Re: non-prefixed @rel

Shane McCarron wrote:
> 
> I think that instanceof has NO non-scoped values.  It takes curies, and
> there are no XHTML-defined nor RDFa-Syntax defined values that would be
> in the xhtml/vocab# space.

Right, I should be clear that I don't mean @instanceof="next" to yield
xh:next, that's obviously not a good idea.

It's less clear what to do with @property="next", though since we are
pre-processing for specific backwards-compatible values of @rel and
@rev, I think we can safely say that @property="next" also doesn't
generate anything.

So I'll restate the proposal accordingly:

=========
rel="X"

  (a) if X is a XHTML 1.1 reserved value, it is re-written in a
pre-processing stage as rel="xh:X", where xh maps to
http://www.w3.org/1999/xhtml/vocab#.

  (b) if X is a prefixed value, it is processed normally.

  (c) if X is a non-prefixed, non-reserved value, it is ignored by any
baseline RDFa parser, though other conformant parsers may choose to
generate a triple *outside* of the default graph, as per [RDFa Parser
Conformance].

The same rules apply to each value in a space-separated set of values for X.

The same rules apply to @rev, which has the same backwards compatibility
issues as @rel.

For @property and @instanceof, only rules (b) and (c) apply.
=========

-Ben