Yes, #2 and #4 are quite related in that they both concern the
abbreviation mechanism for URIs and might be considered alternative
proposals.
On the other hand, on #4, you are opening the gate to independent
entities (be them organizations or individuals) to define the prefixes
they would be using for their pages' metadata: why don't apply this to
#2 as well? IMO, it would be more important for #2 than for #4; since
#4 only provides syntax sugar while #2 enables something that would be
undoable without it (mapping Microdata to arbitrary RDF).
Yes, the idea of distributing the registration could be applied to #2.
About #1, I'm not sure about what you are exacly proposing, so I can't
provide much feedback on it. Maybe you could make it a bit clearer:
are you proposing any specific change to the spec? If so, what would
be the change? If now, what are you proposing then?
Removing the about property, showing how id can be used in this way, and
changing the description of how you transform an HTML5 document to RDF.
Finally, about #3 I'm not familiar with the OWL vocabulary, so I can't
say too much about it. But if your second proposal gets into the spec,
then this would become just syntax sugar, since any property from any
existing RDF vocabulary could be expressed; and if #4 also got in, the
benefit of "built-in" properties would be minimal compared to using a
reasonably short prefix (such as "owl:").
I agree... I'm personally not so attached to reverse domain names, but I
might have missed a lot of the previous discussions on why they are good
to have.
In any case, my intention was to get the discussion restarted around
these issues: it seems to me there was a lot of discussion at the very
beginning on microdata vs. RDFa when microdata was first proposed, but
then the discussion died without necessarily finding the best solution
(for my taste).
Cheers,
Peter
