It's not quite the same thing, but I worked on a project a couple of years ago integrating references/citations into a learning environment (called Telstar , and looked at the question of how to deal with broken links from references.

We proposed a more reactive mechanism than running link checking software. This clearly has some disadvantages, but I think a major advantage is the targetting of staff time towards those links that are being used. The mechanism proposed was to add a level of redirection, with an intermediary script checking the availability of the destination URL before either:

a) passing the user on to the destination
b) finding the destination URL unresponsive (e.g. 404), automatically reporting the issue to library staff, and directing the user to a page explaining that the resource was not currently responding and that library staff had been informed

Particularly we proposed putting the destination URL into the rft_id of an OpenURL to achieve this, but this was only because it allowed us to piggyback on existing infrastructure using a standard approach - you could do the same with a simple script, with the destination URL as a parameter (if you are really interested, we created a new Source parser in SFX to do (a) and (b) ). Because we didn't necessarily have control over the URL in the reference, we also built a table that allowed us to map broken URLs being used in the learning environment to alternative URLs so we could offer a temporary redirect while we worked with the relevant staff to get corrections made to the reference link.

There's some more on this at although for some reason (my fault) this doesn't include a write up of the link checking process/code we created.

Of course, this approach is in no way incompatible with regular proactive link checking.


Owen Stephens
Owen Stephens Consulting
Email: [log in to unmask]
Telephone: 0121 288 6936

On 23 Feb 2012, at 17:02, Tod Olson wrote:

> There's been some recent discussion at our site about revi(s|v)ing URL checking in our catalog, and I was wondering if other sites have any strategies that they have found to be effective.
> We used to run some home-grown link checking software. It fit nicely into a shell pipeline, so it was easy to filter out sites that didn't want to be link checked. But still the reports had too many spurious errors. And with over a million links in the catalog, there are some issues of scale, both for checking the links and consuming any report.
> Anyhow, if you have some system you use as part of catalog link maintenance, or if there's some link checking software that you've had good experiences with, or if there's some related experience you'd like to share, I'd like to hear about it.
> Thanks,
> -Tod
> Tod Olson <[log in to unmask]>
> Systems Librarian     
> University of Chicago Library