On Sat, Jun 22, 2002 at 01:25:50AM +0300, Jarkko Hietaniemi wrote:
> > > Actually, how about looking in locale? That's the storehouse for
> > > preferences like this, isn't it?
> > > 
> > Hm, don't know, maybe. I had some vague memory that the week numbering
> > was part of the timezone info, but I may be wrong there. Someone will
> > probably inform about this.
> 
> No, it's not, and in general I think locales should be avoided like
> plague.  While useful for some rough adjustments they are too blunt
> a tool for small cultural adjustments like this.

Clarifications: when I say "it's not", I am shaving corners, so to
speak...  it *could* be.  As far as I know there is no law/standard
*forbidding* a system from doing so.  But therein lies the greatest
problem of locales, in my opinion: there is no one single definition
for what locale variable affects what and what should be the results.
I'm a cross-UNIX geek, and what I get from Finnish locale settings is
different in every box.

Locales are a grab bag of cultural settings (date/time format/language,
sorting order, what characters are considered letters/digits/...,
application message language, character set and encoding, and so forth),
and when they work it's nice, but when they don't, I prefer LC_ALL=C,
at least that's rather well proven and tested.

-- 
$jhi++; # http://www.iki.fi/jhi/
        # There is this special biologist word we use for 'stable'.
        # It is 'dead'. -- Jack Cohen