The transcription of mediaeval handwritings is a cumbersome and challenging task, but a prerequisite for analysing the contents of those handwritings. The usual approach consists in employing a conventional text processing system which is not well suited for this task. Being the most difficult and time consuming piece of work, the analysis itself is done manually.
The most critical part of the transcription is the subjective nature of the resulting transcript. Instead of relying on the individual judgement of an editor who is transcribing the writing, objective criteria would be important, underpinning the final transcript. Those criteria would help to deal with more sophisticated challenges, like the identification of changes concerning the writers within single texts.
This project investigates computer aided transcriptions which provide the basis for comparing glyphs, identifying differences and similarities of a handwriting, and supporting the different steps within the transcription process. As opposed to conventional document processing and optical character recognition systems, which fail when applied to arbitrary complex handwritings, the present idea consists in the identification of transcription sub-routines which would adapt to the editor’s view of the entire transcription process, supporting him whenever appropriate and possible. This becomes of particular importance when following the character based analysis paradigm.