On 31/01/2019, Doug McIlroy <doug(a)cs.dartmouth.edu> wrote:
I OCR'd a sizable manuscript written on a pretty
shabby portable
typewriter.
I scanned each page twice, making sure to move the paper between scans.
Then I ran both diff (by words, not lines) and spell to smoke out trouble.
The word list for a program listing is quite short and easy to generate.
(Print a list of all the apparent words and visually eliminate the
nonsense.)
And a spell check is an easy pipeline of standard utilities.
doug
Very nice! (I shall remember this technique.)
N.