Quote:
Originally Posted by lkws1
When you send an ePUB file more often than not Amazon won't be able to put it on the Kindle because it needs to convert it into KFX (I think that's what's happening in the background)...
|
Actually it is converting to Master-MOBI (providing both MOBI and KF8 formats) using Amazon's Kindlegen software. So far Amazon is reserving KFX, and the associated Enhanced Typesetting features it provides, to published books.
Quote:
Originally Posted by lkws1
...but it fails for a lot of different reasons (which it never tells you).
|
Kindlegen is a publishing tool and is designed to be intolerant of errors in the source EPUB. It would be nice for those of us who understand the inner workings of EPUB if Amazon would provide the actual error log that Kindlegen produces when rejecting a book. However I don't see that happening as it would just be confusing to most customers.
If you want to know why a book was rejected you can use Amazon's Kindle Previewer software to find out. The most common reason I have seen is a bad table of contents. Amazon's publishing guidelines require the TOC to have working links that point to non-hidden content and entries that follow the same sequence as the book's content.
Quote:
Originally Posted by lkws1
The 30% of time that it works, there are still formatting errors and it's overall a worse experience.
|
The main cause of formatting errors is lack of a declaration of the character encoding being used in the EPUB. The default for EPUB is supposed to be UTF-8 but for historical reasons Amazon's conversion software assumes Latin-1 encoding which messes up characters such as curly quotes. The Modify EPUB plugin for calibre has a function to insert the proper encoding to fix this problem.
It would be nice if there was a calibre plugin that would deal with these issues more automatically.