Thursday, January 10, 2008
Archiving digital copies of electronic journals
There was a time, back in the 'Middle Ages', where monks would cross the continent of Europe to study and copy precious manuscripts. Umberto Eco captures this world in his novel 'The Name of the Rose'. In making an accurate copy and taking it back home with them, the monks were both distributing and preserving the work. Having a copy of a work in one's archive made your institution more prestigious, and helped to preserve the work from a variety of medieval threats. But not everything is copied everywhere, use of manuscripts is also a cause of them wearing out and becoming unusable for future readers.
Later scholars may puzzle over the complex social model, the reciprocal relationships, the transmission history of any individual manuscript and its relation to the original work. Which manuscripts have integrity as authentic copies? What 'trust' should be placed in a Saxon Manuscript found in the collection of a monastery in Lombardy? Where two or more copies exist detailed comparisons would need to take place to remove doubt about integrity: (all copies agree on every word of chapter 1, but there are variant readings of sentences in chapter 3).
Fast forward to the present and similar problems of integrity when it comes to preserving electronic material, such as the articles published in electronic journals. There are various schemes which libraries might look to or participate in and some are based on very different models. Locking files away in a safe place for ever, or until they are the only remaining copies is the strategy of schemes like Portico. Libraries might support this model as insurance against future loss of access.
Another model is the foundation of the LOCKSS strategy. Here the emphasis is on maintaining the integrity of a widespread network of copies. There is not just the safety feature of having multiple copies stored as a defence against the many threats of the modern age. It also enables on ongoing work of comparison to take place, ensuring that variant readings of texts do not get a chance to develop.
Comparing electronic objects to medieval manuscripts will only get you so far in thinking about digital preservation, but considering the social model behind any scheme is an important way of deciding between there relative claims. There is more to digital preservation than an effective backup and disaster recovery program. The threats faced by digital materials include format obsolescence (no viewers for certain file types), temporary or permanent lack of access to original versions (the publishers and their archives have 'gone away'), degradation of media (no one can play 12 inch laser discs anymore) or digital degradation (the saved bitstream is 'corrupt' and will not load), lack of context (the metadata describing the object is no longer attached).
LOCKSS may not have all the problems of digital preservation solved, but it does have some advantages for libraries. There is a sense of taking up the responsibility for preservation and keeping it within your control, it relies on open source software with a track record, it allows librarians to decide for themselves on their collection development policy for preservation.
Some UK Libraries, including De Montfort University, have been involved in testing LOCKSS. As this Pilot comes to a close in February 2008 libraries will be deciding on whether and how to take their participation forward. It will be interesting to see which technology for preservation, and which social model is behind the strategies that do get adopted.
SEADLE, M. (2006). A Social Model for Archiving Digital Serials: LOCKSS. Serials Review, 32(2), 73-77. DOI: 10.1016/j.serrev.2006.03.007