Raw

Sam Ruby's doing wonders with his branch of the Planet Planet code, Venus. The core data is maintained as Atom (containing XHTML). A nice touch is that it now has FOAF blogroll support.

I'm getting particular pleasure from what Sam's been doing because it pretty much epitomises the "content model layer" notion I've occasionally rambled about (I've a proper writeup in the pipeline). Basically that people are tending towards a conceptually view/way of working with Web data that can be seen in Semantic Web terms as a content-oriented vocabulary, but from a more traditional point of view is a highly generic model (of content). It's a kind of intermediary between Web of Documents and Web of Data. I tihnk this is nice-to-have in itself as a bridge. But in practice what's happening is much richer, because of all the thought going into protocols (or to be more precise, the decision to leverage URIs and HTTP), and the potential for the content to be used as a transport for data (either directly or via microformats), with the potential for versioning support ( atom:updated) built in. Not only is this kind of thing heading in the general direction of the Semantic Web, it's also offering a lot that can enrich/inform how to make that Semantic Web better.

Anyhow, if my content model layer hand-wave analysis makes sense, then (derived from the SemWeb intuitions) the approach should greatly simplify useful apps and encourage innovation. I think Sam's latest post gives me a data point - he's got a couple of dozen lines of code that when applied to Venus' data extracts the links corresponding to current topics of conversation within Sam's circle of friends. Neat. He ends the post with a challenge:

...why can’t every public (and private) planet produce a meme feed?