Friday, January 9, 2009

Introduction to Link Resolvers

What is a link resolver?

Jargon words italicised and explained below.

A link resolver is a piece of server software that translates an OpenURL request from source into a URL that will retrieve an item from a specific target.



Is a standard way of representing citation-type information as a URL (OpenURL entry on wikipedia) e.g.

The first part ( is the link resolver's base URL and the rest conforms to the OpenURL standard (version 0.1 in this case, there is also a version 1.0 which is much more sophisticated).


A source has two features it contains citation information (i.e. metadata about a bibliographic item) and it has the ability to create an OpenURL by appending the citation information (in OpenURL form) to the base URL of a link resolver.


Uniform Resource Locater, or as we commonly say 'a web address'.


A target is a web server that stores bibliographic items (preferably full text). Good targets allow you to 'deep link' to specific items like journal articles or conference papers. Bad targets only allow you to deep link to the journal's home page or conference proceedings' home page - some targets are so bad they don't even allow that level of linking (Westlaw springs to mind).

OK, So What's a Link Resolver do again?

The Link Resolver gets an OpenURL for an item from a source, checks to see which targets we have access to can provide that item then creates a URL that deep links to that item. So effectively if you find a citation for an article in a source you can click on the OpenURL and the link resolver will check all your e-subscriptions and display the article, even if the source is purely an indexing and abstracting database with no full text component.

Sounds Cool, can we get one?

Actually we've had one since 2004 - which you might know as the 'Find It' button. For four years our Link Resolver software has been SFX. We are now transferring to some new software called 360 Link. Fundamentally nothing changes (except we expect the new software to be much more accurate in searching our e-subscriptions).

Is a Link Resolver good for anything else?

I'm glad you asked. It can accept a request from anything that can generate an OpenURL. Apart from our databases Citation Linker can create an OpenURL from the citation details you provide.

EndNote (thanks Nicole) can also pass OpenURL to our Link Resolver, so that clicking on OpenURL for a citation will search our e-subs and display the item. Zotero and other bib/ref tools can also pass OpenURLs to your Link Resolver.

, a way of marking up citations on a web page so that browsers with COinS-enabled plugins can determine the address of the user's link resolver and create an OpenURL for it.

What's really cool about OpenURLs is they bypass broken links caused by deeplinking via a static URL. For example, consider that in Reserve Online we deep link to course readings. What happens when we transfer the subscription to another publisher, or the publisher updates their site, or they are swallowed by another publisher (like Wiley's acquisition of Blackwells journals last year)? The links break, and Document Services staff have to detect the broken links, find new ones, and change the data in Masterfile. If we used OpenURLs instead it wouldn't matter if the links changed - and if we had multiple subscriptions and the prime one didn't work there would be options to try the others.

And lecturers could use it to create persistent links from citations in reading lists in Blackboard directly to the item. I even created a simple OpenURL creation tool build an OpenURL from a citation on JustUs, which is currently using SFX but I will change to 360 Link soon. In any case 360 Link has a function on its 'more information' screen that will show you the relevant OpenURL for the item you requested.

Let me know if this was useful to you, or if it was too simple or too complex.