W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 2001

Re: DASL - who wants to use it? (requirements spec?)

From: Alan Kent <ajk@mds.rmit.edu.au>
Date: Tue, 18 Dec 2001 16:53:21 +1100
To: WebDAV <w3c-dist-auth@w3.org>
Message-ID: <20011218165321.I24222@io.mds.rmit.edu.au>
On Mon, Dec 17, 2001 at 04:28:59PM -0800, Lisa Dusseault wrote:
> I was looking at the same draft.  But I think Alan wanted to know if DASL
> addressed body content searching at all, which it does.  If the server
> supports it, the client can use it (as usual).
> 
> lisa

This may be a bit of a can of worms, but is there any interest in using
the structure of content in queries? For example, being able to query
on the content's "Title" or "Author". I am thinking of Dublin Core
metadata sort of approach - not XPath or XML Query since the content
can be anything (PDF, GIF, etc). Allowing something like Dublin Core
would allow a level of structure to be imposed.

If this was considered useful, would the better approach be to define
additional live properites that were automatically extracted from the
content (hence its really just querying on properties after all, not
content itself), or to allow querying on properties or content.
I guess I am sort of asking should <allprops> return the Dublin Core
like elements too? (It could be expensive to compute).

I do a fair bit of work with a search and retrieval protocol, Z39.50.
It is based on the Dublin Core idea of separating your access points
for searching from the physical representation of the data. It allows
querying on "Title" as an abstract concept, without requiring that
it be possible to return a "Title" out from the data. It takes a
different view on query capabilities too - it says you have terms
that can have constraints applied. Specifying a property name to
search within is a constraint. Specifying > (greater than) is a 
constraint. It does not use 3 value logic. It more says 'return
all records that contain some data that match the condition'.
WebDAV might not have the same problem of repeating values as
properites have only one value.

Personally I would like to search on parts of the content using
Dublin Core (or other) metadata standards. Should I do this using
WebDAV properties bound to parts of the content of the resource?
Any other ideas?

Thanks
Alan

ps: I am not likely to do this any time soon. Just thinking about
the future.
Received on Tuesday, 18 December 2001 00:53:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:43:59 GMT