Quote:
Originally Posted by Doitsu
Using unmanifested files with KindleGen might work, but you're not following best practices; it'll also ensure that an epub book assembled from these files will be invalid.
|
It does work and is documented in the “official” examples, but wasn’t that the point: me giving an example of something valid for KindleGen that produces an invalid ePub?
Quote:
Originally Posted by Doitsu
Even though most epub2 apps will ignore the guide section, adding the guide items recommended by Amazon won't invalidate an epub book, because they're a part of the epub2 specs. BTW, Sigil has a dedicated menu option for adding guide items (Add Semantics...). |
True, but KindleGen makes certain assumptions based on the <guide> section in the OPF.
Quote:
Originally Posted by Doitsu
Epub2 books are supposed to contain utf-8 or utf-16 encoded (X)HTML files and, AFAIK, KindleGen does not require deprecated character encoding declarations.
|
Character encoding in HTML5 is done with <meta charset="UTF-8">, but KindleGen still requires the depreciated HTML 4.01 method of <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">, even though the file is read as HTML5 if defined as such.
Quote:
Originally Posted by Doitsu
In fact, according to section 2.2.1 of the Kindle Publishing Guidelines, Amazon recommends using valid epub books.
|
Not in version 2017.4 of that document.
------
I must say this thread has become tedious. After five pages nothing constructive has been said about creating mobi files - I do not count pushing automated software for a round-about method as a viable solution. However, I've taught myself how to do it using source files fed directly into KindleGen by reading the Publishing Guidelines and perusing the myriad example provided by Amazon.
