August 2008 - Haskell-Cafe

FRP question
by Tim Newsham 21 Aug '08

21 Aug '08

I'm interested in FRP, have read several of the papers and am wondering if it could be applied to writing multi-client server programs. What would be the input type events for such a system? What would be the output type events? How would the system handle the fact that it has to multiplex several IO streams (sockets)? I'm trying to answer these questions myself and having a hard time. Should input events include new connection requests and shutdowns? Individual bytes? Streams of bytes? or higher level PDUs? What kind of output should be generated? Full PDUs? streams of bytes? Pairs of (connection identifier,data)? How would such a system effectively hide the multiplexed IO going on? Is this sort of problem poorly suited for FRP? Tim Newsham http://www.thenewsh.com/~newsham/

5 6

Re: [Haskell] Compiler Construction course using Haskell?
by Johannes Waldmann 21 Aug '08

21 Aug '08

Martin Erwig wrote (on a similar course he taught): > * Many students did not know Haskell before, so I had to teach some > Haskell and was forced to avoid some more advanced features, such as > type classes and monads. Yeah, in my case I know they don't know Haskell... Still, I trust them to pick it up quickly, and understand type classes (by analogy to interfaces in Java) and - they have to learn monads. These are master students, they should be open to a challenge ... NB: The course is declared "optional". As a prerequisite, there is a (mandatory) course in Principles of Programming Languages. J.W.

1 0

Re: [Haskell] Compiler Construction course using Haskell?
by Johannes Waldmann 21 Aug '08

21 Aug '08

> This has me worried slightly: do you intend to showcase > monads, or do you intend to teach compiler construction? The subject is Compiler Construction ... Doing tree traversals via monads would, in OO-Speak, just correspond to using some Visitor pattern for the AST. The students know about these patterns. J.W.

1 0

Haskell Weekly News: Issue 82 - August 20, 2008
by Brent Yorgey 21 Aug '08

21 Aug '08

--------------------------------------------------------------------------- Haskell Weekly News http://sequence.complete.org/hwn/20080820 Issue 82 - August 20, 2008 --------------------------------------------------------------------------- Welcome to issue 82 of HWN, a newsletter covering developments in the [1]Haskell community. Announcements Lava2000 on Hackage. Emil Axelsson [2]announced that [3]Lava2000 has been uploaded to Hackage. Lava is a structural hardware description library embedded in Haskell. This version of Lava focuses on verification, and connects to a number of different verification engines (although only Smv and Satzoo are maintained in this version). QuickCheck and HPC Tutorial. Andy Gill [4]announced a half day tutorial on QuickCheck and HPC to be held at [5]DEFUN, taught by Andy Gill and Koen Claessen. The latest version of QuickCheck will also be released in conjunction with the tutorial. Compiler Construction course using Haskell?. Johannes Waldmann [6]announced that he plans to give a course in compiler construction, using Haskell as the implementation language. Any comments are appreciated, especially from anyone who has given or taken such a course. witness 0.1, open-witness 0.1. Ashley Yakeley [7]announced the release of these new packages and the draft paper on which they are based. posix-realtime 0.0.1. Vasili Galchin [8]announced the release of the posix-realtime package. benchpress 0.2.1. Johan Tibell [9]announced the first public release of [10]benchpress, a micro-benchmark library that produces statistics such as min, mean, standard deviation, median, and max execution time. wavconvert 0.1.1. Tim Chevalier [11]announced the release of [12]wavconvert, a tool which takes a directory tree as an argument, and converts any .wav files in it to .ogg (using an external OGG encoder) while filling in the ID3 tags based on the directory names. darcs hacking sprint (25-26 October). Eric Y. Kow [13]announced a darcs hacking sprint, to be held on the 25-26 of October, in Cambridge and Portland (and over IRC). The perfect opportunity to get involved, if you are interested in doing some hacking on darcs. Let Eric know if you are interested. X Haskell Bindings. Antoine Latter [14]announced that he is slowly [15]porting XCB (the X C Bindings) to Haskell, and would like input from any interested parties. logfloat. wren ng thornton [16]announced the release of the logfloat package for manipulating log-domain floating numbers. The main reason for casting numbers into the log-domain is to prevent underflow when multiplying many small probabilities as is done in Hidden Markov Models and other statistical models often used for natural language processing. The log-domain also helps prevent overflow when multiplying many large numbers. Roguestar 0.2.2. Christopher Lane Hinson [17]announced the release of version 0.2.2 of Roguestar. Google Summer of Code Progress updates from participants in the 2008 [18]Google Summer of Code. Hoogle 4. Neil Mitchell (ndm) is working on [19]Hoogle 4. [20]This week, he worked on getting out the last kinks to make [21]Hoogle 4 usable as a replacement for [22]Hoogle 3. Although this is the last official week of the GSoC, he plans to continue working on it when he has time. DPH physics engine. Roman Cheplyaka (Feuerbach) is working on a [23]physics engine using [24]Data Parallel Haskell. GHC plugins. Max Bolingbroke is working on dynamically loaded plugins for GHC. Cabal dependency framework. Andrea Vezzosi (Saizan) is working on a [25]make-like dependency analysis framework for Cabal. Language.C. Benedikt Huber (visq) is [26]working on Language.C, a standalone parser/pretty printer library for C99. Generic tries. Jamie Brandon is working on a library for efficient maps using generalized tries. GHC API. Thomas Schilling (nominolo) is working on [27]improvements to the GHC API. Jobs Platform Architect at Peerium, Inc.. Don Stewart [28]forwarded a job opportunity for a Haskell programmer at [29]Peerium, Inc. Peerium is building a new software platform for direct, real-time communication and collaboration within graphically rich environments, and is looking for a platform architect proficient in Haskell to work with the founders to design and implement components of the runtime and supporting libraries for our software platform. Resumes should be forwarded to hr [30]peerium.com. Blog noise [31]Haskell news from the [32]blogosphere. * Don Stewart: [33]The 500 Packages: Haskell, Distros, and Maintainership. * Neil Mitchell: [34]Hoogle New Features. Neil has finished his Google Summer of Code work on the [35]new version of Hoogle, but still intends to continue working on it when he gets the chance. A number of new features have been added. * Well-Typed.Com: [36]Solving the diamond dependency problem. Duncan explains how the newest version of Cabal (mostly) solves the DDDP (dreaded diamond dependency problem). * Real-World Haskell: [37]Two new chapters, preorder links, and a page count estimate. Two new beta chapters are up! * Dan Piponi (sigfpe): [38]Untangling with Continued Fractions: Part 1. Dan continues his fascinating series on rational tangles, showing how to describe a tangle using do-notation. * Christopher Lane Hinson: [39]ANN: Roguestar 0.2.2. * Luis Araujo: [40]Himerge (and Haskell) out to the street!. * Neil Mitchell: [41]GSoC Hoogle: Week 12. * Twan van Laarhoven: [42]A generic merge function. * Luke Palmer: [43]Slicing open the belly of the IO monad in an alternate universe. Luke explores methods of safely performing some I/O without the IO monad in the context of FRP. * Ketil Malde: [44]The wee beginnings of a biohaskell tutorial? -- and some thoughts on programming productivity. * Douglas M. Auclair (geophf): [45]Using Difference Lists. * Real-World Haskell: [46]Mid-August status update. Quotes of the Week * SimonM: People don't seem to believe me when I say this. In retrospect we shouldn't have called it forkOS, we should have called it forkReallyExpensiveOnlyNecessaryForCallingOpenGL_IO. * quicksilver: unsafeInterleaveIO gives you a way to cheat if you're happy to descend into a mire of programs with no semantics. * dmhouse: Eurgh, metastereo quotes. * quicksilver: [on @yhjulwwiefzojcbxybbruweejw] it's the noise a haskell developer makes if you kick him in the de bruijn index. 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://article.gmane.org/gmane.comp.lang.haskell.general/16381 3. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/chalmers-lava2000 4. http://article.gmane.org/gmane.comp.lang.haskell.general/16380 5. http://www.deinprogramm.de/defun-2008/ 6. http://article.gmane.org/gmane.comp.lang.haskell.general/16379 7. http://article.gmane.org/gmane.comp.lang.haskell.general/16376 8. http://article.gmane.org/gmane.comp.lang.haskell.cafe/43295 9. http://article.gmane.org/gmane.comp.lang.haskell.cafe/43292 10. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/benchpress 11. http://article.gmane.org/gmane.comp.lang.haskell.cafe/43282 12. http://hackage.haskell.org/cgi-bin/hackage-scripts/package/wavconvert 13. http://article.gmane.org/gmane.comp.lang.haskell.cafe/43240 14. http://article.gmane.org/gmane.comp.lang.haskell.cafe/43230 15. http://community.haskell.org/~aslatter/code/xhb 16. http://article.gmane.org/gmane.comp.lang.haskell.cafe/43211 17. http://blog.downstairspeople.org/2008/08/16/ann-roguestar-022/ 18. http://hackage.haskell.org/trac/summer-of-code/wiki/SoC2008 19. http://code.haskell.org/hoogle/ 20. http://neilmitchell.blogspot.com/2008/08/gsoc-hoogle-week-12.html 21. http://haskell.org/hoogle/beta 22. http://haskell.org/hoogle 23. http://haskell.org/haskellwiki/Hpysics 24. http://www.haskell.org/haskellwiki/GHC/Data_Parallel_Haskell 25. http://code.haskell.org/~Saizan/cabal 26. http://www.sivity.net/projects/language.c/wiki/ 27. http://hackage.haskell.org/trac/ghc/wiki/GhcApiStatus 28. http://article.gmane.org/gmane.comp.lang.haskell.cafe/43175 29. http://www.peerium.com/ 30. file://localhost/home/brent/haskell/hwn/at] 31. http://planet.haskell.org/ 32. http://haskell.org/haskellwiki/Blog_articles 33. http://cgi.cse.unsw.edu.au/~dons/blog/2008/08/21#the_500 34. http://neilmitchell.blogspot.com/2008/08/hoogle-new-features.html 35. http://haskell.org/hoogle/ 36. http://blog.well-typed.com/2008/08/solving-the-diamond-dependency-problem/ 37. http://www.realworldhaskell.org/blog/2008/08/18/two-new-chapters-preorder-l… 38. http://sigfpe.blogspot.com/2008/08/untangling-with-continued-fractions_16.h… 39. http://blog.downstairspeople.org/2008/08/16/ann-roguestar-022/ 40. http://araujoluis.blogspot.com/2008/08/himerge-and-haskell-out-to-street.ht… 41. http://neilmitchell.blogspot.com/2008/08/gsoc-hoogle-week-12.html 42. http://twan.home.fmf.nl/blog/haskell/generic-merge.details 43. http://luqui.org/blog/archives/2008/08/14/slicing-open-the-belly-of-the-io-… 44. http://blog.malde.org/index.php/2008/08/14/the-wee-beginnings-of-a-biohaske… 45. http://logicaltypes.blogspot.com/2008/08/using-difference-lists.html 46. http://www.realworldhaskell.org/blog/2008/08/13/mid-august-status-update/ 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/

1 0

Re: [Haskell] Compiler Construction course using Haskell?
by Derek Elkins 21 Aug '08

21 Aug '08

On Wed, 2008-08-20 at 16:03 +0200, Johannes Waldmann wrote: > Hello. > > I plan to give a course in compiler construction, > using Haskell as the implementation language > (not as source or target language). > > Something along these lines: > 1. combinator parsers (Parsec), > 2. simple interpreter (arithmetical expressions) > 3. add algebraic data types, functions > 4. type checker > 5. code generator. > Ideally, 2..5 would be using the very same tree traversal code > and just change the monad for evaluation. > > Any comments appreciated. Have you given such a course? Taken? [I've replied to Haskell-Cafe which is a better list for this discussion.] 2 & 3 are going to have different trees so using the same tree traversal code would require some finesse, though the code for 2 should only need extension not change. One thing you may want to look at is attribute grammars which can be cast into a monadic framework and gives a natural example of using the backward state monad and allows you to connect to other formalisms used for compiler construction. Another possibility is abstract interpretation. Both code generation and type checking can be viewed as examples of abstract interpretation (as well as, of course, actual interpretation.) Also many analyses fit into the model of abstract interpretation. I'd also recommend transforming the initial AST to some intermediate form before doing 2-5. Not only is this a virtually universal practice, but it will the later stages, particularly the interpreter and the code generator easier to write and, in the code generator's case, able to produce better output. Ultimately, trying to have the same code produce all of these is going to lead to some compromises. You are either going to have to forsake some quality in the output, or you are going to have unnatural or, at least, non-trivial implementations of some the functions. The recommendation to for the use of a intermediate language mitigates this somewhat.

4 3

500 Haskell packages in Arch Linux
by Don Stewart 21 Aug '08

21 Aug '08

Monday was something of a landmark, as the 500th Haskell package was added to the Arch Linux distribution. I've written about the approach we took to fully automating the construction of native packages from Hackage, http://cgi.cse.unsw.edu.au/~dons/blog/2008/08/21#the_500 As well as a survey of 10 cool packages that are a little unknown. And some fun stats. On Hackage we have: * 723 unique packages * Over a million lines of Haskell code * 2600 pages of API documentation And in Arch, there are: 6 Erlang packages 21 OCaml packages 28 Lisp packages 101 Ruby packages 495 Python packages and now, 504 Haskell packages :) -- Don

1 0

Cyclic Inclusions
by C.M.Brown 20 Aug '08

20 Aug '08

Hi, I have a question about cyclic inclusions. It appears in the Haskell 98 report that mutually recursive modules are allowed, however GHC complains at any Haskell project that has cyclic inclusions (implicit or explicit). Am I right in thinking that this is a GHC limitation? http://www.haskell.org/onlinereport/modules.html Kind regards, Chris.

8 24

Re: [Haskell-cafe] Unboxing VT_VARIANT in hscom
by Krasimir Angelov 20 Aug '08

20 Aug '08

Aha. I got it. You should compile with -fglasgow-exts option. The extra type signatures doesn't matter. On Wed, Aug 20, 2008 at 8:20 AM, Praki Prakash <praki.prakash(a)gmail.com> wrote: > Krasimir - thanks for your reply. I had tried explicit typing but I still > get the same error. I have some code below and the error message. > > import Control.Exception > import Foreign > import Foreign.COM.Client > import Foreign.COM.Automation > import System.IO > > data WmiConnection = > WmiConnection { > execQuery :: String -> IO () > } > > main = withCOM $ do > conn <- wmiconnect "." "root\\cimv2" > execQuery conn "Select * From ..." > print "done" > > wmiconnect :: String -> String -> IO WmiConnection > wmiconnect comp cimroot = > do > clsid <- progid2clsid "WbemScripting.SWbemLocator" > (bracket (createInstance clsid iidIUnknown) release $ \iunkn -> > bracket (queryInterface iidIDispatch iunkn) release $ \idisp -> do > dispId <- getMethodID "ConnectServer" idisp > withBSTR comp $ \pComputer -> > withBSTR cimroot $ \pCimRoot -> do > (res,args) <- invoke dispId InvokeMethod [Variant VT_BSTR > pComputer, > Variant VT_BSTR > pCimRoot > ] idisp > print res > let conn = WmiConnection{ execQuery = queryFunc res } > return conn) > where > queryFunc :: Variant -> String -> IO () > queryFunc (Variant VT_DISPATCH idisp) query = do > dispId <- getMethodID "ExecQuery" idisp > return () > {- > queryFunc (Variant vt idisp) query = > if vt == VT_DISPATCH > then do > dispId <- getMethodID "ExecQuery" idisp > return "" > else > fail "error" > > return () > -} > > C:\>ghc --make test.hs > [1 of 1] Compiling Main ( test.hs, test.o ) > > test.hs:35:47: > Couldn't match expected type `IDispatch a' > against inferred type `a1' > `a1' is a rigid type variable bound by > the constructor `Variant' at test.hs:34:22 > In the second argument of `getMethodID', namely `idisp' > In a 'do' expression: dispId <- getMethodID "ExecQuery" idisp > In the expression: > do dispId <- getMethodID "ExecQuery" idisp > return () > > Any further suggestions? > > Thanks, > Praki > > On Tue, Aug 19, 2008 at 1:49 AM, Krasimir Angelov <kr.angelov(a)gmail.com> > wrote: >> >> This looks like a GHC bug to me. I am pretty sure that this worked >> before. Variant is defined like this: >> >> data Variant = forall a . Variant (VarType a) a >> >> data VarType a where >> .... >> VT_DISPATCH :: VarType (IDispatch ()) >> >> From this it clear that val is of type (IDispatch ()) because the >> VarType has value VT_DISPATCH. A workaround is to add explicit type >> singnature for val: >> >> someFunc (Variant VT_DISPATCH val) query = do >> dispId <- getMethodID "MethodName" (val :: IDispatch ()) >> >> I don't know why this doesn't work without the signature. >> >> Regards, >> Krasimir >> >> >> On Tue, Aug 19, 2008 at 7:09 AM, Praki Prakash <praki.prakash(a)gmail.com> >> wrote: >> > I am a Haskell newbie trying to do COM automation using Haskell. I am >> > using >> > hscom (Krasimir's implementation of COM automation). I have run into a >> > problem >> > and need some help. >> > >> > I have a Variant returned from a COM method invocation. When I print it, >> > it >> > shows up as below. >> > >> > Variant VT_DISPATCH <interface 0x00039f00> >> > >> > I need to invoke methods on the wrapped interface. My attempt to unbox >> > it as >> > below runs into 'rigid type' error. >> > >> > someFunc (Variant VT_DISPATCH val) query = do >> > dispId <- getMethodID "MethodName" val >> > >> > The code above generates this error. >> > >> > Couldn't match expected type `IDispatch a' >> > against inferred type `a1' >> > `a1' is a rigid type variable bound by... >> > >> > I am probably missing something pretty basic. Any help on this is >> > greatly >> > appreciated! >> > >> > Thanks >> > >> > >> > >> > >> > >> > _______________________________________________ >> > Haskell-Cafe mailing list >> > Haskell-Cafe(a)haskell.org >> > http://www.haskell.org/mailman/listinfo/haskell-cafe >> > > >

1 0

ANN: Tutorial on information visualization and visual analytics in Haskell
by Jefferson Heard 20 Aug '08

20 Aug '08

This is the tutorial I'll be presenting at DEFUN 2008. I'll be building a site around it until then, complete with compilable code examples, but I thought I would let everyone get a sneak peek at the long version of the tutorial before I'm done with it. The code is as yet untested, and keep in mind, advanced Haskellers, that I'm purposefully simplifying some things to be understood by the beginner to the intermediate Haskeller. Comments and questions are welcome and encouraged. Please do ignore typos in the inline code for now, though, as I'll be spending this week testing it out and making sure everything works. The link is: http://bluheron.europa.renci.org/docs/BeautifulCode.pdf -- Jeff

10 12

darcs hacking sprint (25-26 October)
by Eric Y. Kow 19 Aug '08

19 Aug '08

Dear Haskellers, As Jason mentions, there has been recently been a lot of interest in darcs and an influx of new contributors. Thanks, everybody! Things are moving faster now with a few extra brains on board. We have some very ambitious plans, though and we still need all the help we can get. Do you think you might be interested in participating, but don't really know where to start? We may have just the solution for you: Darcs hacking sprint -------------------- You are cordially invited to hack on darcs with us for a weekend, specifically, that of 25-26 October. We will primarily be working on performance hacking of all sorts http://wiki.darcs.net/DarcsWiki/Roadmap If there are sufficient new hackers, I plan to prepare an overview and a set of starter tasks to help you get to grips with the darcs code. Venues ------ The venues are not yet fixed, although we have some tentative offers for space in London (thanks to Ganesh) and Portland (thanks to Galois). We will also be available via IRC, with some possibility for video-conferencing of some sort. For more information, please see http://wiki.darcs.net/index.html/Sprints If you have any suggestions for possible venues in London, please let me know. Our current London host can accommodate a handful of people, but we would love to have more people joining in on the fun! Also, if there are people who cannot make it to either London and Portland, perhaps you could gather into satellite teams? I hear Boulder, Colorado is a possibility. RSVP ---- If you would like to participate in any way, please let me know * when you are available (25/26 October) * roughly where you are based * if you are willing to travel Again, if you have suggestions for a good London venue, please send them my way. Many thanks! -- Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow> PGP Key ID: 08AC04F9

2 2