Page:Wikipedia and Academic Libraries.djvu/292

This page has been validated.
Wikisource as a Tool for OCR Transcription Correction
279

An opportunity to explore this workflow suddenly arose in March 2020, when the Library closed its doors as the United Kingdom entered lockdown in response to the COVID-19 crisis. The Library’s ten-person digitization team, whose work almost exclusively required them to be on-site using cameras and scanners to digitize books and other items from the Library’s collections, needed work that would have impact and advance projects, would be large enough to keep them occupied throughout lockdown, and would require minimal access to the Library’s network or physical building space. This unique situation allowed the Library to test Wikisource at scale; within twenty-hour hours of lockdown the entire team was editing transcriptions for the Library’s recently digitized Scottish chapbook collection (Hagan, 2016) on the platform.

Chapbooks are small printed booklets that were sold for a penny or less on the streets, at fairs, and at markets. Typified by the use of woodblock illustrations and covering a range of subjects such as murders, disasters, love stories, and biographies of famous people, they were staple reading material for much of the population in a time before modern communication systems were invented (Hagan, 2019). This depth of content makes them a particularly useful primary source for social historians of the period, while their engaging content matter and size—3,000 separate books ranging from eight to twenty-four pages in length—meant it was an ideal collection for exploring Wikisource as a tool for OCR correction. Within a few days of lockdown, it became apparent that there were several other members of Library staff who also had time to work on the Wikisource project. Like the digitization team, staff in roles that were public facing or required access to the Library building were also limited in what they could do while working from home. By the end of March there were over fifty members of staff taking part in the work; this number increased to seventy, which was over 20 percent of the entire Library staff (National Library of Scotland, 2020b), at the project’s peak.

It was exciting to have this unexpected, possibly once in a lifetime, opportunity to focus staff resources on a Wikimedia project, but rolling it out to several dozen staff in just a few weeks created many challenges. Arguably the greatest of these was that the Library, and the