Teaching the use of Automated Text Recognition online. Ad fontes goes ATR

Ad fontes, a popular e-learning platform for handling archival sources provided by the University of Zurich, is currently being expanded to include Automated Text Recognition (ATR) modules, enhancing its wealth of already available exercises, resources, and introductions. At the conference, we will present our ideas for this ATR teaching module. Working with automatically recognized text is complex and requires much effort from interested researchers, who often must familiarise themselves with the subject on their own. To reduce this initial workload, the DIZH-funded project PATT is now developing a new Ad fontes module to train students, researchers, and other interested parties (e.g., citizen scientists) in the use of ATR, looking at this new set of tools in combination with conventional qualitative methods of source processing. ATR has significantly improved in recent years. However, many non-standardised fonts and layouts still result in high error rates, which is why it is essential to clean up the data by hand and reflect on how and when to use ATR, as well as to get an idea of where to look for potential errors. The transcription practice and, more general, the work with archival sources today occurs at the intersection of analogue and digital skills. An ATR teaching module, therefore, must not only explain how and when to use automated recognition the most effectively but also teach techniques on how to proceed with the ATR output and how to reflect on data produced and the process used.