I have been hitherto rather ignorant of the Semantic Web, and RDF - a seemingly incomprehensible language that caused me to bodyswerve RSS 1.0.
I read, yesterday, a short profile of Sir Tim Berners-Lee, on the occasion of his knighthood (I'm not sure whether or not I should be against such baubles, but I know in my heart that I'd jump at the offering), wherein he referred to the Semantic Web, and all of the lovely things that this would make possible. It made me think that metadata might well be worthwhile for things other than Search-Engine Optimisation....
I think my idea began to germinate last night, when I bemoaned my inability to review everything that had happened on my website. It developed today, especially during a conversation with Gary about FOAF, a means of specifying what people one knows, and other information about oneself. Specifically, there's a new website called PLINK (people.link), which provides an attractive means of viewing the raw RDF that makes up a FOAF file, and also facilitates navigation between people, based on forward and reverse lookups of who claims to know whom.
All of which brings me to my concept, not that it's particularly ground-breaking. I'd like (or like to build, perhaps) a system that allows me to encode as much as I choose of my life experience in RDF, be it films that I've seen, songs that I've heard, places that I've visited, books that I've read, or any conceivable object in a sentence in which I am the subject (a tad ego-centric, I know, but that's life, or, rather, blogging).
This data would be useless without a couple of things:
- Machine-readable versions of popular database-websites, such as the IMDB, the All Music Guide, and others, which could provide URIs for films that I have seen etc., further metadata about the particular films etc., and a focal point for other people to use as their URIs (thereby allowing two people's experiences of a single film/song/etc. to be compared). Also, it would be nice if more personal websites were machine-readable to some extent (in which area I guess FOAF and RSS 1.0 are useful developments)
A user-friendly means of displaying the RDF graphically, similar to PLINK for FOAF files. I'm not entirely sure how this would work, but I guess it would be along these lines:
There would be a template (possibly more than one) (similar to a blog template) for each kind of information in the collection of RDF files. This would define how to render a page about a film, a book, a CD, a place, etc. An example for a film: there would maybe be some information about the film from IMDB (such as a poster, the leading actors/actresses), the date that I went to see it (if that is known), and the text of, an abstract of, or a link to a blog post in which I review the film. For a place, there could be a gallery of photographs that I took in that place, some blog posts that I made from that place, or whatever would be more appropriate. (Obviously, a bit of funky caching would be necessary
With this (admittedly under-specified) idea, it would be possible to generate a huge site using a mixture of syndicated material, and my own content. Not only that, but if other people ran with the concept, the potential for syndication and search applications would be an exciting proposition. The more I think about this, the more it appeals to me. Maybe I'll just go and do it.