|
What I tought about was the possibility of a python script (because Fontforge is accepting it) which scans an epub for all used chars + variants (as mentioned smallcaps monospace naumerals and ligatures) per every used font passes said charlist to Fontforge which throws everything not on list out of the fonts. The script renames the stripped fontfiles and the replaces the still full-char fonts in the epub with the stripped ones.
Combined with the spec-compliant font encryption as shown on MR by pdurrant this would be a solution not only to save space but also to embed commercial fonts in an epub without distributing whole fontfiles, which should be in accordance with the licenses, being kind of the same thing that's done when you do embedding in PDFs
|