Jeni's Musings

(thanks for the twitter heads-up)

“Given that a significant proportion of the people who work with the semantic web and who want to use metadata embedded within web pages are currently using RDF, I would have thought you’d use a “pave the cowpaths” approach.”

The goal wasn’t to address “the semantic web”. The goal was to address the needs of a variety of people who want to annotate their HTML pages for various reasons (see http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2009-May/019681.html for some of the scenarios that I had in mind; I’m going to be posting more detailed comments on more scenarios and how we approach them shortly). That is, in designing microdata I was trying to address specific concrete problems, I was not trying to plug into the RDF world. None of the problems needed the RDF world as far as I can tell.

Having said that, I did try to bridge the gap to RDF, by defining how this stuff maps to RDF. (Also JSON, for similar reasons, and now iCalendar, vCard, and BibTeX.)

For dates and times, the DOM interface does have ways to interact with the “time” element using native JS Date objects (timeElement.time, .date, and .timezone).

For XML literals, there really weren’t any use cases that needed them. For event descriptions, vEvent (from iCalendar, used as the event vocabulary in HTML5 now), doesn’t support structured markup, so it’s not clear that XML literals would actually be useful there.

Is there any way for me to test what RDF processors do when they find an object is a “time” when they expected a “date”, or when they find a string when they expected a datetime? I really would like to study this more, but I don’t know how to study it. What tools process RDF event data, for instance?

For validation in general (of custom vocabularies used in microdata), I’m thinking the best plan might be to leverage the RDFS/OWL world and just say that tools should use those mechanisms to validate microdata, by translating the microdata markup to triples (as defined by the spec) and validating the triples. This is more or less the same story that RDFa would have for validation and editing.