FWIW, I agree with Rob that we need the capability to handle large records
if SRW is going to make it outside of the libraries community. Z39.50 made
some attempt to accomodate this in Espec-1, but the mechanism was
ultimately too complex and too married to GRS-1 to take off. The ability to
carry a server-defined subset of XPATH seems in my eyes to be the perfect
solution. Ralph doesn't have to implement it (if SRW may only contain
things that everybody agrees to implement, then it truly becomes a lowest
common denominator, and a dead duck. IR is much too complex an application
to be bound in like that), and Matthew doesn't have to use off-the shelf
XPATH processors if he's scared of security issues. Indeed, I would predict
that profiles over SRW will name explicit Xpath expressions valid in a
given application area, and simple applications may support that using
lookup tables without any intelligent processing at all.

I wholeheartedly support the ability to use Xpath as an element selection
mechanism. In fact, I'm still keen to see Xpath introduced as an element
selection mechanism in search expressions as well, as an alternative to the
flat lists of elements, which are basically the bastard children of
attribute combos and unlikely to resonate with someone raised on XML tools.

--Sebastian

At 14:38 23-07-2003 +0100, Mike Taylor wrote:
> > Date: Wed, 23 Jul 2003 14:18:16 +0100
> > From: Robert Sanderson <[log in to unmask]>
> >
> > > I'm more likely to return the Mods record for the item with a link
> > > to the data, than the data intact (or at least support
> >
> > Very 1990s, don't you think?  Can't handle the data properly so
> > ... I'll just give you a link to it and hope that the client can
> > somehow deal with it.  Why don't we do this for all of our records?
> > That would make life much easier if the only contents of recordData
> > was a URL to the record.
>
>:-)
>
> > The world has moved on from this model, IMO, and not being able to
> > gracefully cope with large records is going to be an increasingly
> > significant failing.  -Everything- is going XML, for better or
> > worse, and XML isn't a compact data format.  Having to return only
> > short metadata records is, again IMO, an extremely short sighted
> > decision that will hinder take up worse than a single, optional,
> > parameter that references a well defined, well understood and
> > broadly implemented standard.
>
>I couldn't have said it better myself.
>
> > But, I've registered my proposal, people apart from Mike think it's
> > somehow too complex even though it's expainable in a single
> > sentence, so I'll just have to assign namespaces for all the
> > different schemas I need.
>
>Actually, I think that the set of "people apart from Mike" who've
>expressed that opinion currently consists only of Matthew.  I'd like
>to hear some more opinions before we throw in this particular towel.
>
>  _/|_    _______________________________________________________________
>/o ) \/  Mike Taylor  <[log in to unmask]>  http://www.miketaylor.org.uk
>)_v__/\  "I wish you all happiness and harmony, and that you will
>          all find something you can agree on, however small" --
>          Robin Reed's farewell message to the Dinosaur Mailing List.
>
>--
>Listen to my wife's new CD of kids' music, _Child's Play_, at
>         http://www.pipedreaming.org.uk/childsplay/

--
Sebastian Hammer, Index Data <http://www.indexdata.dk/>
Ph: +45 3341 0100, Fax: +45 3341 0101