pipeline stall counts. Workflow-0.1. Alberto G. Corona [12]announced the release of [13]Workflow, a library for transparent execution of computations across shutdowns and restarts. Reactive library (FRP) and mailing list. Conal Elliott [14]announced the release of [15]Reactive, a library for functional reactive programming (FRP), similar to the original [16]Fran but with a more modern interface (using standard type classes) and a hybrid push/pull implementation. It is designed to be used in a variety of contexts, such as interactive 2D and 3D graphics, graphical user interfaces, web services, and automatic recompilation/re-execution. There is also now a [17]mailing list and a [18]feature/bug tracker. ANN (sorta): OpenGL with extra type safety. Neal Alexander [19]announced a modification of the hOpenGL (and GLFW) source tree to force extra type checking on its various IO actions using the -XGeneralizedNewtypeDeriving extension. The main motivation was for writing concurrent OpenGL applications; the second motivation was to enforce static type checking on commands that can only be executed in certain OpenGL contexts (sending vertex data for example). Hopefully the code will be uploaded to Hackage as a separate package soon. FieldTrip library (functional 3D) and mailing list. Conal Elliott [20]announced the release of [21]FieldTrip, a library for functional 3D graphics. It is intended for building static, animated, and interactive 3D geometry, efficient enough for real-time synthesis and display. FieldTrip also has a [22]mailing list and a [23]feature/bug tracker. gitit 0.2 release - wiki using HAppS, git, pandoc. John MacFarlane [24]announced the upload of an early version of [25]gitit, a Haskell wiki program, to HackageDB. Gitit uses HAppS as a webserver, git for file storage, pandoc for rendering the (markdown) pages, and highlighting-kate for highlighted source code. You can [26]try it out here. Comments and patches welcome. Discussion Proof that Haskell is RT. Andrew Birkett [27]asked whether there exists a formal proof that the Haskell language is referentially transparent. Such a thing cannot exist, since Haskell has no formally defined semantics, but an interesting discussion about referential transparency and semantics ensued anyway. What *not* to use Haskell for. Dave Tapley [28]asked how people answer the question, "what does Haskell not do well?" Unfortunately, it seems that there is no good answer to this question and the thread degenerated into a discussion of all the great things you can do with Haskell. If only Haskell sucked more. Blog noise [29]Haskell news from the [30]blogosphere. * Braden Shepherdson: [31]Pimp Your XMonad #2: SmartBorders. The second in Braden's series on customising xmonad explains how to use the NoBorders extension to get rid of window borders when you don't want them. * Real-World Haskell: [32]Real World Haskell Electronic Edition Now On Sale. * Conal Elliott: [33]More beautiful fold zipping. * >>> Stephen: [34]Endless Cavern. A procedurally-generated tessellated-cavern flying game. * Twan van Laarhoven: [35]Arrays without bounds. Unbounded arrays (with O(1) amortized access) in Haskell! * "The GHC Team": [36]Comparing concurrent linked-list implementations in Haskell. * Ketil Malde: [37]454 sequencing and parsing the SFF binary format. * Tom Schrijvers: [38]Type Invariants for Haskell. Tom, Louis-Julien Guillemette, and Stefan Monnier's paper on Type Invariants for Haskell, which was recently accepted at PLPV'09. * Conal Elliott: [39]Another lovely example of type class morphisms. Conal elaborates on Max Rabkin's recent post on composable folds. * Real-World Haskell: [40]Beautiful Parallelism: Harnessing Multicores with Haskell. Don Stewart will be talking about programming mainstream multicore systems with Haskell at SC'08 next week in Austin, Texas. * Luke Palmer: [41]Sketch of Udon (Version Control/Packaging system). Luke sketches a high-level design for a distributed storage system that could be used as the basis for solving a number of interesting problems. * Galois, Inc: [42]Galois awarded NASA research contract. * Mark Jason Dominus: [43]Representing ordinal numbers in the computer and elsewhere. * The GHC Team: [44]Bootstrapping cabal-install. * Jason Dagit: [45]Phantom Types, Existentials and Controlling Unification -- Part 1. * Ben Moseley: [46]Why does Functional Programming matter?. Quotes of the Week * BMeph: In a functional world, students would ask how that index shadowing works in those funny 'for' statements... * digit: i'm almost annoyed at how brilliant xmonad is. * _pizza_: i think Haskell is undoubtedly the world's best programming language for discovering the first few dozen numbers in the Fibonacci sequence over IRC. * adu: let uncat3 [] = [] ; uncat3 xs = (let (ys, zs) = splitAt 3 xs in ys : uncat3 zs) ; getFrom x y = map (x !!) $ map (fromIntegral . ((\x -> fromIntegral $ foldl (.|.) (0::Word8) (zipWith (\c n -> if c then bit n else (0::Word8)) x [0..2])) :: [Bool] -> Int)) $ reverse . uncat3 . reverse . concat . map (((\x -> map (testBit x) [7,6..0]) :: Word8 -> [Bool]) . fromIntegral . ord) $ y in getFrom " HWdelor" "e\184-\235" * Beelsebob: ((:[]) "pigs eat") <^(++)^> ((:[]) " robot monkies") About the Haskell Weekly News New editions are posted to [47]the Haskell mailing list as well as to [48]the Haskell Sequence and [49]Planet Haskell. [50]RSS is also available, and headlines appear on [51]haskell.org. To help create new editions of this newsletter, please see the information on [52]how to contribute. Send stories to byorgey at cis dot upenn dot edu. The darcs repository is available at darcs get [53]http://code.haskell.org/~byorgey/code/hwn/ . References 1. http://haskell.org/ 2. http://urchin.earth.li/aegean/baby/ 3. http://www.haskell.org//pipermail/haskell-cafe/2008-November/050626.html 4. http://people.collabora.co.uk/~wjt/bustle/releases/bustle-0.1.tar.gz 5. http://resiak.livejournal.com/59590.html 6. http://www.haskell.org//pipermail/haskell-cafe/2008-November/050634.html 7. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/haskell-src-exts… 8. http://www.haskell.org//pipermail/haskell-cafe/2008-November/050378.html 9. http://darcs.net/darcs-2.1.1rc2.tar.gz 10. http://article.gmane.org/gmane.comp.lang.haskell.general/16612 11. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/hpapi 12. http://article.gmane.org/gmane.comp.lang.haskell.general/16607 13. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/Workflow 14. http://article.gmane.org/gmane.comp.lang.haskell.general/16604 15. http://haskell.org/haskellwiki/Reactive 16. http://conal.net/Fran 17. http://www.haskell.org/mailman/listinfo/reactive 18. http://trac.haskell.org/reactive 19. http://article.gmane.org/gmane.comp.lang.haskell.cafe/47458 20. http://article.gmane.org/gmane.comp.lang.haskell.cafe/47416 21. http://haskell.org/haskellwiki/FieldTrip 22. http://www.haskell.org/mailman/listinfo/FieldTrip 23. http://trac.haskell.org/FieldTrip 24. http://article.gmane.org/gmane.comp.lang.haskell.cafe/47384 25. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/gitit 26. http://johnmacfarlane.net:5001/ 27. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/47569 28. http://thread.gmane.org/gmane.comp.lang.haskell.cafe/47491 29. http://planet.haskell.org/ 30. http://haskell.org/haskellwiki/Blog_articles 31. http://braincrater.wordpress.com/2008/11/15/pimp-your-xmonad-2-smartborders/ 32. http://www.realworldhaskell.org/blog/2008/11/15/real-world-haskell-electron… 33. http://feeds.feedburner.com/~r/conal/~3/453765396/ 34. http://www.maths.tcd.ie/~icecube/2008/11/endless-cavern/ 35. http://twan.home.fmf.nl/blog/haskell/UnboundedArray.details 36. http://ghcmutterings.wordpress.com/2008/11/14/comparing-concurrent-linked-l… 37. http://blog.malde.org/index.php/2008/11/14/454-sequencing-and-parsing-the-s… 38. http://tomschrijvers.blogspot.com/2008/11/type-invariants-for-haskell.html 39. http://feeds.feedburner.com/~r/conal/~3/452607365/ 40. http://www.realworldhaskell.org/blog/2008/11/12/beautiful-parallelism-harne… 41. http://lukepalmer.wordpress.com/2008/11/12/sketch-of-udon-version-controlpa… 42. http://www.galois.com/blog/2008/11/11/galois-awarded-nasa-research-contract/ 43. http://blog.plover.com/math/ordinals.html 44. http://ghcmutterings.wordpress.com/2008/11/10/bootstrapping-cabal-install/ 45. http://blog.codersbase.com/2008/11/10/phantom-types-existentials-and-contro… 46. http://nattermorphisms.blogspot.com/2008/11/why-does-functional-programming… 47. http://www.haskell.org/mailman/listinfo/haskell 48. http://sequence.complete.org/ 49. http://planet.haskell.org/ 50. http://sequence.complete.org/node/feed 51. http://haskell.org/ 52. http://haskell.org/haskellwiki/HWN 53. http://code.haskell.org/~byorgey/code/hwn/

3 2

HAPPS on a major hosting provider?
by Jefferson Heard 16 Nov '08

16 Nov '08

I was wondering if anyone's ever tried to run Haaps on a major hosting provider, like oh, say Site5? I have an app I'd otherwise use Rails for and I thought I'd give Haaps a try... -- Jeff I try to take things like a crow; war and chaos don't always ruin a picnic, they just mean you have to be careful what you swallow. -- Jessica Edwards

2 1

Good name for this operator?
by Ariel J. Birnbaum 15 Nov '08

15 Nov '08

Hi all! When working with Applicative, I often find myself defining and using this operator: (<%>) :: (Applicative f) => f (a -> b) -> a -> f b f <%> x = f <*> pure x and always keep wondering whether it is already known by another name. Can't find it in Control.Applicative. Hoogle didn't give me much either. NB: In general, there's a weird obstacle I notice when writing Haskell: the unshakable feeling that whatever I write, there's already a better name for it and a more elegant implementation lying around. It can get downright paralyzing at times *shudder* Have Fun, -- Ariel J. Birnbaum

3 3

ANNOUNCE: haskell-src-exts 0.4.1
by Niklas Broberg 15 Nov '08

15 Nov '08

Fellow Haskelleers, it is my pleasure to announce the new release of the haskell-src-exts package, version 0.4.1: http://hackage.haskell.org/cgi-bin/hackage-scripts/package/haskell-src-exts… darcs get http://code.haskell.org/HSP/haskell-src-exts This is a new "major" release, the first of the 0.4 branch, and as such it is not backwards compatible with the releases in the 0.3 branch. This is due to two major changes to the design of the library: 1) The AST has been cleaned up, by splitting Haskell expressions in the AST from the "expressions" used while parsing to represent both expressions and patterns. This means that the datatype for expressions will no longer contain constructors for things like wildcards and irrefutable patterns. It now only contains constructors for actual Haskell expressions. If you have been using haskell-src-exts previously, this change should not bite you at all unless you have abused these ugly hacks in the AST, so it is *almost* backwards compatible. However... 2) ... I've finally decided to take the plunge and get rid of the ugly prefixes on all datatypes in the AST. I am of the firm conviction that disambiguation between datatypes and functions with the same name should be handled through the module hierarchy and qualified imports, not by adding prefixes. And the longer I wait, the harder the switch. So now the datatypes that make up the AST are called things like Exp, Pat, Module (the former Module is now ModuleName), Type, QOp etc instead of HsExp, HsPat... . The migration from 0.3.x to 0.4.1 should be really simple: i) replace all occurences of Module with ModuleName ii) remove all occurences of Hs This release also contains a number of smaller, fully backwards compatible bug fixes. These are available for the 0.3 branch in the 0.3.12 release that will mark the end of that branch: http://hackage.haskell.org/cgi-bin/hackage-scripts/package/haskell-src-exts… No further bug fixes or features will be implemented for it, and I urge you to migrate as early as possible, it really isn't hard. :-) Cheers and Happy Haskelling, /Niklas

3 4