Tuesday, February 10, 2015

Reference rot in web-based scholarly communication and link decoration as a path to mitigation

Reference rot in web-based scholarly communication and link decoration as a path to mitigation.
Martin Klein, Herbert Van de Sompel. LSE Impact of Social Sciences blog. February 6, 2015.
The failure of a web address to link to the appropriate online source is a significant problem facing scholarly material. The ability to reference sources is a fundamental part of scholarship. "Increasingly, we see references to software, ontologies, project websites, presentations, blogs, videos, tweets, etc. Such resources are usually referenced by means of their HTTP URI as they exist on the web at large. These HTTP URIs allow for immediate access on the web, but also introduce one of the detrimental characteristics of the web to scholarly communication: reference rot." Reference rot is a combination of two problems common for URI references:
  • link rot: A URI ceases to exist; the page is not found
  • content drift: The resource identified by its URI changes over time and is not what was originally referenced
Their study shows that articles published in 2012 suffer from link rot: 13% of arXiv, 22% of Elsevier, and 14% of PubMed Central. For articles published in 2005, the numbers are higher: corresponding numbers are 18%, 41%, and 36%.

The typical strategy to address the problem is to link to a snapshot of the web page (instead of the original web page) created at the time and stored in a web archive, such as the Internet Archive, archive.today, and perma.cc.

 There are problems with the approach. The link copy may not remain in place either. The linking URI is lost, as is the any information about the page or changed page. Link decoration can be used, with the URI of the original, the snapshot, and datetime of linking. Memento can provide this information but there are discussions needed to decide how to best convey the information.

No comments: