[XeTeX] Mixed Roman and Indian alphabets for Sanskrit

[Javier Bezos]

    ​
> \fontspec[Language=Hindi,Script=Devanagari]{FontName}
> \fontspec[Language=Sanskrit,Script=Devanagari]{FontName}
>
> \fontspec[Language=Sanskrit,Script=LatinScript]{FontName}
>
> \fontspec[Language=Sanskrit,Script=Sarada]{FontName}
> \fontspec[Language=Sanskrit,Script=Grantha]{FontName}
> \fontspec[Language=Sanskrit,Script=BanglaLipi]{FontName}

> * Saying \selectlanguage{sanskrit} isn't going to be any use in my
> document, because "{Sanskrit}" could be any of several scripts.

Yes, I understand your point. Different language/script pairs
will be treated like separate languages, much like bcp47 and the
CLDR do - eg, sa-Deva, sa-Gran, and so on. What "sanskrit" (as babel
language) means should be user definable.

> "Bangla" and "Latin" are awkward, like "Arabic,"  because they're the
> names of both a language and a script.  "Bangla lipi" just means
> "Bengali writing," and is what Bengalis say when they specifically refer
> to the script.

Script names will be visible only when necessary and in specific
contexts, like \setbabelfont[*devanagari] (here * meaning a script;
tentative syntax). I think

\fontspec[Language=Latin,Script=Latin]{FontName}

poses no problem at all.

Javier