With all your help I managed to get my submission, A Study Of The Bhagavata Purana Or Esoteric Hinduism into pretty good shape.  This is the latest:

http://epubmaker.pglaf.org/cache/20120406103034/

The book is such a big and complex project that rather than just inflict the whole thing on some unfortunate white washer I'd like to get some comments from anyone willing to give it a quick look first.  I expect that the way I handled family trees (and this book has many) will be controversial.  If I get a consensus that the book doesn't suck (or if I don't get any feedback at all) I'll submit it.

I would be especially interested in Bowerbird's opinion on this.  No, really.

Bowerbird gave me some good advice early on with this book, and while I did not take all of it it did lead me to try RST, which I am convinced saved me a ton of work.  I like RST better than the ZML that he had originally suggested.  My reason is that while ZML documents look like they aren't marked up at all, that is not necessarily an advantage.  For instance, to make a heading in ZML it sounds like you have to insert some number of blank lines before the text.  Counting blank lines is something I'd rather not deal with.  RST lets you underline the heading with something and not worry about blank lines before or after.  The RST converter deals with that.

I also learned that you don't re-wrap your text before proofing.  You let the converter deal with re-wrapping.  Next time I'll know better.

Thanks,

James Simmons