Originally Posted by KevinH
Actually, the mobi format is simply a database file of records and record types that a "compiled" ebook is stored in. It was the original Palm Database File format. It could be used to store any information. They choose the first record to be a header record that points to where to find the records and what type they are. Therefore almost any file format could be stored in a .mobi file. It is really the mobi header record that determines almost everything. The original mobi format was used to store a simplified html 3 dialect with links changed to record offsets, index records for toc, etc. Mobi KF8 uses the same database with extended header record to store epub2 level fragments.
So saying something is not "mobi" is not really correct as you could use a variation of mobi to store almost any compiled set of information.
The shared KindleUnpack source code is a good reference to show how both old html3 pieces and newer epub pieces can be stored in the same binary Plam database structure that we know as Mobi.
|