next up previous
Next: Resolving URNs Up: UNIFORM RESOURCE IDENTIFIERS & Previous: Problems with URLs

URNs

It can be seen that there is a need for a more permanent way of naming networked resources. This could make use of URLs to point to specific instances of a resource, but would not be subject to the dangling link problem. This section introduces a new technology which attempts to provide this higher level of naming - the Uniform Resource Name (URN).

As currently envisaged[18], URNs consist of:

This provides a mechanism which may be used both to identify on-line resources, and as a way of incorporating existing widely deployed naming schemes such as International Standard Book Numbers. It is intended that URNs should be usable in the same contexts as URLs - e.g. magazines, books, newspapers, radio, and television programmes.

To see how this might be done, consider the ISBN number 0-586-03716-0 [19]. This could be translated into the URN

  isbn:0-586-03716-0

In other words, the naming scheme would be isbn, and the ISBN number would become the element ID.

The use of existing naming systems such as ISBNs and ISSNs raises intriguing possibilities such as searching for books and periodicals using their ISBN number. The reader is left to consider the implications of this development, e.g. in the use of the Internet to facilitate inter-library loans.

When URNs are to be transmitted via a medium where corruption is likely, such as paper, it is recommended that they be wrapped in the characters <URN and >. This will make it simple to distinguish URNs from any surrounding material, e.g.

  <URN:isbn:0-586-037
  16-0>

Information providers should be able to use URNs wherever they are currently using URLs, e.g.

  <a href="urn:isbn:0-586-03716-0">Ubik</a>

Exactly what happpens when the user follows such a URN link is at the discretion of their browser, e.g. it might display a suitably rendered list of URLs, or choose a URL automatically according to a set of rules or weightings. An example of this would be preferring URLs which match as much as possible of the Internet domain name of the user's computer, giving preference to the `nearest' copy of a resource. There is also the question of what exactly a URN lookup returns - this is discussed in the section on URCs below.



next up previous
Next: Resolving URNs Up: UNIFORM RESOURCE IDENTIFIERS & Previous: Problems with URLs



Martin Hamilton
Tue Jun 20 12:43:30 BST 1995