Re: bug? with texi2dvi and udotaccent in section

[Dumas Patrice]

Hi,

I disclaim that I am quite ignorant about all that but it never hurts to 
give an opinion...

> I fear that this is all leading nowhere, however.  Even if we somehow
> special cased the accent commands so that they were really supported,
> the next problem would be that hyphenation didn't work.  And I feel sure
> there are other problems as well.

Maybe, but all that is solved in TeX, ie with stuff like \'e you can 
create absolutely correct documents, thus chances are it can be
solved here too. 
 
> The only real way to support languages other than English is to allow
> real 8-bit input, I think, so preaccented letters can be used.  The
> existing accent commands are just Knuth's kludge to allow accents in
> occasional foreign words in English.

I don't agree totally, as 8 bits characters are really hard to deal with 
because they can be encoded in a lot of fashions. If you produce a 
document in a given 8 bit locale, it will be unreadable with another 
locale. You could specify the encoding within the document and let the 
formating tools retrieve the informations, but I have no idea whether 
it is harder to implement or easier than the kludges which allows to 
use only well normalized 7 bit ascii. It seems that latex can handle 
both cases (with \'e for 7 bits and \usepackage[latin1]{inputenc} for
use of 8 bits, here encoded with latin1)..

In my opinion using UTF8 would be an even superior approach because
it avoids the pitfalls of 7 bits (not enough characters) and 8 bits
(multiple encodings).

Anyway the required work may be different for the different approaches,
and I have no idea about what is easier/harder.

Pat