The Importance of Versioning

A few months ago, I checked up on a digital history project to which I was introduced in my introductory digital history course, Richmond’s Voting America project. It is a great project and a wonderful teaching resource, but I was really interested to see that it had been updated. In addition to some cosmetic changes, since the last time I had visited the project had added the presidential data for 2008 and more analysis of their maps. I had originally explored the website to review it for part of the DH course I was taking. In the review (in December 2009) I noted that I was disappointed the project had not incorporated the 2008 data and it was not clear that the project was going to incorporate this data. My review, which suggested the Voting America project had become outdated after the 2008 election, was now outdated itself.

In the grand scheme of things, my review is not very important. However, it does demonstrate that digital history and the digital humanities must begin to establish standards for publishing, especially versioning. Why should a journal bother to review a digital project, if the review will simply be moot in a year or two? On the other hand, why should digital publishing be beholden to the conventions of print publishing?

Returning to Voting America, the promise/perils of digital publishing are clear. As a project that can add new data every two or four years it has the potential to become a truly great and lasting work of scholarship. At the same time, the temptation to purge analysis that might have been subsequently proved wrong or simply outdated, such as speculation on an upcoming election, is greater than ever.

While the ability to update scholarship to with new data and theories without reprinting the entire work is an important aspect of digital work, scholars must not throw the baby out with the bath water. Mistakes and misinterpretations are what build historiography and subsequent generations of scholars and students ought to have access to these building blocks of the field. Additionally, using and citing digital projects becomes difficult if analysis or other text utilized can simply disappear without any record.

The formalization of publishing of digital work will certainly help maintain versions of digital scholarship, as digital work will be presumably frozen at one stage of the projects development. However, scholars cannot sit back and wait for someone else to establish standards of publishing. Even after digital standards of publishing are more or less established, it would be a shame if scholars waited until publication to share their work. Instead, scholars should take control of preserving their work and provide users with some sort of access to prior versions. By providing access to older versions of text, visualizations, or other narrative and analysis, academics can ensure their digital work stays a useful part of scholarship while still developing, tweaking, and building their project.