On Fri, 31 Jul 1998, Brian Kelly wrote:
> Hi Matthew
>
> > ... Within RDF
> > it is hoped to use this sort of construct to enable a standardise
> > protocol for searching for information on the web. At this point it
> > begins to overlap with some of the objectives of Z39.50 (e.g. searching
> > catalogue level descriptions or EAD records).
>
> Where is this happening?
ILRT Towers, 8 Woodland Road, University of Bristol, Bristol BS8 1TQ. ;-)
> At the W3C AC meeting I asked various W3C people
> if there was any work planned for developing distributed searching protocols
> (e.g. an RDF application in HTTP/NG). I was told that this was possible but
> that W3C wasn't doing any work in this area. So who is doing the work?
The W3C activity in this area hasn't been chartered yet, but it is not
revealing any secrets to say that improving resource discovery (searching
AND browsing) is high on the list of reasons for having built a Resource
Description Framework in the first place.
The Z39.50 ZIG minutes from January 1998 are quite revealing (and
realistic?) in this context: their conclusion seems to be that Z39.50 will
remain a mostly library-specific protocol and there is an expectation that
the museum/archives community will bail out of Z in favour of a
Web-oriented solution as soon as RDF is looking useable. (Did someone
mention the word Legacy...?)
In the absence as yet of an official W3C RDF Services activity, I'm
cobbling together a few RDF Searching proposals based on requirements from
ROADS / WHOIS++ experience, from talking to the IMS people about their
CORBA/XMLish architecture and on what I want to get built under the
DESIRE2 project to support thesaurus-servers and distributed browsing.
Don't hold your breath though, don't believe the hype and wait till my
demo can run for a day without crashing!
FWIW the distinction between 'collections level' and 'normal(?)' resource
description floating around on this list is a little bogus. All the
eLib ANR Subject Based Gateways have quietly been doing large
scale "collection level" resource description for 2-3 years, so it's not
as if this is some strange new problem phase-3 has run into.
The only major difference is that many of the collections catalogued by
the Subject Gateway network lack machine interfaces (Z39.50, WHOIS++
etc), with the associated properties (ports, protocols, access rules) that
these things bring.
I mailed a proposal to the ROADS list a month or so ago about augmenting
the ROADS Service template type to include clusters of information
relating to machine interfaces, simple access-permissions stuff. We had a
little useful discussion but it might be worth migrating the proposal to
the wider eLib list, since I'm not sure how many eLib3 people are on
open-roads. Seems like there's a danger of re-inventing the wheel here --
it took long enough to agree the cataloguing fields for the eLib ANR
gateways to describe collections of various types, so it's pretty
important we don't roll back from this and begin again from scratch.
We call it 'phase 3' because it builds on the previous two, right? ;-)
Dan
--
[log in to unmask]
Research and Development Unit tel: +44(0)117 9288478
Institute for Learning and Research Technology http://www.ilrt.bris.ac.uk/
University of Bristol, Bristol BS8 1TN, UK. fax: +44(0)117 9288473
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|