Hi Geoff,

I'd be interested to hear your thoughts on Stephen Downes' take on all this:

http://blog.jonudell.net/2007/01/26/a-conversation-with-tony-hammond-about-digital-object-identifiers/#comment-862

Well - setting aside for the moment the theories about our motivations - I think that the root of this is a flawed technical understanding of the W3C's "Cool URLs Don't Break" guidelines to mean there is never a need for levels of indirection (e.g. using Handles and DOIs.) We too think that cool URLs should never change, the problem is that the W3C examples are dealing with cases in which one has control over (or cooperation from somebody who has control over) the URLs in question. This isn't always the case.

If Josiah Carberry, the editor of the Journal of Psychoceramics (the study of cracked pots) decides to move from publisher A to publisher B, he can not count on publisher A maintaining URL redirects to publisher B. The "commercial churn" argument isn't a smokescreen, it is a very real example of where even cool URLs can break if there isn't a level of indirection introduced.

Note also that this is different from the case cited where Blackboard and WebCT merged but still had control over all the URLs in question.

And you are right, a similar situation occurs outside of academic publishing. If I switch from the hosted MetaMegaBlogger service to the SemanticMegaBlogger service, I might very well want links referring to my past entries to persist. I can not rely on MetaMegaBlogger to provide this service at all, to say nothing of "indefinitely." I could easily see a service similar to CrossRef appearing for other communities. Bloggers could certainly use it.

--G