Hello, Nicolas Goaziou wrote: >> It changes behavior for your setup in that you can define a LANGUAGE >> that isn't a known abbreviation. > > While I think your patch is overall an improvement, I'm not convinced by > this particular point. Indeed #+LANGUAGE: expects a language code as > value, not just any string. This is important since latex backend is not > the only one to use that keyword. > > For example, imagine a user in need for german smart quotes. How do you > explain to him than #+language: german will not work, but #+language: de > will? > > I think special Babel needs can be handled elsewhere.
That makes me think of another point: how do I say that the HTML exported document is in American English? By having a #+LANGUAGE: en-us specification. But that's not a valid language for Babel when exporting to LaTeX. So, I think we'd need some sort of alist with allowed values, and their equivalent for LaTeX and for HTML (and for other uses?), don't we? Best regards, Seb -- Sebastien Vauban
