sigh. Thanks for the explanations, even if they fail to increase my joy.
Seems like we'd need some reasoning that knows that ranges & domains of
something are an information resource, not an information resource, or
either. With a bit of logic that says that statements on an information
resource which can't be on an information resource should be treated as
applying to their primary-topic rather than be disregarded.
I get a headache trying to work out the details, though. dc:creator can
certainly refer to a real world thing (eg. the statue of David).
Working out that the following refers to two different entities is
pretty clever...
<http://totl.net/>
foaf:member<http://users.ecs.soton.ac.uk/cjg/>;
dc:license<...>;
I can then detect the problem, as it implies the URI is both a document
and a group and things can't be both. I can detect the issue, but then
what? There'll be lots of situation where the ambiguity is absolute. eg.
<http://en.wikipedia.org/wiki/David_%28Michelangelo%29>
dc:creator<http://en.wikipedia.org/wiki/Michelangelo> .
Did he make the statue or the webpage?
I think the other common conflation we'll get is between places & legal
entities. eg. "The Royal Society", has both members and a lat/long and
people will naturally muddle them, but we'll need a way to unpick that.
(I assume)
On 13/06/11 14:41, Richard Cyganiak wrote:
On 13 Jun 2011, at 13:02, Christopher Gutteridge wrote:
Option one is to read that as
<http://totl.net/> foaf:member<http://<http://users.ecs.soton.ac.uk/cjg/> .
which is not true
Whether it's true or not is up to the URI owner, because they get to decide
what http://totl.net/ identifies.
and prevents us making any statements about the documents directly (ie.
license, creator, last modified)
How so? Look:
<http://totl.net/>
foaf:member<http://users.ecs.soton.ac.uk/cjg/>;
dc:license<...>;
dc:creator<...>;
dc:modified "..."^^xsd:date.
Web pages don't have members and groups of people don't have licenses. Anyone
with a minimum of intelligence -- human or machine -- can work that out.
What might work better is if you have new predicates which explicitly means<the
primary topic of this document has a member who is the primary topic of...>
That would work, and I've used that pattern in the past [1], but try explaining
that to someone outside this mailing list or writing it down in JSON.
This is more or less what schema.org seems to be doing, if I've understood
correctly...
No, they just don't give a damn that the same URI ends up being used for a
document and a thing.
Best,
Richard
[1] http://vocab.sindice.com/xfn
--
Christopher Gutteridge -- http://id.ecs.soton.ac.uk/person/1248
/ Lead Developer, EPrints Project, http://eprints.org/
/ Web Projects Manager, ECS, University of Southampton,
http://www.ecs.soton.ac.uk/
/ Webmaster, Web Science Trust, http://www.webscience.org/
