W3C home > Mailing lists > Public > www-archive@w3.org > October 2003

RE: catching up on DASL/WebDAV

From: Jim Whitehead <ejw@cse.ucsc.edu>
Date: Mon, 6 Oct 2003 12:42:43 -0700
To: "Dan Connolly" <connolly@w3.org>
Cc: "Massimo Marchiori" <massimo@w3.org>, "Ralph R. Swick" <swick@w3.org>, <www-archive@w3.org>, "Liam Quin" <liam@w3.org>, "C. M. Sperberg-McQueen" <cmsmcq@acm.org>, "Lisa Dusseault" <lisa@xythos.com>

Hi Dan,

Thanks for the update! Note that I'm cc'ing Lisa Dusseault, Co-Chair of the
WebDAV WG (and, incidentally, author of a book on WebDAV soon to be
published by Prentice-Hall).

In the WebDAV community, we've recently started a discussion on what the
future of DASL should be. There seems to be some consensus around shipping
the current specification more or less as-is, since it's quite useful for
searching unstructured DAV properties (that is, it provides a good way to
perform SQL-like queries on properties that map well to single relational DB
cells). While this isn't as general a protocol as I might like, in practice
it's pretty useful.

There are multiple interoperabile implementations of DASL currently in
existence, so we'll likely go to Proposed with the current specification,
though there is some discussion of going to Experimental. Follow-on efforts
for this protocol are likely to go in the direction of adding data typing to
DAV properties to permit better handling of user-defined ("dead") properties
(i.e., properties where the underlying relational database doesn't know how
to assign a RDBMS type). This will improve handling of conditionals, sort
ordering, etc. More discussion on typing of DAV properties can be found in a
recent paper of mine, "The WebDAV Property Design":

One question we're not likely to address in the WebDAV WG is how to remotely
perform XML and RDF queries. The DASL architecture is quite open -- a
request is a SEARCH verb, followed by a query specification in the request
body. This request body, if formulated correctly, could probably be used
with either SOAP marshalling or the SEARCH verb. So, the open question is
how to marshall XML and RDF queries that are executed over a network, either
in the presence of resources that have DAV properties, or not (ideally the
protocol should handle both cases).

IMO, this question has greater relevance to the W3C community than the DAV
community, and I'd be interested in moving consideration of this issue to an
appropriate W3C venue, of course subject to the typical concerns of whether
it's an appropriate W3C activity, if the W3C members are interested, etc.

- Jim
Received on Monday, 6 October 2003 15:44:26 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:42:32 UTC