> Ah. So the reason otf -> pfa usually works is because most fonts
> contain "regular" character outlines that are just packaged as CFF,
> without actually making use of the new features.
Yes, AFAIK. Actually, Type 1 fonts can contain *any* PS code (and
some fonts like `Optima' actually make use of that), but many parsers
like FreeType don't have a built-in PS interpreter and thus won't
understand the additional PS stuff.
> (Of course the conversion loses all the OpenType tables, but this
> can probably be made up for by extra intelligence in the typesetting
> program.)
The OpenType tables are not used in PS or PDF fonts and get thus
stripped off completely.
>> Using -dbackend=eps, you don't need this since a bounding box
>> gets emitted.
>
> I noticed that for very short fragments the bounding box is wider
> (it includes the whole width of the line) than the actual output,
> which is probably useful if you import several pieces and want them
> all nicely aligned...
IIRC, this can be adjusted too, but I forgot how to do that.
Werner
