I monitor the www-tag discussion list, but more than half of it goes right over my head–so I was pleased when a colleague forwarded URNs, Namespaces and Registries to me. Don’t let the 2001 in the URL fool you, it has been updated quite recently. This finding provides an interesting counterpoint to rfc 4452 which I wrote about earlier.
Essentially the authors go about examining the reasons why folks want to have URNs (persistence) and info-uris (non-dereferencability) and showing how URIs actually satisfy the requirements of these two communities.
I have to admit, it sure would be nice if (for example) LCCNs and OCLCNUMs resolved using the existing the infrastructure of http and dns. Let’s say I run across an info-uri in a XML document identifying tbl as info:lccn/no9910609. What does that really tell me? Wouldn’t it be nice if instead it was http://lccn.info/no9910609 and I could use my net/http library of choice to fetch tbl’s MADS record? Amusingly Henry Thompson (one of the authors of the finding) is holding http://lccn.info and http://oclcnum.info for ransom :-)
Instead, in the case of info-uri, OCLC is tasked with building a registry of these namespaces, and even when this is built the identifiers won’t necessarily be resolvable in any fashion. This is the intent behind info-uris of course–that they need not be resolvable or persistent. But this finding raises some practical issues that are worth taking a look at, which seem to point to the ultimate power of the web-we-already-have.