[Haskell] Weaving the Web with Haskell

[Simon Peyton-Jones]

Graham 

Having solid libraries for XML and HTTP would be a jolly good thing.  As
you know, there is plenty of raw material around, but it obviously
doesn't do what you want, yet.

If you, together with other interested parties, were to get together to
do this, the rest of us would be very grateful.  You could do that by
directly working on the libraries distributed with GHC and Hugs (we
could give you CVS access).  Or you could make a separate project of it,
and distribute the stuff separately, which would free you from the GHC
and Hugs's release cycles.  

Simon

| -----Original Message-----
| From: haskell-bounces at haskell.org [mailto:haskell-bounces at haskell.org]
On Behalf Of Graham Klyne
| Sent: 28 January 2004 12:37
| To: Haskell Mailing List
| Subject: [Haskell] Weaving the Web with Haskell
| 
| For the past week or so, I've been wrestling with various bits of
Haskell
| libraries trying to get an XML parser running under Windows.  Each
time I
| think I've resolved a problem, another pops up to take its place.  So
I
| felt it was a good time to stand back and review my goals and
approach.
| 
| 
| Background
| 
| I believe that Haskell has a number of characteristics that make it
| eminently suitable for prototyping and deploying a range of web
| technologies.  My own interest is in the Semantic Web area, but I
think
| Haskell could also be usefully applied to Web Services and other web
| applications, and (as a language) potentially stands up well against
the
| likes of Java and Python.   The Web is spawning a host of little (and
| not-so-little) languages, and Haskell seems to have the right features
to
| handle these in a clean, principled fashion.  This is my working
premise.
| 
| Contemporary web applications are largely based on XML and HTTP.  So a
| prerequisite for widespread use of any language for Web applications
is
| good XML and HTTP support.  Given the right libraries, Haskell could
make
| handling these aspects very straightforward.  But there seems to be a
lack.
| 
| 
| Requirements
| 
| Ideally, I think that the following requirements should be satisfied:
| - full XML parser supporting all features of the current XML core and
| namespace specifications.
| - full HTTP support, including content negotiation, redirection,
| read/write, GET, HEAD, PUT, POST, etc.
| - available across all Haskell compilers on all platforms
| 
| Practically, I think the following would be enough to make a big
| difference, and this is what I'm currently looking for:
| - XML support, not necessary validating, but accepting full XML syntax
and
| supporting including internal entity definitions and substitution and
XML
| namespaces.
| - capability to access the representation of a Web Resource using HTTP
GET
| - supported uniformly by GHC and Hugs, on Linux and Windows platforms.
| 
| 
| Where I am
| 
| I have been trying to use the HXML toolbox, because I understand it's
the
| only XML parser for Haskell that supports XML namespaces.
Unfortunately,
| it seems to be rather dependent on older versions of GHC (unless I'm
| missing something), which is making it more problematic to adopt than
I had
| hoped.  Other packages, such as HaXml, seem to be more portable but as
far
| as I'm aware are missing key functionality (notably XML namespace
support).
| 
| 
| Looking forward
| 
| I shall plug away for a little while with HXML toolbox -- currently
I'm
| stuck for an MD5 library module -- and it may be that, with a little
help,
| I'll win through:  if I get it running with Hugs under Windows, I'll
| consider that to be significant progress.
| 
| But maybe I'm missing a trick:  is there other software I should be
| considering to satisfy my goals?  Constructive suggestions would be
welcome.
| 
| I'm anticipating that as the library infrastructure project settles,
we'll
| see some effort to move more of this kind of support code into a
common and
| generally usable function library.  I hope I shall be able to
contribute to
| such an outcome.
| 
| I suggest that the XML and HTTP support I've mentioned would be a key
| enabler to allow Haskell to be more widely used for real-world
projects as
| well as academic research into language design issues (which currently
| seems to be its strongest developer community interest).
| 
| #g
| 
| 
| ------------
| Graham Klyne
| For email:
| http://www.ninebynine.org/#Contact
| 
| _______________________________________________
| Haskell mailing list
| Haskell at haskell.org
| http://www.haskell.org/mailman/listinfo/haskell