In message
<[log in to unmask]>, Mia
<[log in to unmask]> writes
>
>Then there's the whole question of whether museums should bother publishing
>individually or work to publish aggregate data. Developers are reluctant to
>learn new structures and methods for each publisher, so there are benefits
>to publishing either through one end point or with one agreed data standard,
>but there's also a bigger overhead to this.
If you're publishing URLs, you have to work within the web's system of
domain names. So part of the question is "do you keep control over your
collection's URLs by using your own domain name space, or do you rely on
an external provider to keep the relevant domain registered into the
distant future?"
If you publish individually, you can do various things to make it easier
for others to harvest your Linked Data. You can always add RDF as a
supported metadata format to your existing OAI/PMH interface. Or you
can publish a VOiD [1] description of your resource: a sort of
Collections Level Description for Linked Data. And in the near future
(if not already) I would expect to see "semantic spiders" which crawl
the web following Linked Data in exactly the same way that current
spiders crawl linked web pages.
>> Actually, this touches on point 4 below: does it really qualify as "linked
>> data"
>> if it is only "linkable" rather than actively linking to other
>> data sources?
>
>Good question - I'd say so, but I'm no expert. Anyone?
I'd say "yes". The fact that it's a URL means that you can easily use
HTTP to look it up if you want to (it's "dereferenceable" in the
jargon). So it's inherently "linkable" (to).
When we talk about "linking", all we really mean is that we can find the
same URL, standing for the same concept, in a number of places, and draw
appropriate conclusions. It's just a string match really. The
difference is between matching on "Light, Richard", which could
represent lots of people, and indeed dogs for all I know, and e.g.:
http://everywhoeverlived.org/234968574
which is just me (say). (I was hoping to find a URI for me in the OCLC
Worldcat catalogue, but only books are worthy of URLs, it seems. [2])
Richard
[1] http://semanticweb.org/wiki/VoiD
[2] http://www.worldcat.org/search?qt=worldcat_org_all&q=richard+b+light
--
Richard Light
****************************************************************
For mcg information visit the mcg website at
http://museumscomputergroup.org.uk/
To manage your subscription to this email list visit
http://museumscomputergroup.org.uk/email-list/
****************************************************************
|