Re: BibTeX-mode: Key generation when latin-1 characters appear in author field

[Roland Winkler]

> > Maybe it would be possible to add umlauts and accented
> > characters to the default?
> 
> so that we get `ä' in that list, too. (And all the other
> dots/hats/accents/you-name-it too, of course.) It would be great
> to see this added to the default list. There is of course a whole
> lot of `funny characters', but it would be desirable to have them
> in that list.
> 
> > I don't know what would happen in unibyte mode with stuff like
> > this, though.
> 
> I hope someone can answer that...

I have been thinking about this for a little while. I think, it
goes beyond BibTeX. 

Some time ago, I have written a little package umlaute.el, see
http://www.tfkp.physik.uni-erlangen.de/~winkler/src/umlaute.el 
The idea is to have `translation tables' so that one can go back and
forth between different `representations' of a character, depending
on the emacs mode, for example (I live in southern Germany)

  text-mode:            Grüß Gott
  tex-mode:             Gr\"u{\ss} Gott
  german latex-mode:    Gr"u"s Gott
  html-mode:            Grüß Gott
  other modes (7 bit):  Gruess Gott

BibTeX keys probably should use the 7-bit `representation'. 
For German umlaute the 7-bit `representation' is fairly well defined
- even though I expect that some people would prefer a translation
table that simply drops the double dots, but does not add the extra
`e'. The best choice might depend on the context.
However, I do not know how this could be done for other languages.
Some spanish people use `Señor' -> `Se~nor'. Certainly, this is not
a good choice in the context of BibTeX keys.

I do not know whether such a feature would be useful for other emacs
packages, too, nor do I know how it could best be implemented. In
the meanwhile it is probably best if people customize
bibtex-autokey-name-change-strings.

Roland