[Booki-dev] epub as data exchange format

[Douglas Bagnall]

adam hyde wrote:


> would it be sane to keep epub as the data exchange format for imports?
> 

Not quite.  Real world epubs have a characteristics that don't suit
Booki, and Booki can produce documents that don't natively suit epub.

Epubs:
* can't have xhtml files bigger than 300k (for e.g. Sony reader)
* have files with arbitrary paths
* have no necessary relationship between files and chapters
* must be well formed XHTML
* have indexes/metadata scattered across arbitrarily named XML files
* can contain all kinds of content that Booki doesn't support.

Booki:
* has a 1 to 1 chapter/file relationship (so files might be big)
* has constrained paths for files
* has limited content types
* uses editors that can produces dodgy XHTML
* wants easily accessible/editable contents and metadata

Perhaps there's more that I've forgotten. So while epubs and booki-zips
are superficially very similar, underneath they are shaped by quite
different constraints.  If Booki and Objavi were to talk to each other
via epub they would each have to incorporate the whole of Espri and the
epub part of Objavi.  I don't think it would be worth the trouble.

Douglas