In the last few days I've been working on developing a wiki that will, using Kake's CGI::Wiki modules, support embedding metadata in the <head> section of the rendered pages in RDF, in the format that zool has developed for her spacenamespace mapping project.
Last night, while I was working on this, I was discussing the Right Way To Do Things with people on #rdfig, the IRC channel of the W3C RDF Interest Group. The main focus of the discussion was my trying to understand the abstraction of subject from object in the context of URIs. This is, apparently, the httpRange-14 discussion (see this whacking great huge thread on www-tag). It took me a very long time to get my head around the concept (and I can't help feel that I made an ass of myself in the process, by taking out my frustration with myself on other people), but zool and DanC were very helpful.
Eventually I performed the mental equivalent of passing a kidney stone and understanding dawned. What it took me the time to grasp was that a URI is not the same as a thing, and that if http://example.com/foo is talking about "foo" then to refer to foo itself you should refer to http://example.com/foo#thing, or #it or #object or some other identifier. For a good essay on the subject, read Tim Berners-Lee's What Do HTTP URIs Identify?.
With that out of the way, I was able to get the RDF sorted out. So now, if you look at this node on my development wiki, you'll see the various kinds of metadata that have been entered for it, and viewing the HTML source will reveal some embedded RDF using that metadata. This is the way that Grubstreet will be moving in the near future, in order to be part of the Semantic Web. We're also thinking about allowing people to contribute via SMS, by using deginge's "sms^narf" bot in combination with my very own wikibot, and DrHyde has convinced us to add an email interface as well. Things are starting to get very interesting.
