Cheers Bill,
glad I got the right list & thanks for the reply, inline comments from
here ..
1) That's a good question. Ultimately it's up to you. In the semantic
web, there is no single "correct" URI for a thing. In many cases it
will make sense to use an existing commonly used URI, such as the
dbpedia one in your example.
makes sense, the idea of needing a single GUID is hard to get out of
ones brain and in all honesty I've spent the last 3 days doing nothing
but reading vast amounts of very technical documentation on all of this,
it'll sink in I'm sure; quite possibly time for me to stop reading and
get doing :)
2) The philosophy of RDF is that it is decentralised, so there is no
central authority to define what might be an authoritative URI for a
thing. Although it means less certainty, ultimately the decentralised
approach should be better for scalability.
makes sense, again practise will probably answer a lot of my questions
about this, especially the immediate one of "how do I query the entire
cloud of linked data, and what happens when X servers are down, won't it
mean I'll need to cache everything I /might/ need" - a feeling that's a
whole other subject which is already discussed at length around the net
though.
3) sameAs indicates that two URIs refer to the same thing (ie they are
two different identifers for the same thing - don't mix it up with URIs
of web pages that have information about that thing) - and this allows
you to infer that if A owl:sameAs B then any statement about A is also
true about B.
Indeed the mix-up between URI's as identifiers for *anything* and
expecting there to be something at said URI keeps plaguing me; because
although I realise it's an identifier, a part of my brain assumes there
is information about said thing located at URI endpoint, and further
that even if not then it's simply placeholder endpoint at which info
will eventually be located.
Further from an application programming perspective it will be
programmed in to many (if not all) apps to http request the URI for more
information and thus some response will be "assumed"?
3a) In web jargon, I think you might be mixing up 'information' and
'non-information' resources. What I mean is that you are mixing up a
URI as an identifier for a thing, and a URI for a document on the web
which happens to hold a bunch of RDF. I don't think sameAs is really
relevant for the mirror example you presented.
Concurred, I guess I'm thinking of a few months down the line when some
kind of rdfa editor is embedded in a big cms (say drupal 7.x) and
average joe blogger comes along, visits URI identifier
http://domain.com/some/thing which then 30*'s them to a browser friendly
bit of info about URI (identifier) - to which average joe copies the new
URI (for the resource / description) and sticks it in the rdfa editor.
Assuming this would lead to an incorrect triple, also assuming that
times that by say 3 triples per post, 10 posts per blog, many thousands
of blogs and you've got a big problem (or at least a big waste?)
I'll have to stop there for now - hopefully you'll get further responses
to address your other questions.
Hope that helps
Certainly does, many thanks for the quick response!
Nathan
