Dealing With Link Rot – Are DOIs the Cure?

Over the past decades the publishing industry has developed standards to provide unique identifiers to text products. The most well known is the ISBN, the International Standard Book Number, which now comprises 13 digits, and ensures the same titles published in different parts of the world can be identified separately. The version used for periodicals is the 8 digit ISSN – International Serial Serial Number. Then there is the International Standard Text Code (ISTC), a numbering system for the unique identification of text-based works, which links different text works within books, audio books, etc. All of these standards ensure when an author or purchaser is citing a work, the original work would be correctly identified, and confusion between similar titles, for example, would be avoided.

These standards are not applicable to citations using links from the internet. The issue of link rot, where a link used in an article which was written some years ago may no longer link to the original piece, is becoming more prevalent as time passes. The problem has been addressed by others, including blogs written on SLAW, and there are studies and reports outlining the problem. If you write an article, and provide URLs in your footnotes, chances are that these URLs may change over time. Servers move, websites upgrade, websites are redesigned and take on new URL identities. At the time of writing your piece, you are providing the reader with a great service, the chance to link directly to a source you have quoted. But with the passage of time, if the URL changes, short of republishing your article with new links, there is little that you, as the author, can do.

There are several solutions offered to counter the problem, using the concepts of international standards, similar to the ISBN. These involve allocating a permanent link to an article. One service, (with which we at the law library have affiliated) is called Perma, and was discussed by Simon Fodden on SLAW last year . It aims at authors, allowing them to create a permanent archival link to their articles, and for these links to be included in law reviews, and has buy-in from many prominent law faculties and institutions.

A project like Perma is well intentioned, seeking to achieve a consistency for law related articles, and giving the author a role. This may be most useful where the journal itself does not automatically create permanent links using an accepted standard already.

However, a number of registration agencies exist which use the DOI standard of a structured character string to provide unique digital object identifiers to all manner of digital objects; mainly journal articles, but also book chapters, video, and theses, etc. The DOI becomes the persistent citation to an article. Many publishers have adopted the DOI standard and this citation can be found alongside the titles in the journal.

There is one registration agency I thought it might be useful to mention, CrossRef, formed as an association of scholarly publishers which now includes libraries, standards organisations, NGOs, IGOs, etc. They are catholic in our definition of “publisher”. Established in 1999, its aim, as a not-for-profit organisation, is to facilitate scholarly communications. They offer several services to publishers, such as CrossCheck, software to prevent scholarly as well as professional plagiarism. The have been issuing DOI standard permanent links for journal articles for some time.

From a presentation on their website, they state:
“A DOI persists throughout changes in copyright ownership or location because it’s just a name used to look up an address in an easily updateable directory”

With over 65 million items indexed with a DOI and a free look up section, CrossRef has proven to be an invaluable tool for us recently. We needed to urgently obtain persistent links to a couple of hundred journal articles written by our academics, which were submitted as part of the Research Excellence Framework – REF – exercise universities undergo every few years in the UK. We had titles and authors, and sometimes the ISSN of a journal, but not all the hard copies of the journals, nor even subscriptions to all the titles.

On the look-up section we were able to locate the DOI – which looks like this example – using titles or authors.

It is also possible to look up the actual DOI on the CrossRef inquiry page, using only the digits: 10.1017/S1472669612000643. CrossRef now provides a lookup for a batch of multiple references which is a useful addition.

The next step will be for researchers to encourage their publishers who do not yet do so to implement DOI on their publications, which can then be confidently used by authors when citing to a journal article.

But of course this does not tackle the issue of online legislation, government reports, and cases. There is a need to include these digital objects as well; those websites also change over time. Government departments in the past have been the worst offenders, not archiving reports, etc consistently on the internet, and making wholesale changes to websites when governing parties change at election time. And how nice would it be to have this sort of standard available within databases such as Lexis or Westlaw, especially for cases where the citation is to the paper volume – e.g. [1932] A.C. 562 – and not a digital citation. The URL in Lexis Library for me when I access this case is currently:
How user friendly is that?

CrossRef are currently testing a service that will allow one to assign a DOI to any content on the web. The pilot, tentatively called “OpCIt” was discussed at their annual meeting in 2013. This may be an answer to the link rot problem.

There is an urgent need to have the same reliability of persistent and consistent access to the digital version of cases, government reports and publications as we have to our physical, leather bound volumes, such as 32 Henry VIII, (ie, the acts passed in 1540 in England), a citation unchanged by the passage of time, the object still fully legible, and through its class mark, easy to locate on an assigned shelf in our physical library.

Comments are closed.