W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 2003

RE: duplicate properties (was rfc2518bis DAV DTD ...)

From: Julian Reschke <julian.reschke@gmx.de>
Date: Sat, 18 Oct 2003 01:55:52 +0200
To: "WebDAV Discussion List" <w3c-dist-auth@w3.org>
Message-ID: <JIEGINCHMLABHJBIGKBCIEIAINAA.julian.reschke@gmx.de>

> From: w3c-dist-auth-request@w3.org
> [mailto:w3c-dist-auth-request@w3.org]On Behalf Of Dennis E. Hamilton
> Sent: Friday, October 17, 2003 11:48 PM
> To: WebDAV Discussion List
> Subject: duplicate properties (was rfc2518bis DAV DTD ...)
>
>
>
> This note seems to have gone into a hole somewhere [it is not on
> the archive] so I am resending it for completeness.  Now if I
> could figure out why I receive two of everything, that would be a
> good thing. -- dh

People pressing "reply to all"...

> Julian,
>
> Thanks for the comment on the example.  I did that intentionally,
> because there is an existing example (8.2.1 in bis-04) that
> repeats a property, in effect, but inside of a wrapper, and I
> wanted to see what attention this one would attract.

Could you be more specific? I don't see a duplicate property name there.

> (There is also an error in my <error> element example.  I left
> out the "/" before ">" of the empty element.  It should be
> <forbid-external-entities />.)
>
> There is an issue that will arise with the unique property
> requirement.  Other uses of ad hoc properties in XML applications
> are not limited to single occurrences of property elements for
> the same property.  (Property-asserting elements are seen as
> predicates or establishment of a relationship, and many-to-many
> is commonplace.)  The example I gave would be appropriate if I
> was mapping something from RDF to DAV, for example.  This may
> become a problem as the Semantic Web and DAV collide.

Well, that's how WebDAV works. If you need multi-valued properties (such as
when mapping DC to WebDAV properties), you need to come up with a
WebDAV-compatible way to do it. See for instance:

<http://ftp.ics.uci.edu/pub/ietf/webdav/dc/draft-ietf-webdav-dublin-core-01.
txt>

> I assume that the appropriate DAV treatment would be to wrap a
> bundle of RDF properties inside a single DAV property and pray.
> This is a new topic and probably becomes something to identify as

No, it's absolutely not new. See above.

In fact, our server already supports extraction of DC-encoded metadata from
HTML content and mapping to WebDAV properties. If there are more people
interested on a common format, please email me.

> an open/future item without resolution in 2518bis.  It is part of
> the overall question of relationship to other applicable
> standards, such as XML digital signatures and encryption, XML
> Schema, SOAP/XML-HTTP, UDDI, RDF-XML, and other Web Service and
> Semantic Web specifications that may have some bearing on DAV
> applications.
>
> -- Dennis
>
> Side Note: From my background with DMS and especially DMA, I
> would find it natural to prohibit duplicate assertions about the
> same property.
>
> Having worked with other XML-based specifications for a while, I
> see that is not the course being taken.  What I learn from that
> is that I have been subjecting the protocol to the constraints of
> an envisioned storage model: I am using ideas about
> storage-system implementation to place tacit constraints on the
> request/response protocol for which there is no abstract
> justification.  That's a lesson for me.
>
> It's apparently a lesson in W3C work too.  Later specifications
> are being abstracted above the XML InfoSet and even above XML
> altogether.  The difference in specification models between SOAP
> 1.1 and SOAP 1.2 is informative in that regard.  The same arises
> in the Working Documents for revisions of RDF and in other
> Semantic Web specifications (e.g., OWL).  I don't know what the
> lesson might be for DAV. -- dh

At the end of the day, that's just an aspect of marshalling in WebDAV. As
properties can have arbitrary XML content, there's no problem mapping these
multi-valued properties to WebDAV. The only things that's missing is a
common format. So far, the interest in agreeing on one seems to be little.


Julian

--
<green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760
Received on Friday, 17 October 2003 19:56:17 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:44:05 GMT