Embed book identification information in index header

[strk]

Maybe as a comment, but would be also nice for the comment to be machine-parseable.
Useful information would be:

• Number of pages in the book
• A set of MD5 sums of known versions of the book (MD5 changes with resolution/size..)

[aspiers]

I think this could be covered by an index.csv or (perhaps better) index.yaml file which lists all the books. Fields should include:

• book name
• publisher?
• number of pages
• checksum

If variations of the same book are referenced by different indices, then each book should have its own entry.

[aspiers]

That said, I really think that #39 is the best way to go here, rather than switching to YAML / XML or some other format.