[XeTeX] Separate fonts for Latin and Greek with fontspec?

BPJ bpj at melroch.se
Fri Sep 17 17:53:15 CEST 2010


2010-09-17 13:14, Michiel Kamermans skrev:
>> Too bad BTW that "Phonetics" isn't definable as an informal group!
>
> Indeed, but that exposes the inherent problem of multiple
> languages using scripts that overlap with other languages...
[snip]
> Unicode went with "scripts", but that's basically equivalent to
> doing what Polyglossia does, where you indicate what the following
> stretch of text is. It works great, but you have to put in more
> commands, and the source becomes hard to read if you switch
> languages several times on a page, or within a paragraph.

Well, the error was for Unicode to assume that IPA <a>
be identical to Latin <a> while being perfectly happy
that Cyrillic <a> be different.  The IPA a-z (and the
IPA βθχ) have one important property which sets them
off from their Latin (and Greek) counterparts (quite
apart from details of shape in the case of β), namely
that they don't capitalize! This bit me once when a
snippet of IPA got wrongly capitalized in a page
heading.  This is not a problem if you *know* that your
snippet of text contains IPA but any text input which
is going to be handled programmatically must be
tagged/marked up and the program made to look for the
tagging and take measures to apply the right casing
behavior.

I've made a habit of using and defining an \IPA command
-- usually just defined as \let\IPA\MakeLowercase --,
and similarly using an IPA class with proper style
definitions when producing HTML, and using it
consistently just in case I should later apply some
formatting to surrounding text which is unwanted for
the phonetics.

/bpj


More information about the XeTeX mailing list