RE: Proposed resolution for issue "score-pseudo-property"

Elias Anderson wrote:
> I agree that the spec should mention something about the intended
> interaction between DAV:nresults and DAV:score. The most straightforward
> approach would be to state that the server MUST follow the expected
> behavior. As much as I'd like to move the sorting cycles from the server
> to the client, there doesn't seem to be an obvious way to do this while
> still allowing truncated result sets. The fact that a server may
> truncate the result set of a SEARCH, without the client explicitly
> asking it to do so, leads me to believe that sorting is an inherently
> server-side activity.

Thinking about it, the relationship *really* is between DAV:nresults and the
presence of DAV:score -- not the fact that the client requests ordering by
DAV:score. Is there a use case for combining DAV:nresults with ordering by
DAV:score in *ascending* order (lowest scores first)?

If not, we can just point out that combining any operator that produces
scores (right now thats DAV:contains) with DAV:results should cause the
server to return the "n" top scoring results.

--
<green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760

Received on Saturday, 30 November 2002 03:41:42 UTC