Computation and Palaeography: Potentials and Limits

Summary

The Schloss-Dagstuhl Perspectives Workshop on `Computation and Palaeography: Potentials and Limits' focused on the interaction of palaeography, the study of ancient and medieval documents, and computerized tools developed for analysis of digital images in computer vision. During the workshop, the interaction between domain experts from palaeography and computer scientists with computer vision backgrounds has yielded several very clear themes for the future of computerized tools in palaeographic research. Namely,

• Difficulties in communication between palaeographers and computer scientists is a prevailing problem. This is often reflected not only in computerized tools failing to meet the requirements of palaeography practitioners but also in the terminology used by the two disciplines. Better communication should be fostered by joint events and long-term collaborations.
• Computerized palaeographic tools are often black boxes which put the palaeography scholar on one end of the system, only receiving a systems output, with little opportunity to directly influence how the system performs or communicates with it using natural palaeographic terminology. The long-term desire is to have the scholar at the center of the computerized system, allowing interaction and feedback in order to both fine-tune performance and better interpret and communicate results. This is crucial if palaeography is to become a truly evidence based discipline. To this end the use of high-level terminology, natural to palaeography, should be integrated into computerized palaeographic systems.
• Palaeographic data, scarce to begin with, is even more restricted by accessibility and indexing problems, non-standard benchmarking techniques and the lack of accurate meta-data and ground truth information. Multiple opportunities were identified for acquiring data and disseminating it both in the palaeographic research community and outside to the general public.
• Palaeographic research is largely restricted to the domain of experts. Making palaeography accessible to non-experts by using computerized tools has been identified as an effective means of disseminating valuable cultural heritage information while at the same time potentially giving rise to crowdsourcing opportunities, such as those proved successful in other domains.

The manifesto which resulted from this work elaborates on the existing challenges and limitations of the field and details the long-term recommendations that have emerged in the workshop.