jim said:
Not that I totally disagree, but when you take out the easy stuff, the stuff that's left is harder to find.
where's your data that validates that?
Simply my personal experience finishing a book taking this exact approach. When you take out the easy stuff, you are left with the stuff that typically the P2s and P3s find -- hopefully! When you do the machine marking of errors you double the amount of errors that need to be fixed -- since each input file contributes its errors. But now many of the "P1" type errors are marked, which could be a win if the P1 people were presented with a highlighted version of those errors, similar to what is currently presented in WordCheck. To me a surprising amount of work goes into fixing hyphenation/dehyphenation linebreak errors, where it should be possible to make a more intelligent tool to fix most of these problems -- or does someone claim to already have such an intelligent tool? But again, when I'm doing a "solo" project, I don't really spend that much of my total effort on a "P1" type pass to begin with, and a surprisingly large amount of time and effort goes into doing "PP" stuff -- including supporting that GD PG "txt70" format!