Hi, I make sites like
http://ny-pictures.com/
and
http://carpictures.cc/
that are built from Linked Data, and I'd also like to publish Linked Data on them. It's no secret that W3C standards for semantics can be pretty impenetrable to the average webmaster, so I'd like to see the WHAT-WG play a useful role in developing compromises.
I saw this
http://wiki.whatwg.org/wiki/MetaExtensions
which looked good to me at first but bothered me on close reading. One trouble is that it hijacks Dublin Core properties and defines a detailed semantics for them that isn't the one from DCIM. Second, that it contains a number of detailed taxonomies for things like "audience rating", "subject", "license", etc.
I prefer much more to refer to named entities in a "linked data" style, because this lets people use whatever classifications of things that are appropriate for their use.
For instance, there are ESRB ratings for video games and MPAA ratings for movies. These can be mapped to a specific URL namespace. The use of namespaces means that "audiences" can be defined for different kinds of content using already-accepted vocabularies.
Similarly, people might want to specify subjects with dmoz categories, the dewey decimal system, LC Subject headings, dbpedia entries about topics, etc. No one system is going to satisfy everyone, but a linked data system is expandable, keeping things in separate namespaces.
I ~do~ think the efforts of the WHAT-WG towards creating taxonomies could be useful there, but that it should grab a chunk of namespace for it's own "terms".
Personally, I'd like to see at least one simplified RDFa profile in HTML5. Right now I'm working on something I call "RDFa For Turtles" that's mean to be something that can be specified in 1-2 pages and implemented by ANYBODY... Here's some rough notes:
http://lists.w3.org/Archives/Public/pub ... /0094.html