June 2013 - Haskell-Cafe

writing a function to make a correspondance between type-level integers and value-level integers
by oleg＠okmij.org 25 Jun '13

25 Jun '13

Well, I guess you might be interested in geometric algebra then http://dl.acm.org/citation.cfm?id=1173728 because Geometric Algebra is a quite more principled way of doing component-free calculations. See also the web page of the author http://staff.science.uva.nl/~fontijne/ Geigen seems like a nice DSL that could well be embedded in Haskell. Anyway, the reason I pointed out Vectro is that it answers your question about reifying and reflecting type-level integers (by means of a type class).

2 1

Roman Numeral Problem
by Christopher Howard 25 Jun '13

25 Jun '13

Hi. I am working on some practice programming problems, and one is the Roman numeral problem: write a program that converts Roman numerals into their (arabic) numeral equivalent. I imagine I could hack something together, but I was trying to think about the problem a bit more deeply. I don't know much about parsing, but I was wondering if this problem matched up with some kind of parsing or grammar or other generic approach to thinking about the problem. (I once read something about Context Free Grammars, which was rather interesting.) I can do more of my own research if I can get some initial guidance. -- frigidcode.com

4 3

ANNOUNCE: haskell-names-0.1
by Roman Cheplyaka 25 Jun '13

25 Jun '13

I am pleased to announce the first public release of haskell-names, a name resolution library for haskell-src-exts AST. Namely, it can do the following: * for a module, compute its interface, i.e. the set of entities exported by the module, together with their original names. * for each name in the module, figure out what it refers to — whether it's bound locally (say, by a where clause) or globally (and then give its origin). Thanks to haskell-packages, this library is fully integrated with Cabal, so that you can easily generate name interfaces for any Cabalized package. See more details in the README: http://documentup.com/haskell-suite/haskell-names This library is based on the code written by Lennart Augustsson in 2010. Little of that code survived, but nevertheless it's been a great help. Roman

6 8

tangential request...
by Mark Lentczner 25 Jun '13

25 Jun '13

As you may know, I've been writing a shell in Haskell called Plush<https://code.google.com/p/plush/> . I'm working on the look of the terminal input and output, and I want to do a little data gathering. I want to know what your terminal looks like! You can help me by: 1) Download this file: http://www.ozonehouse.com/mark/plush/TerminalDisplay.txt 2) Open a new terminal window 3) Do not resize the terminal window 4) cat the file 5) Take a screen shot of the whole terminal window 6) E-mail it to me: mark.lentczner+term(a)gmail.com The file contains some sample text, including some screen codes to change colors and bold. This looks like: - in my terminal program <http://www.ozonehouse.com/mark/plush/term.png> - in Plush <http://www.ozonehouse.com/mark/plush/term-plush.png> (bleeding edge, not checked in yet) Thanks!

10 16

[OT]Help with translating Error Monads to C++
by Larry Evans 24 Jun '13

24 Jun '13

Currently in the c++ developers ml, there's a discussion of a proposed Expected<T> or an Either<L,R> template: http://article.gmane.org/gmane.comp.lib.boost.devel/242496 A lot of the discussion appears, at least to me, to not be aware of how haskell handles similar situations. The above link was to my post suggesting they try to emulate the way haskell solves the problem using Monads. However, I've only briefly dabbled in haskell; hence, my responses on that thread are probably missing something. I invite anyone on this list to please provide insights on this particular thread in the boost.devel mailing list. -kind regards, Larry

1 0

Munich Haskell Meeting
by Heinrich Hördegen 24 Jun '13

24 Jun '13

Dear all, if you plan to join our monthly Munich Haskell Meeting on Wed, 26th of June, please go to http://www.haskell-munich.de/dates and hit the button. See you soon, Heinrich -- -- Funktionale Programmierung Dr. Heinrich Hördegen Gutenbergstr. 26 80638 München FON: +49 (89) 12 59 79 49 FAX: +49 (89) 12 59 79 50 hoerdegen(a)funktional.info www.funktional.info --

1 0

A friendly reminder: Ghent-FPG meeting on 26 June, 2013
by Andy Georges 24 Jun '13

24 Jun '13

Hello, This is to remind you that you are kindly invited to attend our next meeting. The original email follows below. The "Functional Programming Group Ghent" (GhentFPG) [1] is a friendly group for all people interested in functional programming, with a tendency towards Haskell. It is organised as part of Zeus WPI [2]. We are pleased to announce that we will hold a next meeting on Wednesday, 26th of June, starting at 19h00! There will be three talks. The main presentation, by Adam Bergmark from Silk [3] is about Fay [4]: Fay is a proper subset of Haskell that compiles to JavaScript. There is a compiler with the same name written in Haskell. Web browsers only speak JavaScript but more and more people find that they want to compile to JavaScript instead. Why do we want to compile Haskell to JavaScript, and what advantages does Fay have compared to other compilers? What are the challenges in compiling Haskell and supporting a language ecosystem, and how do we do it? What can Fay currently do, and what is planned for the future? This will be a broad overview about Fay for prospective users, followed by an in-depth look at interesting parts of the compiler internals. Additionally, there will be two short talks by two students who did an Msc. Thesis about functional programming languages: "Genetic Algorithms in Haskell" by Matthias Delbar "Automatic Detection of Recursion Patterns" by Jasper Van der Jeugt The meeting will take place in the Jozef Plateauzaal at the following address, Faculteit Ingenieurswetenschappen Universiteit Gent Plateaustraat 22 9000 Gent As mentioned above, we aim to start at 19:00. After the meeting we can go for drinks in a nearby pub (this latter part is, of course, completely optional) We hope to see you all there! Regards, On behalf of the GhentFPG organising committee. [1]: http://groups.google.com/group/ghent-fpg [2]: http://zeus.ugent.be/ [3]: http://www.silkapp.com/ [4]: http://www.fay-lang.org/

1 0

New Functional Programming Job Opportunities
by Functional Jobs 24 Jun '13

24 Jun '13

Here are some functional programming job opportunities that were posted recently: Senior software developer/Functional programmer at Vector Fabrics http://functionaljobs.com/jobs/154-senior-software-developer-functional-pro… Cheers, Sean Murphy FunctionalJobs.com

1 0

Fwd: How to make this data type work?
by Miguel Mitrofanov 23 Jun '13

23 Jun '13

Forgot to reply all, as usual. -------- Пересылаемое сообщение -------- 21.06.2013, 12:52, "Miguel Mitrofanov" <miguelimo38(a)yandex.ru>: Actually, this is not the real error you should care about. Try removing FromJSON instance completely, and you'll get a lot more. And these are fundamental: you have to decide what "j" to use when serializing. Haskell won't automagically substitute some suitable type for you. So, that's a classic mismatch: for serializing (ToJSON) you need your "j" type to be known to the AD value (meaning: it should be quantified existentially), but for deserializing you need it to be any type (quantified universally). All in all, AD seems to be the wrong type. 21.06.2013, 12:18, "Magicloud Magiclouds" <magicloud.magiclouds(a)gmail.com>: > data ActionData = AD { oldData :: (FromJSON j, ToJSON j) => j > , newData :: (FromJSON j, ToJSON j) => j} > instance ToJSON ActionData where > toJSON (AD o n) = object [ "oldData" .= o > , "newData" .= n ] > instance FromJSON ActionData where > parseJSON (Object v) = AD > <$> v .: "oldData" > <*> v .: "newData" > parseJSON _ = mzero > > I got when compile: > No instance for (FromJSON (forall j. (FromJSON j, ToJSON j) => j)) > arising from a use of `.:' > Possible fix: > add an instance declaration for > (FromJSON (forall j. (FromJSON j, ToJSON j) => j)) > In the second argument of `(<$>)', namely `v .: "oldData"' > In the first argument of `(<*>)', namely `AD <$> v .: "oldData"' > In the expression: AD <$> v .: "oldData" <*> v .: "newData" > > -- > 竹密岂妨流水过 > 山高哪阻野云飞 > > And for G+, please use magiclouds#gmail.com. > , > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe(a)haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe -------- Завершение пересылаемого сообщения --------

3 2

putting the result of a function in `infix` declaration
by TP 23 Jun '13

23 Jun '13

Hi everyone, Consider the following code: -------------- import Prelude hiding ((+)) data Expr = Plus Expr Expr | Minus Expr Expr | Expr String deriving Show f :: Expr -> Int f (Plus _ _) = 1 f (Minus _ _) = 2 -- infix (f (Plus undefined undefined)) + -- (*) infix 6 + (+) :: Expr -> Expr -> Expr e1 + e2 = Plus e1 e2 main = do let a = Expr "a" let b = Expr "b" print $ a + b -------------- I would like to declare the infix precedence of (+) with a function (commented line (*)) instead of directly as above. Is there any means to do that? Do we need some kind of preprocessing machinery? How to do that in Haskell? In a more general way, let us suppose I have a Haskell library able to perform some calculations: how to use it in a pre-processing step, before compilation of the executable? It could be a function to compute symbolically roots of a polynomial of second degree to be used at runtime. We would put some placeholders in the code where the result of the pre- processing calculation would enter. Thanks in advance, TP

3 2