Build the semantic web, certainly -- but using which standard?
For practical reasons, schema.org is an excellent choice: it simply works with the major search engines (Google, Bing, Yahoo, Yandex). But at first schema.org supported only one semantic markup style (microdata) -- relatively easy to use, but, in some people's opinion, not as powerful as RDFa.
That limitation is now disappearing. In an official schema.org blogpost on Friday, Dan Brickley announced that schema.org will now be adopting a simplified form of RDFa:
As a result of our continued discussions and collaborations with publishers, implementers and standards-makers, we're pleased to give advance notice of a new way of adopting schema.org's structured data vocabulary. W3C's RDF Web Applications group are right now putting the finishing touches to the latest version of the RDFa standard. This work opens up new possibilities also for developers who intend to work with schema.org data using RDF-based tools and Linked Data, and defines a simplified publisher-friendly 'Lite' view of RDFa...We hope that our support for 'RDFa Lite', alongside Microdata, will allow publishers to focus more on what they want to say with their data, rather than on the details of its specific encoding as markup.
RDFa 1.1 Lite is pretty new itself -- Ben Adida first proposed it in late September. If you aren't familiar with the 'Lite' version (I wasn't), watch Ben's presentation or, more briefly, read semanticweb.com's summary -- the short version of which is:
RDFa 1.1 Lite is a simple subset of RDFa consisting of the following attributes: vocab, typeof, property, rel, about and prefix.
Caveat developer, however: the RDFa 1.1 Lite format schema.org just promised to support is based on the current working W3C draft, but the latest editors' draft has already introduced some significant changes, which are slated to enter the 'working' document fairly soon. As Ivan Herman said on the W3C blog:
The most important new feature, as far as the Schema.org examples also goes, is the changed behavior of the @property attribute: in the overwhelming percentage of the RDFa usage, it becomes synonymous to @rel. (What essentially happens is that, in the presence of, say, an @href attribute, the value of that attribute is bound to @property instead of a possible literal.)
Caveats aside, integrating RDFa (in some form or another) on schema.org definitely sounds like a good thing -- one more semantic web developer's worry fading away.