[Fwd]: Response to Mark's message]

[Martin Duerst]

At 14:56 03/04/09 -0400, John Cowan wrote:
>Martin Duerst scripsit:
>
> > In effect, 'orthography' could go as far as including things such as
> > number formatting or date formatting, and maybe even sorting.
>
>The trouble is that it's not even clear that numbers and dates are tied
>to the language or not.  I want my documents spell-checked as en-us,
>but I like ISO dates rather than bogus American ones.  Must I set my
>"locale" to Japanese?

[others said something similar]

I think I was pushing things a bit. But I still think I'm not totally
wrong. In particular, I think that number/date formatting in running
text is quite a bit part of the language, or at least very much
influenced by the language. This should be obvious in examples such
as "Today is April the 10th, 2003." If we get to "Today is 04/10/03."
then I have to start to agree with you that this becomes less of
a language issue. But I guess to some extent it still is, because
the average en-us reader would correctly interpret 04/10/03, and
would be confused by other orderings.

The fact that you prefer ISO date formats doesn't mean that date
formatting isn't part of (written) language conventions; the easiest
way to explain it is that you don't like the general en-us conventions
and therefore use something different. I also don't like 04/10/03
because it is confusing, and I think that in today's networked world,
the chance for confusion is much higher than it was a while ago.
Also, I suspect that this list has an unusually high percentage
of people who prefer ISO dates over 04/10/03 :-). There are other
conventions in en-us that I don't like, and don't observe, such as
"quoting." (putting the period inside the quotes even when logically,
it is outside).

Anybody can of course create his/her preferred conventions for
writing in a language. While I think it is highly desirable to
be able to do that in locale-related areas, I think one potential
drawback of using language codes for some locale aspects is that
it is difficult to address such personal conventions in such
a framework.


>In addition, collation is not really a data property, it genuinely is a
>user property.  If I'm looking at Danish names, I want them collated in
>English order, not Danish order, or I'm never going to be able to find 
>anything.

Fully agreed. I don't want to claim that collation is a data property.
But I'm very sure collation is very much related to language. You
just showed that by saying 'English order' and 'Danish order'.

The limit here is that there are cases, some of them well-known
(such as German telephone-book ordering) where there are several
orders for a single language.


Regards,    Martin.