Did some programming for analysis of the Middle English Dialect project way back when. This was based on corpus of 320 words as they were spelt by scribes in different parts of the UK. Often the same scribe would spell a word 3 or 4 different ways in the same document. But there were regional variations etc. Fascinating topic. Helped to draw maps like:

http://www.lel.ed.ac.uk/ihd/elalme/elalme_frames.html

http://www.amc.lel.ed.ac.uk/?page_id=563