Re: [PATCH] Fix first_weekday/first_workday for EU locales

[Petr Baudis <pasky at ucw dot cz>]

On Wed, Jun 06, 2012 at 02:41:47PM -0600, Jeff Law wrote:
> On 06/05/2012 12:05 PM, Petr Baudis wrote:
> >On Tue, Jun 05, 2012 at 11:08:15AM -0600, Jeff Law wrote:
> >>I'm digging out, but plan to do the en-mass fixing of addresses;
> >>then I'm happy to participate in setting a reasonable set of rules
> >>so that locale issues can move forward in the future.
> >
> >   I have tried to document the current consensus at
> >
> >	http://sourceware.org/glibc/wiki/Contribution%20checklist#Qualification_.28Locales.29
> >
> >some time ago. Any comments / suggestions are of course welcome!
> What about referencing CLDR; Uli claimed it wasn't accurate, but
> never provided any details.

Indeed.  I still don't feel comfortable with blindly updating data from
CLDR (since I'm not very familiar with the repository and its policies
yet), but given its wide userbase I see nothing wrong with taking CLDR
reference as a *very* strong supporting argument for a change.

In very long term IMHO it would be sensible to slowly reduce differences
to CLDR (by appropriate changes on both sides) and finally switch to
CLDR providing locale data. Avoiding two redundant databases seems even
more appropriate given the recent legal opinion that locale data are not
copyrighted and therefore there are no licencing and philosophical
(freedom) considerations here.

However, if we want to go that way, this should probably happen
very gradually so that we can get familiar and comfortable with CLDR
processes and policies - right now, I don't really know much about it
personally. For starters, a tool that quickly gives a diff between
glibc and CLDR locale data would be awesome (and likely straightforward
by reusing existing tools). I will put that on my TODO list, but it
will be probably fairly easy to beat me to it.

-- 
				Petr "Pasky" Baudis
	Smart data structures and dumb code works a lot better
	than the other way around.  -- Eric S. Raymond