XFML—eXchangable Faceted Metadata Language (via Guide to ease). Now this is interesting. It’s an open XML format designed to facilitiate the publication and distribution of metadata—it uses a load of terms that are currently way over my head (hierarchical and faceted taxonomiest, topicmaps?) but the general principle looks fantastic. I wrote a metadata system last year that used a relational database and it was something of a nightmare—XFML looks like it solves some of the problems I faced, although my biggest challenge was how to grab and present usable information from the huge amounts of metadata collected which is a problem that falls outside the scope of XFML. XFML is best summarised by the following quote:
Say you run a weblog on a certain topic. You put your posts into categories you have invented yourself. A friend of your also runs a weblog on the same topic, and also puts their posts into categories they invented themselves. Many of these categories are really the same thing, but they are called something else. You both publish your metadata in XFML format. XFML allows you to relate these categories, so they still keep their own names, but your system knows they are really the same. So you import your friends XFML file, and manually relate the categories that are the same. You set your weblog software up so it will import your friends file daily to check for changes. Now your software can automatically generate links to stories about the same topics on your friends weblog that show up with your stories.
Incidentally, Peter Van Dijck (the guy behind XFML) runs a superb blog covering information architecture and other web related topics—Guide to Ease. I’ve only been reading it for a few days but I’m hooked already.