Manu Sporny wrote:
...
We are currently working[1] on features to dynamically extending the
base set of reserved words and the set of pre-defined prefixes through a
mechanism called RDFa Profiles[2]. It is proposed that this mechanism
would allow authors to do this in their documents:
<div profile="http://example.org/myprofile.html";> ...
<span property="description">A description for this page.</span>
<span about="#me" property="name">Manu Sporny</span>
</div>
Note that 'description' and 'name' are not prefixed, but would be mapped
to a full URI in the document listed by @profile. This allows the ease
of Microformats-like markup but with all of the rigor of RDFa.
...
So, to process the HTML fragment above, a recipient would always need to
look up the profile resource, and extract RDF triples from *that*
document (*) first?
How do you prevent the problem we've seen with DTDs (consumers not
willing or able to cache the resource, and thus re-requesting the same
resource over and over again)?
BR, Julian
(*) Which specific media types would a recipient be required to support
for profiles?
