Cool URIs for the Semantic Web
Cool URIs are a regular feature on this blog so it was with some interest that I read the W3C working draft announced just before Christmas, Cool URIs for the Semantic Web.
It is perhaps worth noting that this document is a
First Public Working Draft of an intended W3C Interest Group Note giving a tutorial explaining decisions of the TAG for newcomers to Semantic Web technologies
which I take to mean that it is still very much under development. I've sent my detailed comments to the appropriate list but I just wanted to note a couple of things here.
Firstly, it strikes me that this document doesn't really say a great deal about cool URIs as such. Rather it talks about the use of URIs in the context of the Semantic Web and, in particular, how URIs for what it calls variously 'real-world objects' or 'non-information resources' should be constructed and dereferenced.
Secondly, and more importantly it seems to me, the document highlights some current problems with terminology in this area. Back in 2004, the Architecture of the World Wide Web, Volume One introduced the notion of 'information resources' as being resources for which
all of their essential characteristics can be conveyed in a message.
Given this terminology, it seems logical that those things which are not 'information resources' should be termed 'non-information resources' (though the Architecture document itself doesn't use that term). Unfortunately, the term 'non-information resource' isn't exactly snappy at the best of times, especially when one considers that the printed version of a Web page (which is an information resource) is itself a non-information resource (by virtue of being a physical object).
To try and bypass some of this terminological soup the new working draft introduces the use of 'Web document' and 'real-world object' instead (though not always in a particularly consistent way). One can understand why... though I'm not totally convinced that the result in any clearer. Is the conceptual notion of the colour red a 'real-world object' as far as your average man or woman in the street is concerned?
I'm reminded of the early days of discussions around the Dublin Core, circa 1998 I guess, where the term 'document-like object' was introduced as a short-hand for those classes of things that it made sense to describe using the Dublin Core metadata elements. The term never really took off and, in some cases at least, only served to confuse things further than they were already confused.
The W3C seems to have a similar problem here... distinguishing those things that are, in some sense, on the Web from everything else. In our work on the DCMI Abstract Model, we often partitioned the world into three basic classes - 'digital resources', 'physical resources' and 'conceptual resources' - of which the first, IMHO, shares a lot of similarities with the W3C's 'information resources'?