Re: Dumb Question

Hi Shane,

Fair points, all.

So what about coming at this a different way?

>  ...I know that we don't like having people duplicate things...

If we decide to go the route that I've been suggesting, where we have
RDFa-core and then a language-specific 'interpretation', then we could
avoid the question of attribute duplication altogether; if we consider
something to be important then it needs to go into the core attribute
list, because those are our core 'concepts'. If that feature is then
duplicated when RDFa-core is hosted by another language, it doesn't
matter, because in many ways duplicate attributes is what it's all
about; the host language 'interpretation' is all about saying, 'given
this language and this mark-up, what could we reasonably say it
means'.

We might therefore consider that representing rdf:type is of such
fundamental usefulness that RDFa-core needs to have an attribute to do
it. That would be independent of any host language that might
additionally define one of its attributes to be a mapping for
rdf:type, just in the same way that HTML/XHTML defines @rel to be a
mapping for @property when used with @href (which is itself a mapping
for @resource).

As you can see, what I'm getting at is that we probably need an
attribute for 'type' indication that is independent of any decision we
make about @role or @class.

Now, I happen to disagree that we should avoid defining a mapping for
@class, but if we put a new attribute into RDFa-core to represent
rdf:type then we can fight the great 'what does class mean in HTML'
battle another day, since we'd have achieved the most important
requirement which is to have a way of marking up RSS feeds
efficiently. :)

What do you think? And if you like the idea of putting the attribute
into RDFa-core, what shall we call it? Naming attributes...we always
seem to have a problem with that. But given that @type is bound to
clash in just about any host language, we might have to think
laterally...what about @isa, which I think is the N3 way of doing it?:

  <a>
    <b about="s" isa="foaf:Person" />
  </a>

etc.

Also, now I think about it, in the modern RDF-world, is owl:sameAs
just as important as rdf:type? Whilst we're looking at RDFa-core,
should we also be thinking along these lines:

  <a>
    <b about="[a:b]" sameAs="foaf:Person" />
  </a>

Any thoughts from RDF bods?

Regards,

Mark

-- 
  Mark Birbeck, formsPlayer

  mark.birbeck@x-port.net | +44 (0) 20 7689 9232
  http://www.formsPlayer.com | http://internet-apps.blogspot.com

  standards. innovation.

Received on Thursday, 31 May 2007 14:08:45 UTC