the newest iteration over at distributed proofreaders will be _3_ proofing rounds and 2 formatting rounds, with provisions for skipping some of these rounds... with this new change, i think we can safely say that d.p. has wasted a lot of time studying its workflow and _still_ not come to the point of perfecting it... so it's time for me to once again interject my opinion. 1. pre-proofing clean-up programs could handle _many_ of the problems that are found in your o.c.r. (careful image handling could solve most of the rest.) 2. if d.p. used zen markup, it could save itself from the drudgery of those "formatting rounds". conversion from plain-ascii to html is now routine. (pushing out each page to check its formatting is a tremendous waste of bandwidth. but who cares?) 3. no matter how many rounds you add, it will _still_ be the case that some pages will have needed more. (some _pages_, *not* some _books_; it's silly to treat all of the pages in a book as being of equal difficulty.) d.p. needs to go "roundless", treat pages individually. 4. duplicate proofings by independent proofers can be crosschecked to quickly and easily spot any differences, which can then be dispatched with a minimum of effort. this double-key strategy can be used on individual pages. again, these are all things that i've been saying for years. if all the energy that's been spent on "research" would've been used to implement these recommendations instead, it would've been a lot less work, and d.p. would now have a good workflow. as it is, it will probably take a year or so for the problems in the newest system to reveal themselves, and then more work after that to install all of my suggestions. -bowerbird