Let's scrap it. I leave in page numbers and breaks but take out the
author's name and the title at the top of pages carefully letter by letter
as I go. The messiness at the top makes bookshare books come across as
messy output.
I think we can learn to do this well so the stripper is not
necessary. Obsessing about it means validating books takes much too long
as witness the discussion on the list. If we can get the ocr package
people to get the remove headers to work that would be good. What is the
judgement about the one in version 9?
E.