[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Apology to german users required in the release notes

Roland Bauerschmidt <roland@hbg-bremen.de> writes:

> Thomas Bushnell BSG wrote:
> >   "The XXX locale for German unfortunately uses a bad way of
> >   representing open quotation marks.  We have retained it this way in
> >   order to preserve compatibility with other Linux distributions, and
> >   we hope that in the future it will be fixed.  Meanwhile, the YYY
> >   locale can be used, which fully supports German with the correct
> >   quotation marks."
> Just to be picky here, "meanwhile" sounds to me as though YYY was an
> inferior solution, which I don't think it is. It implies that one might
> want to switch back to XXX, once that's fixed. I'd suggest the following
> wording instead (or something along those lines):
>   "[...] We suggest that you switch to the YYY locale, which fully
>   supports German with the correct quotation marks, and, using Unicode
>   encoding, has better support for other languages as well."

Mike Stone has concerns that multibyte support may not be up to the
task.  I don't know what his concerns really are, because he hasn't
given details, and the relevant bug reports in coreutils seem on my
brief look to be not terribly serious.  But I have certainly not
attempted any comprehensive look.

(For example the only current one in coreutils that I noticed is
"filenames don't line up in ls output", which some people can't get to
be reproduced at all, and is a cosmetic issue equally with "quotes
look ugly".)

I think it's safest to say that YYY has problems with ugly quotes, and
XXX may have problems which we are unaware of, and let users make up
their own mind about which they prefer.

I'm entirely comfortable with recommending UTF-8; the fact that we
don't guarantee UTF-8 works doesn't carry much weight with me.  We
don't guarantee bugfree packages in general.  We don't guarantee that
single-byte coding systems work either, right?  Indeed, it is true
that there are surely bugs which occur for UTF-8 and not for latin-1.
But those bugs may not matter to users, who might well prefer having
good-looking quotes, and it seems good to just give all the relevant
information so that users can make their own informed choices.

One advantage of advocating UTF-8, of course, is that it will generate
pressure on developers to fix the UTF-8-related bugs which they have
been ignoring on the grounds that "UTF-8 isn't required."  But I only
suggest that in jest.


Reply to: