Re: CfC: Request transition of HTML Microdata to Candidate Recommendation


I am an implementer who works with both RDFa and microdata. I am a
co-maintainer of Drupal core's RDF module (which publishes RDFa) and the
maintainer of Drupal's Microdata module.

I believe that microdata should move forward to REC. Some people assert
that there isn't a difference when using RDFa Lite 1.1 and microdata. I

While RDFa Lite did adopt most of microdata's processing model, it has a
wildly different processing algorithm. It uses the full RDFa Core 1.1
algorithm, which is extremely difficult to understand. And there isn't an
explicit distinction between publishing full and lite... if you
accidentally throw a rel in there, you've switched from lite to full. This
makes it hard for publishers to debug their own data, as even the widely
used distillers disagree about how such RDFa should be parsed. For example,
if you parse this test
distiller <> (written by the new RDFa
Test Suite author) vs the W3C's distiller
<>(written by one of the RDFa 1.1
editors), you get different results. This
isn't a theoretical issue, it's an actual issue posted last
week<>to the issue queue.
This puts a strain on data publishers and
the tool authors (like me) who support them.

Microdata's processing algorithm is vastly simpler, which makes the data
extracted more reliable and, when something does go wrong, makes it easier
for 1) users to debug their own data, and 2) easier for me to debug it if
they can't figure it out on their own.

But RDFa does have its own strengths for particular use cases, such as
those that require explicit datatyping or XML literal properties. I think
RDFa should continue as a REC for those more advanced use cases, and
microdata should be published as a REC for those which don't require the


Lin Clark
Drupal Consultant

Received on Wednesday, 28 November 2012 18:49:21 UTC