DBpedia, Law, and Structured Data

In law we’re used to structured data, although, like the oft-mentioned M. Jourdain, we may not know it by its fancy name. Very roughly, it’s data that’s been labelled in some useful way, so that it can be found or otherwise manipulated using that label. So, when we tell CanLII or our favourite commercial database that we’d like to see R. v. Molière, please, we’re wielding the label “case name,” for example. Like most professions and trades, we analyze our tools and products, naming the parts and their relationships, creating various “ontologies.” It starts simply in law school, where you learn that there are cases and there are statutes, and that cases can be understood to comprise “facts,” “issues,” “rules,” and the like, so as to make your briefs into structured data.

In all of this, law is to some extent ahead of many areas, which are only now grappling with how to structure data on the internet. I don’t suggest that law has laurels it can rest on, though; indeed, its having got there early has created a sense of complacency and a pile of old capital that stand in the way of needed change. Only consider what looks to be a miserable (and miserably late) launching of the online Ontario Reports, at least as judged by Garry Wise today: the potential advantages of having digital data online were largely ignored or suppressed (though compare a subsequent piece on Slaw by Ted Tjaden). As I’ve suggested here, the opportunity is ripe for the profession to collaborate on new ways to structure all legal documents so that legal work can be made more efficient.

It’s useful to look at current efforts to master the online data flood. DBpedia is a German organization (writing in English) that is doing some very interesting things with those bits on Wikipedia that are labelled. It has pulled in all of the more than a million items and makes them available online as structured data, mostly supplying developers. But they’ve got a front-end of their own that lets you experiment with searching their database and refining your search with the “facets” that Wikipedia offers. It’s not the easiest tool in the world to figure out, despite the help page, but if you run some of the canned searches they offer you, you’ll get the idea if not the technique. I managed to construct a search for public universities in the Canadian city of Toronto, so if I can work it, I’m certain you’ll be able to.

There’s not much if anything that’s of direct use to law here. But that’s not the point. My aim is to start you thinking about your ideal sets of structured data, about how you might ideally go about searching them, and — here’s the rub — about how the profession can be mobilized to apply the structure in the first place.