I realize that this thread is a little out of date but as I've been doing some book conversions and I've started thinking, always a dangerous thing
.
It might be easiest to keep my books in a zipped file something similar to and oeb container. That is all of the html files, images and the opf zipped to reduce size and stored that way. I used to convert quite a few books this way so that FBReader would read them. Since when you convert a mobi or lit file to imp you start with this structure and with a PG text that is the structure you get just before creatation of the imp file why not just keep it? It's fairly simple to write a utility that can pull the metadata out of the zip file and if you want to go back to another format that is simple too. You could even store the original book file in the archive if you wanted, although that would make the file about twice as big as it would be otherwise.
Any thoughts, or critiques of this method from anyone?