Re: sru/cql sort

["Ray Denenberg, Library of Congress" <rden@loc.gov>]

 From: "LeVan,Ralph" <levan@oclc.org>
> Well, my original complaint about incorporating sort into the query
> grammar was about the additional complexity it added to the query.

As I recall, your compaint was that having sort as part of the query makes
it difficult to use a retrieval field (as opposed to an index)  as a sort
key.  A complaint that I find rather compelling.  I'm leaning towards Rob's
suggestion, be able to express sort both inside and outside the query.

--Ray



















But,
> our work on trying to use the URL template structure has focused my
> dissatisfaction.
>
> The problem is that we can't describe how to do a sort separate from how
> to do a search, because we've bundled them together, for no particularly
> good reason.  The clear objective in every simplification project I've
> been on is to decompose complex problems into simpler ones.  We've done
> the reverse by bundling search and sort together.  I'd like to see that
> decision reversed.
>
> Ralph
>
> > -----Original Message-----
> > From: Ray Denenberg, Library of Congress [mailto:rden@loc.gov]
> > Sent: Wednesday, October 29, 2008 6:13 PM
> > To: search-ws@lists.oasis-open.org
> > Cc: LeVan,Ralph
> > Subject: sru/cql sort
> >
> > Just looking through some of my personal archives and it seems to me
> > that we
> > never addressed Ralph's complaint about sort.
> >
> > Ralph, could you write up a brief proposal that we can add to the
> > "table"?
> >
> > By Monday for the call, if possible?
> >
> > Thanks.
> >
> > --Ray
> >
>
>