W3C home > Mailing lists > Public > www-webdav-dasl@w3.org > January to March 2002

RE: Call for Participation: new internet draft for WebDAV SEARCH method

From: Jim Whitehead <ejw@cse.ucsc.edu>
Date: Fri, 25 Jan 2002 11:53:41 -0800
To: <www-webdav-dasl@w3.org>
Message-ID: <AMEPKEBLDJJCCDEJHAMIEEBFEBAA.ejw@cse.ucsc.edu>
I'm glad you're taking the lead on finishing a protocol for DAV searching. I
have updated the DASL pages at http://www.webdav.org/dasl/ to point to your
latest draft.

I have send a request in to the W3C to change the list to be more actively
moderated. I'm hoping we can cut down on the amount of spam received on this
list, and hence make it a better venue for discussion.

> Q1) Should the draft attempt to define Query Schema Discovery for
> DAV:basicsearch? Has anybody actually *implemented* QSD?

Well, on the surface it seems reasonable to allow a client to discover
whether a given query will be fast/slow, based on the implementation of the
server. But, that said, I'm not sure what a client woudl do with the
information. If a client's operation depends on executing a given query,
even if it discovers the query is inefficient against a given server, it's
still going to execute the query.

Discovering new operators might be useful, but only for clients that expose
the entire search syntax to a human operator.  I don't see how we could
provide enough information that a computational agent could learn about a
new operator in sufficient detail to know how to use it effectively. Either
such a client alreadyy knows about the operator, or it doesn't.

> Q2) Is there any real interest of defining a generic SEARCH framework that
> can exist independantly of WebDAV? If not, it could make sense just to
> define a REPORT based on property searching, and the complexity
> of the draft could be greatly reduced.

I have said before, and I'll say it again here: I like SEARCH, and prefer
its use for a general purpose DAV searching protocol. REPORT is good for
special-purpose queries.

Removing the SEARCH method from the specification provides only minimal
reduction in complexity for both the protocol specifier, and for
implementors.

- Jim
Received on Friday, 25 January 2002 14:54:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 22 March 2009 03:38:07 GMT