Re: microdata feedback

Manu Sporny wrote:
> 1. Write a Microdata parser from scratch following the exact HTML5 spec
>    language.

http://philip.html5.org/demos/microdata/demo.html (JS) and 
http://james.html5.org/microdata/ (Python) were written that way. (The 
spec language may have changed a bit since then, but I don't remember 
any substantial changes). Are these sufficient for what you're looking for?

> 2. Mark up the entire RDFa Test Suite in Microdata (see which test
>    cases cannot be marked up).

http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-May/019732.html 
and 
http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2009-May/019744.html 
suggest that it's possible to (inelegantly) express any arbitrary RDF 
graph using Microdata, as long as all literals are strings, and there 
are no cycles, and no blank node is used as the subject of two or more 
triples. So that should determine the set of test cases that cannot be 
marked up in a fashion that gives the expected triples. (I might have 
missed some other issues, though.)

-- 
Philip Taylor
pjt47@cam.ac.uk

Received on Thursday, 13 August 2009 15:56:19 UTC