On Mon, 14 Sep 2009, Eric Hellman wrote:

> The original question was whether it's a good idea to use OpenURL for a URL 
> persistence application.
>
> Issues with using PURL are that
> 1. it only works with one website- PURL paths don't travel, though there have 
> been proposals to get around this.
> 2. There's not a really good way to package metadata with the PURL reference.
>
> If there was some standard other than OpenURL that would do the trick, then 
> we'd probably not be looking at OpenURL- there I agree with Jonathan.

Not a standard per se, but issue #1 can be handled by some of of high 
availability system -- it can even be as simple as a few organizations 
federating together to resolve either other's items, and then setting up 
load balancer or the sloppy route of DNS round-robin.

For issue #2 ... there's always the issue of linking straight to the 
object vs. presenting the metadata.  If the goal is to track metadata and 
not just link straight to the object, you might be able to use purls to 
link to OAI-ORE [1] documents. (although, I admit, I've never actually 
used ORE myself)


> You can't maintain a tr.im url or a bit.ly url, by the way. Those services 
> can't support post-creation modification of the target url, because if they 
> did, they'd get killed by spam.

Agreed.


[1] http://www.openarchives.org/ore/


-Joe