Hardware and software recommendations for OCR on a book series

Hi All,

I'd like to scan in a series of novels and convert them to text for analysis
on the PC (occurance of rare words etc). I can pick up PDFs from
peer-to-peer file share applications like Kazaa, but the OCR has introduced
errors into them (especially for words that span lines for ex-
ample that one).

Do you have any recommendations for accurate fast scanners, and for accurate
fast OCR software? I did check the FAQ at
http://www.cfar.umd.edu/~kia/ocr-faq.html
but it was last updated in March 1997!

Ideally, I'd like a "money's no object" recommendation and a "budget"
recommendation.

Cheers,

Tim.