RDFa 1.1

Lately I’ve been looking at RDFa 1.1. Previous versions of RDF and RDFa had been hampered by not being usable in ordinary HTML. RDFa 1.0 could be used only with XHTML. With version 1.1’s usability in HTML5 as well as XML, a lot more possibilities for embedding metadata in documents arise. It’s invisible in the browser but can be extracted more easily and reliably than data can be mined from ordinary Web pages. Dublin Core metadata, for example, is often expressed in RDFa.

The quick explanation, in case you aren’t familiar with it: RDF is an extensible way of expressing arbitrary data relationships by triples, which consist of a subject, a predicate (property type), and object (property value). It makes heavy use of IRIs, which are like URIs but allow the full Unicode character set; IRIs are used rather than simple names to avoid ambiguity. “Title,” for instance, means one thing when talking about books and something else when talking about British nobility; different IRIs can distinguish between them. To create a new RDF vocabulary, you just have to create new IRIs. RDFa is a way of expressing RDF in XML or HTML5 syntax. You’ll often run into the term “linked data” in connection with RDF.

I’ve started work on a possible tool to take advantage of RDFa data in HTML. I start a lot more projects than I ever bring to completion, so for the present I won’t say more about it.