↓ Archives ↓

The missing link

Over the past few weeks I’ve been thinking a lot about how we should provide links from a reference to a resource. (There is perhaps another question of whether providing a link from a reference to a resource is desirable in all cases – any comments on this welcome, but for the purposes of the current discussion I’m going to assume that generally given a reference to a resource that is available online, we should provide a link.)

There are several different scenarios for this depending on the type of resource we are referencing.

Let’s start with something that seems simple – a reference to a website. Typically if you cite and reference a website in a piece of writing, the reference would look something like this (using a modified Harvard style):

JISC Technology Enhanced Learning supporting Students to achieve Academic Rigour (TELSTAR) : JISChttp://www.jisc.ac.uk/whatwedo/programmes/institutionalinnovation/telstar.aspx (Accessed 24 August 2009)

This seems relatively straightforward – you are referencing a website, and you include the URL – what could be simpler?

However, if we used this in the context of an Open University course, we recommend that when linking to external websites that you do so via what I would term a ‘managed’ link. So, rather than link directly to http://library.open.ac.uk/ you would get this URL added to a database (maintained by the library) which would assign a locally managed URL for use – for example it could be something like http://managedurls.open.ac.uk/123456 (note, this is not a real example).

Why do we do this? URLs are often not as stable as we’d like, and over the lifetime of a course (several years), a resource may move where it is hosted on the web. By directing the link via a ‘managed’ URL, we can update where this redirects to as a resource is moved. If that resource is referenced in several courses, we only have to update the managed URL in one place to keep all of the links working. It also allows us to easily run link checking on all the links recorded in our managed URL database. Finally, we can collect statistics on the use of the links.

However, there are some problems. It means that an author (or editor) of the learning material can’t just put in a URL – they have to know what the managed URL is for the service – and if one doesn’t already exist, someone has to setup the managed URL before it can be added to the reference.

It also means that the reference example I used above changes into:

JISC Technology Enhanced Learning supporting Students to achieve Academic Rigour (TELSTAR) : JISC, Available from: http://managedurls.open.ac.uk/123456 (Accessed 24 August 2009)

Now, this is still technically correct as a reference and does the job, and within the original context of the reference is probably fine. However, what about if you were to use this reference outside the context of the Open University? Surely using the actual URL for the resource makes more sense (and avoid random traffic via our ‘managed URL’ service coming from use of this reference in non-OU contexts)?

So, it seems desirable that:

  • The ‘real’ URL is preserved so that it can be transferred with the reference if it is exported to another context
  • The ‘author’ of the reference doesn’t need to know the ‘managed’ URL for the service
  • We can manage the link when used within the context of the OU Learning environment

How would we do this?

How about we use the ‘real’ URL as a key to lookup a ‘managed’ URL? So rather than a URL like:


we instead use something like:


We could do this transformation ‘on the fly’, so we wouldn’t need to store the extra information in the reference – we could simply add in the extra information when we display the reference in the learning environment. Any export procedures would simply use the original URL.

Done correctly, if the URL didn’t already exist in our ‘managed’ service, it would simply redirect to the original URL – but at the same time it could alert a member of library staff to the need to add a new managed URL – and apply link checking etc. to it.

One specific method I’ve been pondering is whether we could use an OpenURL construct to push the original URL to a resolver service that would do the work of looking for a ‘managed’ URL. This would look something like:


The OpenURL resolver would need to have access to a database which mapped the URL to whatever the current address of the resource is, but we could make use of existing services like stats collection.

That seems like enough for one post – I’m interested in comments. I’ll come on to linking to other types of resources in subsequent posts.


  • Aug 25th 200911:08
    by Peter Sefton

    Nice ideas here – extrapolating this to the IR context, you could make all IR (or even institution-wide) links go via a local resolver service which would auto-register the each new URL and correlate with an ID, so something like:


    Then when resource 53485734875 is no longer available on the same URL you could find it somewhere else, even via the handle service if you can afford the $50pa.

    This would mean that for any web service you had a list of URLs you need to remap when you change the way it works.

  • […] written a couple of posts on linking from references to resources, and particularly on how we are planning to use OpenURLs to provide persistent links to web […]

  • Leave a Reply

    Sorry, comments are closed.