Re: RDFa Lite 1.1 Conformance Section - host language attributes (ISSUE-136)

On Mon, Apr 23, 2012 at 11:59 PM, Ivan Herman <ivan@w3.org> wrote:
> Alex,
>
> this time I think I do not fully agree with you...
>
> Using @rel/@rev would push the source out of RDFa 1.1 Lite. Ie, that should not be allowed. I realize that @rel _may_ be used in HTML5, and that creates an additional issue which Stéphane just noted:
>
> https://www.w3.org/2010/02/rdfa/track/issues/135
> http://lists.w3.org/Archives/Public/public-rdfa-wg/2012Apr/0073.html

There are a vast majority of HTML documents that use @rel attributes
on link and anchor elements that do not have RDFa attributes.  If an
author adds RDFa Lite, those documents, as specified wouldn't be
considered conformant.  RDFa "borrows" the @rel and @rev attributes
from HTML and makes them more pervasive.  As such, I would suggest
that we allow a host language to include them given that they already
exist and have been used for a long time in HTML.

>
> But, if we go along option #1 in that proposal, a value of @rel with only predefined HTML5 value is immaterial from RDFa's point of view.

Well, I personally rely upon @rel with predefined HTML5 values to
produce relations between the current document and the target of the
link regardless of whether it is RDFa Lite or not.  As such, I still
think the conflict is in RDFa in step 11.  We have a dual use of the
@property attribute that has unintended consequences in HTML.

Meanwhile, option #1 doesn't address the existence of the @rel and
@rev attributes in HTML.  The conformance clause would have to address
the existence of these attributes.

Also, to implement option #1, we'd have to disallow generation of
triples for certain values.  We don't have anything in the algorithm
nor in the XHTML+RDFa 1.1 specification that does this.  We'd then
have to change how terms are processed and allow a list of disallowed
values to be specified in the context.  I don't find that a pleasant
solution.  Also, we'd have to specify in Step 11 that if the @rel/@rev
attributes resulted in no triples, treat them as if they didn't exist.
 We don't have language like that as of right now.


>
> _My_ proposal would be to amend that paragraph as follows:
>
> [[[
> It must not use any additional RDFa attributes other than vocab, typeof, property, resource, and prefix; it may also use href and src, in case the Host Language authorizes their usage.
> ]]]
>

That still makes HTML documents non-conformant when they use the @rel
attribute, as they are likely to do so.

-- 
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics

Received on Tuesday, 24 April 2012 16:08:45 UTC