I would also suggest some kind of persistent identifier, and especially persistent URIs.
Sure it would be more complicated and longer in order to start using them, but the benefits would be good (as Ethan said, in a perfect web world).
Then, you would be sure all your records or digital objects (or any links) are linked using a persistent identifier, therefore never (or almost) having a broken link.
Brian explained the purpose quite nicely (with the handle system), but I would rather suggest using the full URI as your persistent identifier instead of the ID part itself, if not you (or your third party) would again end up needing a stylesheet or some kind of concatenation at one point or another for getting correct links.

So basically, if you are in a hurry, use a stylesheet for your third party distribution (to get your absolute links out of relative ones, easy if you have one server, a bit harder if you're using different domain names).
If you are not, think about this PID thing.
Actually even if you are doing it the transformation way (with stylesheet or not), take some time checking PIDs, it might become very handy for you, now or later, could be a nice investment.


On 6/21/11 5:43 PM, Nick Sincaglia wrote:
[log in to unmask]" type="cite"> The problem that you describe is a common one. I would like to recommend that you look into The Handle System (http://www.handle.net/). This system would enable you to create permanent public URIs that link to content items. If the links to the content change, the changes get updated behind the scenes but the public facing URIs will never change.


On 06/21/2011 09:19 AM, Ethan Gruber wrote:
[log in to unmask]" type="cite">Hi Nathan,

Many larger institutions probably have stored their media on other servers than their finding aids (for example, in institutional repositories).  In a perfect web world, each media object is bound to a permanent URI, so in your case, I would recommend replacing your paths with absolute ones if your intention is to provide access to the linked objects from the same URI for the long term.


On Tue, Jun 21, 2011 at 9:27 AM, Nathan Tallman <[log in to unmask]> wrote:
Greetings list,

I'm trying to get a handle on whether institutions are using relative or absolute links for linking digital content. Relative is probably the preferred for coding standards, but what if your are sending your EADs off to a consortium or central finding aid repository, in addition to your own website? Won't all the links be broken? Perhaps one can work some find/replace magic to make them all absolute.

Just curious how others are handling this, especially since finding aid repositories are on the rise.


Nathan Tallman
Associate Archivist
American Jewish Archives

Nick Sincaglia
NueMeta LLC
Digital Media & Technology Consultancy
Phone: 630-303-7035
[log in to unmask]
Skype: nsincaglia