Haskell Weekly News: August 23, 2005

Greetings, and thanks for reading the fourth issue of HWN, a weekly newsletter for the Haskell community. Each Tuesday, new editions will be posted (as text) to the Haskell mailing list and (as HTML) to The Haskell Sequence.

New Releases


Category theory monads. Cale Gibbard began a discussion comparing the monads from category theory with the implementation of monads in Haskell. Michael Vanier suggested some of Phil Wadler's papers on monads. Michael went on to say that Haskell monads are very similar to those from category theory.

More on FFI and callbacks. The thread on FFI and callbacks was revived this week. Among other things, Duncan Coutts noted that there is no easy way to do a really correct binding to wxWidgets from Haskell when Haskell programs are multithreaded, due to limitations in Haskell's threading model.

Oracle on Haskell. Brian Strand asked about using Oracle on Haskell, and more generally, about the suitability of Haskell for database programming. Alistair Bayley mentioned that takusen has Oracle support. John Goerzen suggested using HSQL's ODBC support, with unixODBC on *nix platforms. He went on to say that HSQL has been used in production environments. Krasimir Angelov, author of HSQL, added that he's been wanting to add Oracle support to HSQL for awhile. Finally, Brian Strand later followed up and said that takusen has been working well for him so far.

Pros and cons of static typing. Keean Schupke revived an earlier discussion about the pros and cons of static typing and side effects.

Static typing and interactivity. On a similar note, Ketil Malde wrote about not being able to load modules with type errors into ghci. Bernard Pope suggested the type debugger in Chameleon.

Argument ordering. A thread about the order of arguments to functions entertained many different viewpoints this week. Too many to really summarize here.

Decoupling and encapsulation. Terrence Brannon wrote about decoupling program elements from presentation, as in HTML generators. He included a link to his document describing architectural flaws in Perl's HTML::Mason.

Future of The Monad.Reader. Shae Erisson wrote on #haskell today that he is looking for someone to either take over, or help with, editing The Monad.Reader, Haskell's monthly online magazine. Anyone that would like to help should contact him.

How is HWN material found? I've received some questions this week about how I find material for HWN. The most obvious way is if someone sends it to me; see the link at the bottom of each HWN for contribution information. Other than that, I read the main Haskell mailing lists, the Haskell Sequence, and IRC looking for things to write about. I prefer to have stories linked in at least one of these places before covering them in HWN, since it gives readers a convenient place to follow discussion.

Haskell Toolchain

Cabal design. Frederik Eaton started another discussion about Cabal, this time focusing on run time vs. configure time issues, the usage of custom package.conf files, and multiple cabal files in one package. Isaac Jones commented on most of those items. Duncan Coutts suggested the ability to register a package "in place".

Darcs Corner

Colordiff. Dmitriy Morozov asked about using colordiff with darcs. Timo Savola suggested a shell function to accomplish this.

Linus and git. Juliusz Chroboczek posted a link to a discussion about Git, and mentioned that it looks like they're re-inventing some Darcs features.

Success with trac. Pedro Melo posted about his success using trac (a SourceForge-like system) with Darcs.

Quotes of the Week

Seen on #haskell today...

<tuomov> define drug <Itkovian> anything that gets you addicted to <Itkovian> potentially <Itkovian> and that messes with yr senses <Lemmih> Haskell?

About Haskell Weekly News

Want to continue reading HWN? Please help us create new editions of this newsletter. Please see the contributing information, or send stories to hwn -at- complete -dot- org. There is also a Darcs repository available.