Re: [Bug localedata/16777] Incorrect thousands separator in pl_PL locale

[Marko Myllynen <myllynen at redhat dot com>]

Hi,

On 2014-03-31 13:44, Keld Simonsen wrote:
>>
>> Perhaps you want to see whether that's doable, please see the Locales wiki page
>> which has a link to a page describing LC_NUMERIC and grouping in detail. The
>> page contains also information on how to test and submit your locale changes.
> 
> I do not think the "under 10000" is doable with current glibc functionality.
> 
> furthermore I think this is not desirable, for LC_MONETARY.
> 
> The no break space versus the COMMA/PERIOD thousands separator issue
> is a classical problem between linguistic and computer use.
> Linguists tends in many languages, European style, to advocate no break space,
> but for computer use a period is most often used. I have yet to see a financial
> application to use no break space. Bank applications and ledger applications do not
> use no break space.
> 
> Actually we shoud probably provide for both styles, and also do the "no separator
> for under 10000.".

I filed a report about the "no separator for under 10000" case:

https://sourceware.org/bugzilla/show_bug.cgi?id=16782

> I would then advise that we use the current functionality
> in the POSIX style, and compatible with the POSIX/C locale, and then have new keywords
> both in LC_MONETARY and LC_NUMERIC for the linguistic styles.

So LC_MONETARY/mon_thousands_sep and LC_NUMERIC/thousands_sep would
still be used in the POSIX context and the new keywords would be defined
for linguistic context? What would be then used for the currently
defined keywords with locales where no-break space is the recommended
separator and which would used for linguistic context? Are there locales
where such distinction is defined?

Thanks,

-- 
Marko Myllynen