Re: Why DASL need not be DAV-specific

Sounds ok to me. So the question is, does this result in any changes to the
charter? Your two statements below look like stuff for the requirements
document to me.

Also, those two queries look identical. Please keep in mind that DAV has a
totally open ended property framework, so <d:dma-guid> is a property just
like any other.

Alex Hopmann
Microsoft Corporation

Jim Davis wrote:
>At 01:56 PM 3/2/98 PST, Yaron Goland wrote:
>
>> The reason why this working group was formed was to solve search for DAV.
>> ... to provide search facilities which make optimal use of the DAV object
>> model, repository model, variant model, access control model, and
versioning
>> model...Universal Search is a laudable goal but one that will require
>> compromises we are not willing to make.
>
>Perhaps we could all agree as follows:
>
>1) DASL MUST fully support WebDAV repositories efficiently.
>2) DASL MAY also support other HTTP applications, if such support imposes
>only minor costs (e.g. in added complexity.)
>
>To elaborate on the second point, at present, it seems to me the only DAV
>feature that DASL requires is a property store.  Given that we're using XML
>to express queries, it would be very easy to generalize the notion of
>'property' to support  property stores other than that of DAV.  Surely this
>is not objectionable?
>
>We're talking about the difference between (to use an utterly fabricated
>query syntax)
>
><D:contains>
> <D:prop> <X:author/></D:prop>
> <D:value>Knuth</D:value>
>
>and
>
><D:contains>
> <D:prop><d:dma-guid>123762-72738</d:dma-guid></D:prop>
> <D:value>Knuth</D:value>
>
>Something like that.
>
>As I noted in earlier email, it will also require means to specify query
>scope, which may be a bit trickier to generalize, but not too hard.
>
>It may well turn out that when we get into details, there will be something
>about DAV that will be too hard to generalize, or something about e.g. LDAP
>or Z39.50 that would impose a great penalty on DAV.  If either of these
>happen, we should discard the attempt to be universal.
>
>Is this an acceptable framing of the charter?
>
>In the meantime, we can scarcely design DASL's interactions with the
>variant, access control, or versioning models of DAV, before these are
>designed.  I at least haven't seen any sign of progress on any of these
>fronts.  Can you provide any guidance?
>
>Jim
>
>
>------------------------------------
>http://www.parc.xerox.com/jdavis/
>650-812-4301
>

Received on Monday, 2 March 1998 20:33:28 UTC