November 2008 - Haskell-Cafe

Monadic bind with associated types + PHOAS?
by oleg＠okmij.org 19 Nov '08

19 Nov '08

Ryan Ingram wrote: > One thing that often comes up is a desire to do a pass on the > resultant code to optimize it, but it's pretty difficult with the > standard monadic formulation because of embedded functions. You can't > do introspection on functions in Haskell; they aren't elements of Eq > or Show. This has caused, for example, some implementations of FRP to > switch to using arrows instead. However, I find the arrow syntax > cumbersome and I feel it tends to obfuscate what is actually going on > in the code. > > An earlier talk at ICFP mentioned using PHOAS instead of HOAS to > encode expressions. In HOAS, a lambda expression in the target > language is represented by a function in the source language: If one uses a tagless final representation, one benefits from the conveniences of the higher-order abstract syntax and yet gains the ability to show terms, count the number of constructors and even partially evaluate them (i.e., optimize). For the example, please see http://okmij.org/ftp/tagless-final/Incope.hs (in particular, see Semantics2 for the partial evaluator that does not use GADT). We can easily add an interpreter that does a CPS transform, thus letting us embed a CBV object language in a Haskell without any use of monad or the change of syntax of the object language. We can easily extend the language to add state (as we have done in the OCaml part of the project). http://okmij.org/ftp/tagless-final/README.txt

1 0

Haskell library support
by Galchin, Vasili 18 Nov '08

18 Nov '08

Hello, I am looking for something to work on. Where are there perceived holes in the Haskell library support? Regards, Vasili

3 4

FFI and returning structs
by Maurício 18 Nov '08

18 Nov '08

Hi, I have not found examples of this in documentation or hackage packages: how can I deal with funcions returning not pointers to structs but structs thenselves? struct example {... example function_name (... My intuition said I could try a data declaration that is a properly done instance of Storable, and then use that. data Example = ... instance Storable Example where ... But it seems I was wrong. Do you have any tips on how to proceed? Thanks, Maurício

1 0

How to make interface IO() to IO a in a command-driven tool
by Kyongho Min 18 Nov '08

18 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

2 1

clojure's data structures (was: Re: [Haskell-cafe] implementing python-style dictionary in Haskell)
by Evan Laforge 18 Nov '08

18 Nov '08

This actually brings up something I was wondering about lately. I recently stumbled across a language called clojure, which is a lisp-like that runs on the JVM. The interesting thing is that mutations go through a transactional mutable reference system, and the other datastructures are all immutable. The author implements an immutable hash map with a trie on each 5 bits of the hash, so each node potentially has 32 children. This means that lookups are effectively O(1) and alterations have to copy a maximum of 7 chunks of 32 pointers. Extendable "vectors" are implemented similarly. The hash tables sound basically like an IntMap on the hash code, except as far as I know, IntMap's patricia tree splits up the bit space adaptively instead of being hardcoded on 5 bit segments. I'm not sure what the performance difference would be. I haven't run any tests, but the clojure author claims his vector and hash map are quite fast. I suppose the extendable vector corresponds to the current crop of ByteString-like variants, so we sort of already have that, though in a kind of chaotic and inchoate way. I'm also curious about how far the generalized trie SoC project got... it should be more accessible now that we have a mainline release with ATs, right?

1 0

Deploying a Binary Haskell Package
by John Van Enk 18 Nov '08

18 Nov '08

This question isn't directly related to Haskell, but I figure some one might know here. I want to deploy an application. I could either: 1) Tell people how to download GHC, have them check out the repository, have them install all the needed hackage packages, ... or 2) Give them a setup.msi/exe I'm curious if any one has done this with a Haskell package. It seems that it's something that might make sense being integrated into Cabal (runhaskell Setup msi). Either way, has some one deployed a Haskell binary as a MSI package? /jve

3 2

Proof of a multi-threaded application
by Silviu ANDRICA 18 Nov '08

18 Nov '08

Hello, I am very new to Haskell, this is my first day, and I wanted to know if it is possible to prove correctness of a multi-threaded application written in Haskell. Basically, I want to check that a multi-threaded implementation of an algorithm that detects cycles in a dynamically changing graph, actually detects the cycles. Also, the algorithm prevents previously detected cycles from happening by not allowing certain edges to be added. And I want to check that this property also holds. Thank you and I hope you can help me, Silviu ANDRICA.

7 10

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