On 2013-06-01, Jürgen Spitzmüller wrote:
> Uwe Stöhr wrote:
>> We had this topic recently on the list and also in stackexchange. LuaTeX is
>> not able to handle RTL languages properly, no matter if polyglossia is
>> used. There is an experimental package called luabidi but this is not even
>> in beta state according to their developers.
> Well, I browsed the Internet as well, and I figured several people use
> LuaLaTeX succesfully for Hebrew.
>> > It's an alphabetic script. Not every alphabetic script is Roman (e.g.
>> > Cyrillic, Hebrew, Greek).
>> They defined Greek as non-Roman because some words require the change of
>> characters. So for example Fraktur is in their definition also non-Roman
>> because the "s" is a different character depending on the position in the
>> word. Maybe the name non-Roman is misleading but they did not say
>> non-Latin.
> Could you please point me to that discussion? I've read dozens of books
> about script theory and history, and I've never came across a
> categorization of this kind. It strikes me highly odd. In any case,
> this is not the common meaning of the term.
I'd prefer, not to use the term "Roman" in with this meaning in our
documentation.
a) it is easily mixed up with "Latin" (where Greek and Fraktur are not part
of the game).
b) "Roman" is widely used to denote a serif typeface or a serif-upright
typeface.
...
> I would have gotten your point faster if you would have said: "We need
> to implement a way to set a default output format for non-TeX fonts"
> rather than stating "We set the wrong default output format for non-TeX
> fonts".
However, I understood "we *use* the wrong default output format ..."
without mentioning it is *set* somewhere.
This is the problem with "default value" meaning "the value used if
nothing else is configured" ("Standardverhalten in Ermangelung von ...")
as well as "the standard config value" ("Voreinstellung").
