omk3 it doesn't need to go as far as a
completely non-latin set of glyphs.
(say greek, hebrew, or cyryllic)
already latin extended are not fully supported
Since full (or even mostly) unicode covering font files are rather NOT small the need of including them into the epubs bloats them as hell.
(Adobe sells a
full unicode arial variant: the filesize is ca 24MB)
A tool which:
- analyzes the epub
- makes a list of glyphs used from each fontfile
- deletes the remaining ones from said files inside the epub
would be just great
a PDF embeds only the characters needed;
having such a stripping tool would be great for epub optimizing.
but regardless of that Val is completely right: it's a rather poor behaviour by a company so deeply specialized in typography offering a reader app with such a poor set of glyph coverage, be it even as a fallback solution for not self-typeset epubs.