>I can think of cases I've seen where paragraphs have not been handled properly; but I have no idea if they are the same as yours. As it stands, your assertion about broken paragraphs wouldn't be helpful for me as a text provider because I probably think I have marked all my paragraphs properly; you may disagree and you may be right; but there's no information I can use to fix what's wrong.

Historically, there are three ways paragraphs have been formatting: 1) the highest quality books used the formatting Hart followed: no first line indent, line between. 2) medium quality books used: first line indent, no line between.  And 3) the lowest quality books used: no indent and no line between, paragraphs demarcated only by the ragged right of the last line.  The PG sausage-making chain serves paragraphs in many ways, but often “belt-and-suspenders-and-suspenders” where what the end reader sees is: line between AND first line indent AND yet another line between “just for good luck.”  Even this isn’t the end of the world for books with very long paragraphs, but for books with lots of short dialogs it is deadly.  And *why* do this in the first place?

>I don't understand why we are manually constructing tables of contents in the first place. All the chapters need to be located; all the chapter headings need to be identified and positioned; and all it takes is some software to create a guaranteed accurate TOC from the text.

I use such software.  It is quite successful and helps me make and proof complete and useful books.  Unfortunately PG doesn’t support it, requiring instead the “partial solutions” that HTML gives us.

3) Blockquotes. Similar issues to paragraphs.

>Same comment as paragraphs. A problem I see is that some things are marked as block quotes that aren't block quotes, simply because they are formatted like block quotes.

This again is arguing semantics rather than typesetting. Those of us who see “typesetting” would claim that the HTML labels don’t seriously represent semantic labels in the first place, but rather as simple mnemonics to refer to a particular formatting option.

5) Covers.  Every book should have them, and they ought to be easy to implement.

>Shouldn't they be easy to auto-generate from metadata?

Yes if we really had metadata, and not the current rag-tag collection of info dragged hap-hazardly from the four corners of the world.

6) Title page.  Again, pretty much every book should have them, and they should be easy to implement.  The title page info should not have to also be submitted redundantly 12 other places.

>I don't understand the "12 other places" comment.

Copyright clearance, and submission, and “title page”, and metadata, and file names, and WW input, etc.

7) PG Boilerplate.  Should be implemented in an attractive and non-obtrusive manner, which does not scare off the readers, nor make PG look like idiots, and should be written in such a manner as to convince most readers that the boilerplate is actually a good thing to their advantage.

>Agreed. But Greg needs to agree and commit to do something about it.

Well, realistically Greg would need to agree and commit to do anything.

8) Statement that this book is “risen to the public domain” and what that means. The implication that PG is giving away this book is false, because the book is not PG’s to give away.  Rather, the book belongs to the public in the first place.

>Probably true but I'm not sure who would disagree, or what the significance is.

OK but currently PG does not say that a book is “risen to the public domain.”