Versions

Further information

Version Identification Framework

VERSIONS Project (including user study, November 2007)

As repositories have grown in size and scope, the problems caused by versioning have become more prominent. It is now an accepted problem; only 5% of academics and 6.5% of information professionals find it easy to identify versions of digital objects within institutional repositories. Across multiple repositories this becomes 1.8% and 1.1% respectively. Version identification is not about identifying the 'best' or 'right' version of an object. The Versions Identification Framework (VIF) aims to make the version status of a research object and the relationships between linked objects clear to end-users so they can identify which is the most appropriate version for their purpose.

Draft Version

Early version circulated as work in progress

Submitted Version

The version that has been submitted to a journal for peer review

Accepted Version

Author-created version incorporating referee comments and is accepted for publication

Published Version

The publisher-created published version

Updated Version

Updated since publication

Incorporating different versions of learning and teaching materials proves equally challenging, presenting a barrier to deposit: many academics are concerned that people may use earlier versions of their content that may not still be accurate (this is more apparent for those subject disciplines where course content regularly changes and particularly if old information could be dangerous - as in medical disciplines). Version policies which recognise these fears and provide clear guidelines around deposit, updating and using content are essential to gain trust and ongoing deposit. A learning and teaching repository could chose to hold only the most recent version, or could hold previous versions only for legacy purposes and not make these discoverable. This relies on good communication and clear workflows to ensure that new versions are deposited appropriately.

Versioning problems are minimised if some or all of 5 key pieces of information that exist at the time of creation of the object are captured, defined and made sufficiently visible to the user, to a deposit mediator or made machine-readable:

• Defined dates
• Identifiers
• Version numbering
• Version labels/taxonomies
• Text description

It is important to include versioning information both in metadata and within an object itself and one of the following solutions should be used systematically to store version information within a repository:

• The filename
• A watermark
• A coversheet - link to coversheet page
• An ID tag or property field

Advising academic authors about versions

Academic researchers typically produce many revisions and versions of a research output, up to 60 in some cases, and it increases when working with co-authors. Authors vary in how many of these versions they keep: either all, milestone only, or just the latest. They tend to want to limit the number of versions disseminated.

A key message for repository managers to convey is that, because of standard agreements between authors and publishers, keeping only the Published Version is likely to limit the author's ability to make their work accessible in future. As a minimum, keep milestone versions like an author's Accepted Version. Half of authors are not satisfied with the way they manage their own personal collections of digital objects. An institutional repository provides academic staff with a managed environment to deposit their milestone versions, where they can be easily located and retrieved. 81% of authors surveyed stated they would deposit their final Accepted Versions of journal articles in an institutional repository 'if invited to do so'. Adding version information to authors' papers, with clear links to published versions, helps allay concerns about the quality of accepted versions and about potential loss of downloads for published versions. Institutional repository managers can help inform academic authors about publishers' access policies. Researchers remain uncertain and unaware of possible negotiable alternatives. SHERPA RoMEO and SPARC are two excellent resources for providing detailed advice to authors on 'permitted' versions.

The VERSIONS toolkit offers 5 top hints for authors, derived from feedback by active researchers:

• Plan how personal versions of files will be stored and named
• Keep separate, permanent author-created Submitted Versions and final author-created, Accepted Versions of research publications
• Add the completion date to the first page of any versions, especially milestone versions
• Consider carefully how to disseminate work before signing any agreements with publishers
• Keep a copy of signed agreements. Deposit your work in an open access repository and guide readers to latest and published versions