 |
 |
Let me echo Jim in suggesting a transcription tool rather than OCR for handwritten texts. However, a lot depends on the kinds of material you're working with and the uses you plan for the transcripts. Is it structured data, like census records, account books, or an index cards database? Is it free-form text like diaries or letters? Does the text contain a lot of genetic elements like strike-throughs, careted insertions and marginalia? Do you want to index terms so that readers can view all mentions of banjos within the text? At present, there is no one tool that supports all of these. I built and maintain one (AGPL) tool for free-form text to be used in indexing [Self-promotion: http://fromthepage.com/ is the tool; source is at http://github.com/benwbrum/fromthepage/ ] and have spent the last year building another (Apache) tool for converting tabular records into a search database. I think they're great, and am really excited about them both. Nevertheless, last week I pointed a project at Jim's T-PEN instead of my own tools, because the manuscripts were medieval Arabic donation records which needed line-based transcription. I maintain a list of transcription tools used in crowdsourcing projects here: http://tinyurl.com/TranscriptionToolGDoc Currently there are around 30 that I know of, and I'd be happy to give my opinion of what's appropriate for your project on or off list. Ben Brumfield http://manuscripttranscription.blogspot.com/