RE: RE: Future direction for DASL/WebDAV SEARCH

> From: Kevin Wiggen [mailto:kwiggen@xythos.com]
> Sent: Thursday, October 02, 2003 7:39 PM
> To: Jim Whitehead; WebDAV; www-webdav-dasl@w3.org
> Subject: [Moderator Action] RE: Future direction for DASL/WebDAV SEARCH
>
>
>
>
>
>
> I agree with Jim that the current implementation is very useful in its
> current form.  Although there are holes, it is very useful in many
> circumstances.

Yes. However, where these holes affect the predictability of search results,
we need to close them.

> Given the lack of work going on with this specification, I do not

As a matter of fact, I've seen little progress with any of the WebDAV specs
in the last months. Progress occurs when people decide to spend time to
improve the draft. The SEARCH spec has been (albeit slowly) progressing,
mainly because of feedback from Software AG (Tamino) and the Catacomb
developers. If we can get all interested parties (probably that would
include Xythos and Oracle) to concentrate on the remaining issues, then we
can probably get the spec out within a few weeks.

> believe it useful to put some of the "nice to haves" into the
> specification now.  There are a number of production systems being used
> today without extra features and I believe we should get the work that
> is done issued.

Yes. However I don't think that anybody is doing that. Almost all open
issues aren't about nice-to-have additions but about problems with the
existing functionality (for instance, the missing ability to relaibly query
on non-string typed properties, such as DAV:getcontentlength).

> I believe we should make this a Proposed standard.

...when we've closed these issues.

> Julian recently wrote that the remaining issues are marshaling and data
> typing.
>
> Let's go through the marshaling on this list quickly and get consensus
> on those issues.

That would be good.

> Data Typing.  If the Webdav WG can't get a consensus on this topic in
> all these years, I don't think DASL should be held up on it.  In the
> last week I have only seen Julian answer this email, so I don't think a
> lot of people will work on the data typing just to get DASL done.  Let's
> simply get what work has been done into an RFC.

However, without data typing the grammer simply is incomplete. Could you
please explain how you're going to query for resources where
DAV:getcontentlength is between 1000 and 5000 without somehow introducing a
distinction between strings and numbers. And what about queries on dates? If
the spec doesn't answer how queries can be made against the basic RFC2518
live properties, there's no point in submitting it. This is currently
covered in section 5.10. The remaining questions are summarized here:
<http://greenbytes.de/tech/webdav/draft-reschke-webdav-search-latest.html#rf
c.issue.typed-literal>).

> I believe a NEW WG can be made to improve on DASL and bring together web
> searching, XPath, data typing etc. for not only Webdav resources but
> others as well (can you say web service for Google, Verity, etc).



Julian

Received on Saturday, 4 October 2003 05:59:29 UTC