RE: JW11: single scope

Jim Whitehead:

"If I want to do a search only on the src/module2/ and src/module3/
hierarchies, I would have to submit two DASL requests in the current
proposal."

Yes, that's correct. This will probably be generalized on
the next release of DASL. The famous criterion "is it useful
without this feature" applied to this situation resulted in
limiting the scope list to one element.

Your other comments below don't lead me to want to revise this
version of the protocol draft.

Alan Babich

-----Original Message-----
From: Jim Whitehead [mailto:ejw@ics.uci.edu]
Sent: Tuesday, July 20, 1999 2:52 PM
To: Babich, Alan; 'DASL'
Subject: RE: JW11: single scope



> If I remember correctly, we deliberately decided to limit the
> scope list to one element for the first version to avoid defining and
> debating three valued elimination -- the obvious straightforward extension
> of three valued logic used in other industry standards to handle
> cross-repository queries.

I was thinking more in terms of performing a query on, say, multiple
hierarchies on the same server, making the assumption that all of the
hierarchies contained resources with the same properties.

For example, if I have some resources set up as follows:

        src/
        src/module1/
        src/module2/
        src/module3/

If I want to do a search only on the src/module2/ and src/module3/
hierarchies, I would have to submit two DASL requests in the current
proposal.

> That is necessary, because the query
> schemas of multiple repositories don't have to be identical, and
> the issue of handling queries that are partially defined arises
> immediately.

I agree, but I don't see a query directed to a single DASL arbiter running
into this problem.  Perhaps if a DASL search arbiter were sitting on top of
a DMA interface this would occur...

> First issue: Do you want your query to be defined on the intersection
> of the query schemas, or their union? That must be specified.
> The intersection may be too small to be useful, especially for
> repositories that don't implement QSD. The union case must
> obviously handle queries that are only partially defined on its
> constituent repositories.

Since all DAV resources have to support a known set of predefined
properties, doing a search on the intersection can be quite useful.

- Jim

Received on Wednesday, 21 July 1999 15:55:49 UTC