March 16, 2006
DOI and OpenURL Resolution
One of the things that I keep coming back to in the area of technological support for elearning is how best to reference online resources in a persistent way.
Earlier this week, I fired off a sanity checking email to Tim Wales, the Technology Faculty liaison in the OU Library to see if this link pattern: http://libezproxy.open.ac.uk/login?url=http://dx.doi.org/whateverDOI worked in general for sites that were registered with the OU ezproxy, the idea being that the ezproxy would handle authentication with commercial database providers.
The thing that was concerning me was that http://dx.doi.org/ resolves a DOI to another website, and I wasn't clear whether the libezproxy script would be happy with this.
Well, it does seem to work , but perhaps more interesting was the reply that Tim sent back. In particular, he made the very useful point that using the DOI would not necessarily result in a direct link to content in perpetuity for cases where the OU's subscription to the host of the particular article identfied by the DOI lapsed.
[F]or article linking, I was tending towards considering advocating the alternative method of using an OpenURL link with our Linkfinder+ resolver [internal link] as it would automatically take users to an alternative source of the document if one existed thanks to our multiple subscriptions or present an alternative search options screen. With a DOI link, if our subscription to the resource expired, the user would have to know to look on our catalogue for other options or would think they would have to buy it from the host site.
Tim then went on, it seems, to hack a solution:
The ideal would be to combine the two together and this facility does exist in Linkfinder+. I have just worked this out, this is the combi DOI openurl link ... :
http://libezproxy.open.ac.uk/login?url=http://openurl.open.ac.uk/lfp/LinkFinderPlus/Display?&id=doi:theDOI
err? http://libezproxy.open.ac.uk/login?url=http://openurl.open.ac.uk/sfxlcl3?&id=doi:theDOI, which then becomes http://openurl.open.ac.uk.libezproxy.open.ac.uk/sfxlcl3?id= maybe?
So - here's a rather neat recipe for: 1) resolving a DOI, 2) against the OU LinkFinder+/OpenURL resolver, and 3) accessing the article via libezproxy.
As an example, (if you're behind SAMS), Tim concludes:
[Here's a link to a database we do subscribe to:] http://libezproxy.open.ac.uk/login?url=http://openurl.open.ac.uk/lfp/LinkFinderPlus/Display?&id=doi:10.1016/S0040-1625(03)00072-6This is an example of what happens with a DOI'd resource that we do not
have access to with our current Linkfinder+ configuration:
http://libezproxy.open.ac.uk/login?url=http://openurl.open.ac.uk/lfp/LinkFinderPlus/Display?&id=doi:10.1002/0470841559.ch1
You can see that the Publisher link does take the user to the resource in the end.
One thing to note is that not all OpenURL resolved resources will necessarily require libezproxy access (err, is that right?) - anyway, under that assumption, I have set up a OUseful shortcut - http://ouseful.open.ac.uk/doi: - which redirects to http://openurl.open.ac.uk/lfp/LinkFinderPlus/Display?&id=doi:
http://openurl.open.ac.uk/sfxlcl3?&id=doi:, enabling LinkResolver resolution of doi queries.
Just while I'm on the topic, here is an OpenURL link generator, err - this maybe? citation generator?, and here is a link to the OU Open URL resolver.
There are a couple of official LinkResolver graphics, too: 
and 
For finding DOIs, here's a DOI lookup service, and here's a DOI resolver (and another DOI resolver).
I guess the next thing to do is to get these things working with a COinS bookmarklet or Greasemonkey script!
PS a couple of things that came to mind when trawling items for this post: 1) the same document (?terminology) may be available on several commercial databases, and each database will have it's own DOI for the document- is there an xISBN equivalent for DOIs (xDOI)? The LinkResolver+ I guess works by looking up the reference of the artcle referenced by a DOI, and then finding a database that the OU subscribes that serves that document - which is not really xDOI, though it has a similar effect. 2) Using libezproxy results in URLs of the form http://ieeexplore.ieee.org.libezproxy.open.ac.uk/, for example. Does this mean the OU is effectively republishing the ieeexplore.ieee.org site in this case, which is presumably a breach of copyright?