Jeni Tennison recently discussed microdata and RDF on her blog: “As part of the ongoing discussion about how to reconcile RDFa and microdata (if at all), Nathan Rixham has put together a suggested Microdata RDFa Merge which brings together parts of microdata and parts of RDFa, creating a completely new set of attributes, but a parsing model that more or less follows microdata’s. I want here to put forward another possibility to the debate. I should say that this is just some noodling on my part as a way of exploring options, not any kind of official position on the behalf of the W3C or the TAG or any other body that you might associate me with, nor even a decided position on my part.”

Tennison continues, “As I’ve said before, RDFa, in my experience, is complicated not primarily because of the whole namespaces/CURIEs issue but because its processing model tries to be too clever. RDFa was designed to largely fit in with existing markup and turn it into embedded data ‘just’ by adding a few attributes here and there.”

Tennison provides an example of what she means, then states, “Even having written an RDFa parser, having written code to mark up documents with RDFa, having taught it, I still cannot write RDFa past a trivial example and be 100% sure that it will produce what I was aiming to produce. If we were to look at really simplifying RDFa, rather than making cosmetic changes, we need to address this complexity. It would certainly mean backwards-incompatible changes, such as dropping the use of particular attributes and revising the way the processing model works, such that future RDFa processors couldn’t be used on RDFa 1.0.”

Read more here.

Image: Courtesy Flickr/ *n3wjack’s world in pixels