For this special MWG session we are very pleased to welcome Juha Hakala, Director of Database Services at the University of Helsinki, who will be speaking to us about Uniform Resource Names. Mr. Hakala is one of the founders of the Dublin Core Metadata Initiaitive and has led the Nordic Metadata Project.

Uniform Resource Names (URNs) are intended to serve as persistent, location-independent, resource identifiers. Standardization of all aspects of the URN system is now complete. Several organizations, including libraries, have started registering URN namespaces for diverse identification systems. Examples of already registered systems familiar for libraries include ISSN, ISBN and NBN (national bibliography number). From a national library's point of view, NBN can be seen as a fall back mechanism for unique and persistent identification of resources which do not qualify for any more established system such as ISBN. There are at least two good reasons for using URNs instead of internal identifiers: global uniqueness, and global resolution. The latter benefit will only emerge in full when the Internet infrastructure supporting URN resolver discovery is in place, and when Web browsers and retrieval systems such as Integrated Library Systems or Digital Library system such as ENCompass support URN resolution.

Several European national libraries use NBN-based URNs for identification of electronic resources in their digital archives. One example is the E-depot system DIAS system recently implemented in the Dutch national library. Their URNs look like this:

URN:NBN:nl:kb:eDepot-1039428424571

This syntax is as defined in the RFC 3188. According to that RFC, the namespace specific string (in the example, eDepot-1039428424571) can be basically anything as long as the string is unique; the Dutch are using UNIX time of the moment the URN was generated, in the Finnish Web archive MD5 is utilized.

A national library is entitled to divide its own domain (such as urn:nbn:fi: for Finland) further, so that e.g. Cornell might have its own domain (e.g. urn:nbn:us:cornell:) which can be divided further locally as needed.

He will finish his presentation with a few words on the relationship between URN and DOI in one hand, compared to OpenURL and URN on the other. Regarding the former relation, Hakula believes in peaceful coexistence, although he does think that DOI is technically inferior to URN.

And if there is time left, Hakula will present the ISBN revision work which, due to the procrustean bed of the EAN system, is going to deliver a less than optimal solution which will require a lot of work from the system vendors, but will not guarantee the capacity of the system for long enough.