April 2013 - Haskell-Cafe

Automated Differentiation of Matrices (hmatrix)
by Dominic Steinitz 10 Apr '13

10 Apr '13

Hi Cafe, Suppose I want to find the grad of a function then it's easy I just use http://hackage.haskell.org/package/ad-3.4: import Numeric.AD import Data.Foldable (Foldable) import Data.Traversable (Traversable) data MyMatrix a = MyMatrix (a, a) deriving (Show, Functor, Foldable, Traversable) f :: Floating a => MyMatrix a -> a f (MyMatrix (x, y)) = exp $ negate $ (x^2 + y^2) / 2.0 main :: IO () main = do putStrLn $ show $ f $ MyMatrix (0.0, 0.0) putStrLn $ show $ grad f $ MyMatrix (0.0, 0.0) But now suppose I am doing some matrix calculations http://hackage.haskell.org/package/hmatrix-0.14.1.0 and I want to find the grad of a function of a matrix: import Numeric.AD import Numeric.LinearAlgebra import Data.Foldable (Foldable) import Data.Traversable (Traversable) g :: (Element a, Floating a) => Matrix a -> a g m = exp $ negate $ (x^2 + y^2) / 2.0 where r = (toLists m)!!0 x = r!!0 y = r!!1 main :: IO () main = do putStrLn $ show $ g $ (1 >< 2) ([0.0, 0.0] :: [Double]) putStrLn $ show $ grad g $ (1 >< 2) ([0.0, 0.0] :: [Double]) Then I am in trouble: /Users/dom/Dropbox/Private/Whales/MyAD.hs:24:21: No instance for (Traversable Matrix) arising from a use of `grad' Possible fix: add an instance declaration for (Traversable Matrix) In the expression: grad g In the second argument of `($)', namely `grad g $ (1 >< 2) ([0.0, 0.0] :: [Double])' In the second argument of `($)', namely `show $ grad g $ (1 >< 2) ([0.0, 0.0] :: [Double])' /Users/dom/Dropbox/Private/Whales/MyAD.hs:24:26: Could not deduce (Element (ad-3.4:Numeric.AD.Internal.Types.AD s Double)) arising from a use of `g' from the context (Numeric.AD.Internal.Classes.Mode s) bound by a type expected by the context: Numeric.AD.Internal.Classes.Mode s => Matrix (ad-3.4:Numeric.AD.Internal.Types.AD s Double) -> ad-3.4:Numeric.AD.Internal.Types.AD s Double at /Users/dom/Dropbox/Private/Whales/MyAD.hs:24:21-26 Possible fix: add an instance declaration for (Element (ad-3.4:Numeric.AD.Internal.Types.AD s Double)) In the first argument of `grad', namely `g' In the expression: grad g In the second argument of `($)', namely `grad g $ (1 >< 2) ([0.0, 0.0] :: [Double])' What are my options here? Clearly I can convert my matrix into a list (which is traversable), find the grad and convert it back into a matrix but given I am doing numerical calculations and speed is an important factor, this seems undesirable. I think I would have the same problem with: http://hackage.haskell.org/package/repa http://hackage.haskell.org/package/yarr-1.3.1 although I haven'¯t checked. Thanks, Dominic.

3 5

abs on Float/Doubles
by Levent Erkok 10 Apr '13

10 Apr '13

Hello all, I definitely do not want to create a yet another thread on handling of floating-point values in Haskell. My intention is only to see what can be done to make it more "compilant" with IEEE754. (Also, my interest is not a theoretical one, but rather entirely practical: I'm interested in using SMT solvers for reasoning about Haskell programs, in particular using the new SMT logics that cover IEEE-754 semantics. So, any discrepancy between Haskell and IEEE-754 is a cause for concern as we strive to build powerful tools around the Haskell eco-system.) In IEEE-754, there are two zero values: +0, and -0; inhabiting all floating point types. IEEE-754 requires these two compare equal, but be distinguished when used as arguments to functions as they can produce different results. (Unfortunately the standard is not freely available, but there's a nice discussion in wikipedia: http://en.wikipedia.org/wiki/Negative_zero.) As far as I know, Haskell's floating-point handling does *not* explicitly claim to be IEEE-754 compliant, but I think it's safe to assume that any serious implementation will take advantage of the underlying CPU's floating-point facilities; and thus will use IEEE-754 semantics. In particular, Haskell supports both +0 and -0; and you can distinguish them using the isNegativeZero function of the RealFloat class. Furthermore, the show instance for Float/Double will turn these two values to strings differently, preserving the minus sign if it receives a negative-zero. Comparisons follow the IEEE rules; so all these behaviors seem to be in accordance with IEEE754. Unfortunately, the same isn't true for the abs function. In Haskell, the call: "isNegativeZero (abs (-0::Double))" evaluates to "True". IEEE754 requires abs to always return "+0" in this case. The same also holds for the semantics of the newly proposed SMT-Lib logic for IEEE754: See section 3.3.1 (page 8) of: http://www.philipp.ruemmer.org/publications/smt-fpa.pdf The fix is easy. A one line change in the Double/Float instances of the Num class in the Prelude would guarantee the IEEE754 semantics. The reason I'm not simply posting this as a "library" issue is that it appears the topic came up in stack-overflow a while ago; without any clear resolution: http://stackoverflow.com/questions/10395761/absolute-value-of-negative-zero… It appears that the consensus is that this is a historical definition dating back to the times when IEEE754 itself wasn't quite clear on the topic itself, and "so nobody thought that hard about negative zeroes." (The quote is from a comment from Lennart.) I think it's a safe bet to assume IEEE754 semantics (for better or worse) is here to stay, and it would be nice to reduce any discrepancy we have in the Haskell libraries with respect to it. So, if there's consensus, I'd like to propose a library change to implement the proper IEEE754 semantics for the abs function for Double and Float types. Note that the argument is not really about semantics here, but rather with compliance to the accepted industrial standard. So, it would be best if we kept the discussion purely to the IEEE-compliance here, as opposed to relative merits of the semantics as defined. -Levent.

2 2

SBLP 2013: Final CFP, deadline for Abstracs 19/4
by Andre Rauber Du Bois 10 Apr '13

10 Apr '13

[Apologies if you receive multiple copies of this CFP] ======================================================= CALL FOR PAPERS 17th BRAZILIAN SYMPOSIUM ON PROGRAMMING LANGUAGES Brasília, Distrito Federal, Brazil September 29th to October 4th, 2013 http://cbsoft2013.unb.br/en/sblp-en ======================================================== IMPORTANT DATES Paper abstract submission (15 lines): April 19th Full paper submission: April 26th, 2013 Notification of acceptance: May 31st, 2013 Final papers due: June 28th, 2013 INVITED SPEAKERS * Tim Harris, Oracle Labs * More TBA INTRODUCTION The 17th Brazilian Symposium on Programming Languages, SBLP 2013, will be held in Brasília, Brazil, on September 29th to October 4th, 2013. SBLP provides a venue for researchers and practitioners interested in the fundamental principles and innovations in the design and implementation of programming languages and systems. The symposium will be part of the 4th Brazilian Conference on Software: Theory and Practice, CBSoft 2013, http://cbsoft2013.cic.unb.br/, which will host four traditional, well-established symposia: * XXVII Brazilian Symposium on Software Engineering (SBES) * XVII Brazilian Symposium on Programming Languages (SBLP) * XVI Brazilian Symposium on Formal Methods (SBMF) * VII Brazilian Symposium on Components, Software Architecture and Software Reuse (SBCARS) SBLP 2013 invites authors to contribute with technical papers related (but not limited) to: * Program generation and transformation, including domain-specific languages and model-driven development in the context of programming languages. * Programming paradigms and styles, including functional, object-oriented, aspect-oriented, scripting languages, real-time, service-oriented, multithreaded, parallel, and distributed programming. * Formal semantics and theoretical foundations, including denotational, operational, algebraic and categorical. * Program analysis and verification, including type systems, static analysis and abstract interpretation. * Programming language design and implementation, including new programming models, programming language environments, compilation and interpretation techniques. SUBMISSIONS Contributions should be written in Portuguese or English. We solicit papers that should fall into one of two different categories: full papers, with at most 15 pages, or short papers, with at most 5 pages. Full papers submitted in English will be published in a volume of Lecture Notes in Computer Science (LNCS), by Springer. For this reason, all papers must be prepared using the LNCS template, available at http://www.springer.com/computer/lncs?SGWID=0-164-6-793341-0. We encourage the submission of short papers reporting on partial results of on-going master dissertations or doctoral theses. All accepted papers will be published in the conference proceedings distributed in a digital media by the CBSOFT organizers. Submissions should be done using SBLP 2013 installation of the EasyChair conference mangement system at http://www.easychair.org/conferences/?conf=sblp2013. As in previous editions, a journal special issue, with selected papers from accepted contributions, is anticipated. Selected papers from 2003 to 2008 editions of SBLP were published in special issues of the Journal of Universal Computer Science, by Springer. The post-proceedings of SBLP from 2009 to 2012, also with selected papers from the conference proceedings, are being edited as special issues of Science of Computer Programming, published by Elsevier. CBSOFT CHAIRS Genaina Nunes Rodrigues, UnB Rodrigo Bonifácio, UnB Diego Aranha, UnB PROGRAMME CHAIRS André Rauber Du Bois, UFPel Phil Trinder, Glasgow University PROGRAMME COMMITTEE Alberto Pardo, Univ. de La Republica Alex Garcia, IME Alvaro Freitas Moreira, UFRGS André Rauber Du Bois, UFPel (co-chair) Andre Santos, UFPE Carlos Camarao, UFMG Christiano Braga, UFF Edwin Brady, University of St. Andrews Fernando Castor Filho, UFPE Fernando Quintão Pereira, UFMG Francisco Heron de Carvalho Junior, UFC Hans-Wofgang Loidl, Heriot-Watt University Jeremy Singer, Glasgow University Joao Saraiva, Universidade do Minho João F. Ferreira, Teesside University Lucilia Figueiredo, UFOP Luis Soares Barbosa, Univ. do Minho Manuel António Martins, Univ. de Aveiro Marcelo A. Maia, UFU Marcello Bonsangue, Leiden Univ/CWI Marcelo d'Amorim, UFPE Marco Tulio Valente, UFMG Mariza A. S. Bigonha, UFMG Martin A. Musicante, UFRN Noemi Rodriguez, PUC-Rio Peter Mosses, Swansea University Phil Trinder, Glasgow University (co-chair) Qiu Zongyang, Beijing University Rafael Dueire Lins, UFPE Renato Cerqueira, PUC-Rio Ricardo Massa, UFPE Roberto S. Bigonha, UFMG Roberto Ierusalimschy, PUC-Rio Sandro Rigo, UNICAMP Sergio Soares, UFPE Simon Thompson, Univ. of Kent Varmo Vene, Univ. de Tartu

1 0

ghc head creates premature tarball
by Takayuki Muranushi 10 Apr '13

10 Apr '13

Hi all, These few days I've been having trouble in attempt to use GHC head. I follow the following standard steps to create a tarball: $ git clone http://darcs.haskell.org/ghc.git/ $ cd ghc $ # optionally, git checkout <branch name> $ ./sync-all --testsuite get $ cp mk/build.mk.sample mk/build.mk $ ### edit BuildFlavour to quick ### $ ./boot $ ./configure $ make -jN # <N> is the number of cores you have. $ make binary-dist and then sandbox-install the tarball by hsenv. However, the installed ghc exhibits the following strange behaviors: (1) several .so files need to be manually copied to /usr/lib . (2) the installed ghc doesn't find any package, even 'base' or 'ghc-prim' Then carefully looking, I found that `make binary-dist` is outputting some error messages,: > cd bindistprep && "/bin/tar" hcf - -T ../bindist-list | bzip2 -c > ../bindistprep/ghc-7.7.20130409-x86_64-unknown-linux.tar.bz2 > /bin/tar: ghc-7.7.20130409/compiler/stage2/doc: Cannot stat: No such file or directory > /bin/tar: Exiting with failure status due to previous errors The file "./bindist-list" included a filepath that doesn't exist, so that tar command aborted there and every item after that line on the list was excluded from the tarball. So I tested the following trick; $ touch compiler/stage2/doc Then the previous error message didn't show up any more, and the spurious behavior (1) (2) didn't appear any more. I'm wondering if this is an installer bug or I was doing something wrong. I'd appreciate any advice. Best, -- Takayuki MURANUSHI The Hakubi Center for Advanced Research, Kyoto University http://www.hakubi.kyoto-u.ac.jp/02_mem/h22/muranushi.html

1 0

[Haskell-cafe] algorithm-for-finding-numerical-permutation-given-lexicographic-index
by Lone Wolf 10 Apr '13

10 Apr '13

> > > http://stackoverflow.com/questions/8940470/algorithm-for-finding-numerical-… > > How would you rewrite this into Haskell? The code snippet is in Scala. > > /** > example: index:=15, list:=(1, 2, 3, 4) > */ > def permutationIndex (index: Int, list: List [Int]) : List [Int] = > if (list.isEmpty) list else { > val len = list.size // len = 4 > val max = fac (len) // max = 24 > val divisor = max / len // divisor = 6 > val i = index / divisor // i = 2 > val el = list (i) > el :: permutationIndex (index - divisor * i, list.filter (_ != el)) } > > >

3 3

[ANN] hoodle-publish-0.1 : publish your hoodle files on the web
by Ian-Woo Kim 10 Apr '13

10 Apr '13

Hi, all, I am happy to announce a utility hoodle-publish for publishing hoodle files on the web as a form of pdf files. I am planning to support html publishing later. For those who do not know what hoodle is, hoodle (http://ianwookim.org/hoodle) is a pen notetaking program written entirely in haskell. hoodle-publish is a shell command which takes hoodle files in a directory recursively and transforms them to pdf files preserving hyperlinks. For detailed explanation, please refer to my blog post : http://ianwookim.org/blog/posts/2013/04-09-publish-your-hoodle-files-on-the… A sample hoodle publication is also shown in http://ianwookim.org/sasha/hoodle/main.pdf Enjoy! Ian-Woo

1 0

GSoC Project Proposal: Markdown support for Haddock
by Johan Tibell 09 Apr '13

09 Apr '13

Hi all, Haddock's current markup language leaves something to be desired once you want to write more serious documentation (e.g. several paragraphs of introductory text at the top of the module doc). Several features are lacking (bold text, links that render as text instead of URLs, inline HTML). I suggest that we implement an alternative haddock syntax that's a superset of Markdown. It's a superset in the sense that we still want to support linkifying Haskell identifiers, etc. Modules that want to use the new syntax (which will probably be incompatible with the current syntax) can set: {-# HADDOCK Markdown #-} on top of the source file. Ticket: http://trac.haskell.org/haddock/ticket/244 -- Johan

28 45

Prolog-style patterns
by Jan Stolarek 09 Apr '13

09 Apr '13

Hi all, consider this simple reimplementation of 'elem' function: member :: Eq a => a -> [a] -> Bool member _ [] = False member y (x:xs) | x == y = True | otherwise = member y xs If Haskell allowed to write pattern matching similar to Prolog then we could write this function like this: member :: Eq a => a -> [a] -> Bool member _ [] = False member x (x:_) = True member x (_:xs) = member x xs The meaning of pattern in the second equation is "match this equation if the first argument equals to head of the list". Many times I have found myself instinctively writing patterns in this way, only to get a compilation error. I was thinking about implementing language extension for GHC that would allow to write Prolog-style patterns. Would there be an interest in such an extension? Also, if I missed something obvious please let me know. Janek

15 21

Can any hackage library do JWT signing
by Ryan Newton 09 Apr '13

09 Apr '13

I want to enable a Haskell benchmarking bot that would upload its data to a Google Fusion Table each time it runs. That seems to require the "service-to-service" mode for Google's OAuth2: https://developers.google.com/accounts/docs/OAuth2ServiceAccount There are a bunch of libraries on Hackage for different flavors of OAuth. But I haven't yet found one that handles this signing of JWT headers and claimsets. Is there something I'm missing? Thanks, -Ryan

1 0

Re: [Haskell-cafe] [Caml-list] Vector Fabrics is hiring!
by Alexey Rodriguez 09 Apr '13

09 Apr '13

Hi Phil, I am posting this on behalf of Martijn Rutten. Cheers, Alexey 8-<--------------- Fair point. When we started the company in 2007, the experts in Silicon Valley told us that our ambition to develop tools to create an efficient, parallel embedded system from sequential C and C++ code is the "holy grail" of embedded system design. In the same sentence, they told us it could not be done. It is widely acknowledged as well as debated that there can be an order of magnitude difference between a real good and an average programmer (http://programmers.stackexchange.com/questions/14052/is-there-such-a-thing-…)) Often these talented programmers are called a "rockstar", "free electron" (http://www.randsinrepose.com/archives/2005/03/20/free_electron.html) or just plain "awesome" for their ability to quickly sort out the many trade-offs in software architecture and design. Without these talented programmers to design and implement the algorithms to make sense out of the literally millions of data dependencies that can make or break parallelism, our company's vision would still be just that: a vision. You are right, if the talent comes with a rockstar's ego, forget it. No matter how smart, you are always part of a team. As for craftsmanship, what we did learn is the huge difference between functional programming and imperative programming. Our analysis engine is fully written in OCaml. Obviously that allowed us to focus on the algorithmic problems at hand without ever wasting time in debugging stray pointers (the one week debugging such a Heisenbug was spent in the few hundred lines of C interfacing with OCaml). I am sure I am preaching to the choir here when I say we learned that thinking functionally allowed us to much better reason at the algorithmic level and we generally wrote significantly cleaner and more reliable code. My 2 cents, Martijn Rutten, Co-founder of Vector Fabrics. On Mon, Apr 8, 2013 at 10:09 PM, Phil Tomson <philtomson(a)gmail.com> wrote: > On Sunday, April 7, 2013 10:30:40 PM UTC-7, Stefan Holdermans wrote: > >> >> * Your friends and colleagues describe you as a "rockstar" programmer; >> >> your programming ability is way above average; >> > > I thought we (as in industry) were over this "rockstar programmer" thing? What does a "rockstar" do? S/He breaks guitars when not happy. Rockstars are petulant and hard to work with. We need less rockstars and more craftspeople. > > Just my 2 cents. > > Phil > -- Alexey Rodriguez Yakushev O +31 (0)40 8200960 | D + 31 (0)40 8200974 | F +31 (0)40 8200979 Vonderweg 22, 5616 RM | Eindhoven | The Netherlands www.vectorfabrics.com | alexey(a)vectorfabrics.com On Mon, Apr 8, 2013 at 10:09 PM, Phil Tomson <philtomson(a)gmail.com> wrote: > On Sunday, April 7, 2013 10:30:40 PM UTC-7, Stefan Holdermans wrote: > >> >> * Your friends and colleagues describe you as a "rockstar" programmer; >> >> your programming ability is way above average; >> > > I thought we (as in industry) were over this "rockstar programmer" thing? What does a "rockstar" do? S/He breaks guitars when not happy. Rockstars are petulant and hard to work with. We need less rockstars and more craftspeople. > > Just my 2 cents. > > Phil > -- Alexey Rodriguez Yakushev O +31 (0)40 8200960 | D + 31 (0)40 8200974 | F +31 (0)40 8200979 Vonderweg 22, 5616 RM | Eindhoven | The Netherlands www.vectorfabrics.com | alexey(a)vectorfabrics.com

1 0