[XeTeX] Review of OpenType font features and language fitness

Andy Lin kiryen at gmail.com
Sat Jan 2 01:30:48 CET 2010


Hi Kārlis,
Your work is going to be quite useful for a lot of people!

c2sc does not provide faked small capitals. Rather, it converts
uppercase letters to small capitals, in contrast to the behaviour of
smcp, which is to convert lowercase letters to small capitals. You may
also wish to check the behaviour of numbers with small capitals; some
fonts will change the number style to ranging (old style), even if the
default is lining.

nalt is more common in CJK fonts, and you should test it with numbers.
It usually produces an enclosed version of the glyph (in a box, in a
circle, in parentheses, etc.).

mark positioning and mark to mark features are used with diacritics,
as noted by Khaled, you won't find it has any effect on punctuation
marks.

When you list variants (e.g. on p13), the variant 1 appears to begin
one space to the left compared to variants 2, 3, and 4.

If you're using fontspec, you should be aware that
superscript/subscript will be faked if the proper glyphs are not
present in the font. Thus, you may have false positives for
superscript support if you rely on pdf output alone; otfinfo will give
you the correct information here. I wonder if you've read the fontspec
manual, as it details the effects various Opentype features have, and
also demonstrates some fonts you can investigate. Some fonts will also
have an online specimen sheet, which might help you in tracking down
the existence of ligatures/variants present in the font.

You might find the list at http://www.wazu.jp/ to be interesting,
though the emphasis there is Unicode coverage, not Opentype features.

-Andy


More information about the XeTeX mailing list