Researchers in Israel are claiming that they have developed a computer algorithm capable of identifying words, letters, and handwriting styles for use with ancient manuscripts. It is supposed to “learn” over time, gaining precision with each manuscript it is fed, while establishing a “google”-like searchable catalogue of manuscripts. The algorithm also has a predictive feature to help fill in gaps or faded parts of characters and words.
It should be interesting to see how effective this is as a technique; one would think that it could be a decent tool in conjunction with palaeographic experts who could correct it (having done a little manuscript work myself, I’m skeptical of how accurate this will be, but I’m curious). Obviously this technology will not be perfect, but it’s a good idea.