PURL

Purpose: to discuss the role of PURLs, and to offer some guidelines.


A PURL is a Persistent Universal Resource Locator (URL). This is a service offered at http://www.purl.org/ , to enable people to publish a fixed name for a resource, which acts as a relocator to the items current location.

There are 3 important concepts to keep in mind when defining a PURL:

  1. The most important concept relating to a PURL is the DOMAIN. A domain is used to organize a series of related PURLs. A domain does NOT correspond to any particular web site; instead, it provides hooks for pointers to web sites. An example of a PURL Domain is http://www.purl.org/tcl/ .
  2. The next most important concept is the PURL partial redirector. This is what one wants to define when pointing to a particular location within a 'web' of web pages. For instance, the PURL http://www.purl.org/tcl/home/ points currently to the top of the Tcl Developer's Xchange. If that site moves, the owners of the PURL can point to the new home, and no one using the PURL addresses need change anything.
  3. Finally, last resource, and what I consider the misnamed resource <smile> is the simple PURL. In this case, what one is defining is the address of a particular file/web page. Obviously, this one entity might physically print off onto more than one page. And in fact, if it is dynamically generated, it may consist of more than one disk file. Nevertheless, it is perceived by the user as being a single location.

There are several downsides of PURLs of which one should be aware.

  1. If you define a string as a simple PURL, that's it - it can't be changed into a partially redirecting PURL. One might define a similarly looking PURL as partially redirection. In general, people would only define a simple PURL for a paper, front page or single page tutorial, etc.
  2. PURLs cannot be deleted. Ever. You can define them to be null though, if for some reason the item must disappear.
  3. Users of PURLs are going to tend to bookmark the real location under the PURL due to the way PURLs work.
  4. Some Web sites which provide search engine coverage or other similar add-on functionality do not accept redirecting URLs.

See http://purl.oclc.org/OCLC/PURL/FAQ for common questions, as well as the answers to those questions, with regards to PURLs.


One catch I have encountered a couple of times is that as you go through the creation or modification process the next to the last step APPEARS to indicate that your request has been granted. It hasn't. That screen is a last chance "are you certain this is what you want" (and in fact that's what it says if you read carefully enough). Just click the confirm button on this page if the links work as you expect.