Re: @itemid and URL properties in schema.org

On Wed, Nov 9, 2011 at 8:26 AM, Dan Brickley <danbri@danbri.org> wrote:

> On 9 November 2011 14:14, Toby Inkster <tai@g5n.co.uk> wrote:
> > On Sat, 5 Nov 2011 10:43:45 -0400
> > Gregg Kellogg <gregg@kellogg-assoc.com> wrote:
> >
> >> Alternatively, schema:url could be a subProperty of owl:sameAs, which
> >> after a fair amount of reasoning can yield what you want (in RDF that
> >> is). If schema:url really is used to designate the subject, much
> >> better to use @itemid in the first place.
> >
> > Based on the examples at schema.org (for instance, the example at the
> > end of http://schema.org/Person), schema:url is basically the same as
> > foaf:page or rdfs:seeAlso, not owl:sameAs. Perhaps even foaf:homepage
> > (which is an owl:IFP, and thus an indirect identifier).
>
> Or foaf:isPrimaryTopicOf, even? If the thing pointed to is a doc
> primarily about the thing.
>
> This seems to be a variant of the eternal perma-thread of RDF:
> deciding where and how we care to distinguish between identifiers for
> things versus more conventional pages that describe those things. In
> it's hash-versus-slash guise, it had people advocating (largely
> unsuccessfully) for Web server reconfiguration to send http 303
> redirects.
>
> As an online discussion grows longer, the probability of someone
> mentioning URNs, "URLs not being for identifying real things",
> map-vs-territory and http-range-14 approaches 1.
>
> All that aside, it's a pretty common pattern to identify things
> indirectly via pages about them; the question generally is how
> explicitly we handle the indirection / proxying.
>
> http://www.imdb.com/title/tt0083658/ is my favourite movie, etc.
>
> Maybe I should've said http://en.wikipedia.org/wiki/Blade_Runner ? Or
> searched around and found http://dbpedia.org/page/Blade_Runner ? Or
> realised that this was a 303 redirect from the real deal,
> http://dbpedia.org/resource/Blade_Runner ...?
>
> We need markup idioms that allow for people who don't have the time to
> chase down "URIs for the real world thing" every time. Maybe
> schema:url is for that; I don't know, I wasn't there at the
> beginning...
>

Are you including @itemid with schema:url here? In any case, I agree that
HTML authors should not have to worry about these subtleties, which ideally
should be pushed to the RDF processors. Surely humans can tell
that birthDate relates to the real human described in the page of type
http://schema.org/Person, but machines would need a hint to infer
that. Maybe vocabulary authors could state whether a property relates to
the page or the primary topic of the page, and RDF processors could use
that info? [1].

One caveat of the primaryTopic approach though is that you need a page for
each resource, which might or might not be a problem depending on the
application (Drupal for example tend to expose a page for most of its data
items). One could work around the need for a page for everything by
describing multiple items in the same document and using hashes, but they
are error prone and not as friendly to extract from a page for copy-pasting.

Steph.

[1] http://lists.w3.org/Archives/Public/www-tag/2011Oct/0099.html


>
> Dan
>
>