In message 
<[log in to unmask]
.uk>, "Ridge, Mia" <[log in to unmask]> writes
>
>That's pretty much what I was thinking about it this morning.  Project
>and staff time is always tight, and even if management had heard of the
>latest *mumble* *acronym* *mumble*, we (as techs) haven't yet sold them
>on the benefit, and there's not an existing visible or vocal community
>of users out there demanding it.

Rather like SPECTRUM compliance, this is an area where it might be 
reasonable to look to software providers to put a framework in place in 
the first instance.  (Though we should possibly bear in mind the OAI 
situation as described on this list: several vendors offering this 
support, but apparently very few museums actually using it.)

>Hooray!  I also think it's important that the sector plan and do this
>properly - URIs should obviously be implementation- and
>technology-neutral, but ideally they should also be able to survive the
>change of name of the local authority that runs museum X, or the
>re-branding of a family of museums.

We have wise words on this from Father Tim (from 1998!):

http://www.w3.org/Provider/Style/URI

containing an entertaining list of "dog ate my homework" excuses for 
changing URLs, and some more specific (and recent) advice on techniques 
for achieving stable URIs which are suitable for the Semantic Web:

http://www.dfki.uni-kl.de/~sauermann/2006/11/cooluris/

I don't think the museum sector needs to do anything special about its 
URIs beyond putting this advice into practice.  And it comes down to 
individual institutions (or, specifically, domain name holders) - I 
don't see a need for the Collections Trust to get involved, beyond 
providing advice and encouragement.

Conversely, I do think there is scope for the museum community to inject 
its vision for data structuring into this evolving Semantic Web effort. 
If you look at a typical dbPedia entry for a museum object, e.g.:

http://dbpedia.org/page/Bedroom_in_Arles

you'll get an idea of the types of information ("Properties") which have 
been auto-extracted from the Wikipedia page.  These include ambiguous 
concepts such as "p:city".  Properties are picked up from whatever 
namespace is handy, making e.g. "foaf" a major provider of dbPedia 
properties.

Something the Collections Trust and/or CIDOC could do is to publish a 
set of properties for "museum" information which could, at least, be 
used by museums themselves.  Once in the public domain, these properties 
could also be declared to be equivalent to the properties which are 
currently being used (rather like the owl:sameAs reference in this 
example).

This page serves as a nice example of how the "web of data" works: if 
you click on any of the links, you go to the RDF record for the selected 
resource. Software agents can do the same; and when they arrive at the 
destination "page" they can process it reliably, allowing both 
navigation and querying of these RDF resources.

If museums are to play an effective part in this game, one thing they 
will need to accept is that they must use existing URIs wherever 
possible, rather than inventing their own.  This applies most obviously 
to concepts such as places (see http://www.geonames.org/ontology/) and 
people.

Richard
-- 
Richard Light
XML/XSLT and Museum Information Consultancy
[log in to unmask]

**************************************************
For mcg information and to manage your subscription to the list, visit the website at http://www.museumscomputergroup.org.uk
**************************************************