see how easy it is to change the subject-header? *** don said:
When DP started, here was the basic process
the irony was, back in those olden days, it was actually much more difficult to digitize a text, because the o.c.r. was horrific, and thus it was a pure pain to proof. nowadays, even though o.c.r. is vastly improved, it seems to take forever for a book to transit d.p. here's an illustrative datapoint i just churned... in one of the books that rfrank is using for his roundless experiment, even tepid preprocessing (which is what he practices) combined with o.c.r. to produce 20% of the book's 240 pages perfectly. another 30% of the pages had only 1 error on 'em. and most of the errors failed spellcheck, meaning they could've been isolated and fixed immediately, without need of a word-by-word proofing modality. d.p. uses dozens of volunteers, taking hours of time, to do something that one person can do in one hour. which would, you know, ordinarily be a very sad thing. except what makes it funny, in this particular case, is that the people at d.p. think they're being "efficient"... -bowerbird