BarCampAustinIII Session : Practical Semantic Markup

This is a warts-and-all discussion of the background, goals and

approach of the session.

Background: The W3C is finalizing RDFa, a way to embed semantic

information in web pages. Unlike RDF/XML and other RDF-related

notations, you don't really have to know anything about RDF to

use it. In fact, it comes out looking a lot like microformats.

RDFa has at least an outside shot at getting used by regular old

web practitioners, an audience that until now has been lukewarm at

best towards big-S semantic web technologies like RDF.

The goal of this session (or sessions, or workshop, that hasn't

been finalized yet) is to introduce practitioners to RDFa (and

associated technologies) in a way that demonstrates the cool

parts but not only doesn't require a background in Knowledge

Representation research but doesn't even _mention_ RDF, the

Semantic Web, or any of that other stuff. (That isn't actually

possible, of course, but let's see how close we can get)

The (original) outline went something like this (think of a

Disney ride where you're in a boat that runs on underwater

rails. You might get a couple branches, but free will is an

illusion and the outcome is predetermined)

• Pick a domain that doesn't have any microformats support and doesn't show any signs of getting microformats support anytime soon. (Trip itineraries seemed to the consensus over dinner)

• Come up with a little-m model of what you'd want to record, and some simple markup (that happens to confirm to RDFa) that gets you the info. Work off real-life examples from travel blogs, etc, instead of trying to come up with some sort of perfect big-M model.

• Wave hands and claim developers can do magic, then whip out a web service that produces the notation, one that consumes it, and an operator plugin that shows what you can do with it.

The idea is that a relatively small community can come up with

a reusable and useful model of their domain, with relatively

little working going into the underlying technology. (As opposed

to other approaches, which require elaborate custom processing

rules and don't naturally leverage the really pretty impressive

set of tools available for dealing with RDF)

Or something like that.