Re: much smaller libgweather timezones update

[Dan Winship <danw gnome org>]

Roozbeh Pournader wrote:
> Dan (and everybody interested in translating these),
> 
> Just to bring to your attention that the Unicode CLDR standard
> provides localization patterns and best practices for user-friendly
> timezone names in user interfaces:
> 
>   http://www.unicode.org/reports/tr35/#Timezone_Names

Yeah, we were going to use their data, but we decided against it. In
particular, one of their design goals is to have all regions that share
the same offset and DST rules end up having the same timezone name (eg,
"Central European Time", "East Africa Time"), because the CLDR timezone
data is mostly concerned with doing strftime(), and so translating
"Europe/Stockholm" and "Africa/Tunis" both to "Central European Time"
makes it easy to compare timestamps between the two.

But for us, saying "Central European Time" is bad, because it turns out
that a lot of people in that timezone have never even heard of that name
for it. So for libgweather's purposes, it's better to just say that
"Europe/Stockholm" is "Sweden Time" and "Africa/Tunis" is "Tunisia Time".

More notes in the bug, particularly here:

  http://bugzilla.gnome.org/show_bug.cgi?id=529054#c6

FWIW, the arguments there don't apply as much to showing dates in the
evolution calendar, so we may want to use the CLDR data there. (There's
still the problem that AFAIK most distros aren't packaging the CLDR
data, and that most of the CLDR timezone data isn't translated into many
languages yet...)

> Disclaimer: Gnome Foundation is a member of the Unicode Consortium,
> and Behdad and I are our representatives there ;-) This also means
> that you find a fault with the (localized/original) data or the
> specification that you wish to be corrected, pass me and Behdad a
> note, we will contact the people in charge to correct them.

OK. Well, I'm not sure this is really a "fault". It's more of a use case
that is a little bit outside what they're trying to cover. In order for
us to be able to use their data, we'd need a flag in the metazone data
distinguishing timezone names that really honestly are in common use
(eg, basically 100% of Americans know about "Eastern Standard Time") vs
the timezone names that were invented for the tzdata or CLDR databases
which people just *wish* were in common use.

-- Dan