[tz] Suggestion for addition of America/Brasilia entry to time zone database

Random832 random832 at fastmail.com
Wed Oct 19 13:38:04 UTC 2016

On Sun, Oct 16, 2016, at 23:55, Gabriel Gaspar via tz wrote:
> Hello there, and sorry for the intrusion.
> I presently write to this list with a suggestion for addition of an
> entry in the time zone database, America/Brasilia, or at least for a
> clarification on why there may not be one. Being the standard official
> time in Brazil, I firmly believe there should be one such entry
> (unless, of course, matters of greater power disallow it).

This is essentially the same as the Beijing/Shanghai problem that has
been discussed before - the policy is that the timezone name in the
timezone database is based on the largest city that the timezone applies
to, regardless of the existence of another city which may be part of a
popular or official name of the same time zone. Thus China's timezone is
"Asia/Shanghai" despite being called "Beijing Time".

This is considered acceptable because the raw timezone name is not
supposed to be displayed to users; the string meant to be displayed to
users is the description in zone.tab, which is "Brazil (southeast: GO,
DF, MG, ES, RJ, SP, PR, SC, RS)" for America/Sao_Paulo. For its part,
Asia/Shanghai's description is "east China - Beijing, Guangdong,
Shanghai, etc." Perhaps Asia/Sao_Paulo should likewise be "southeast
Brazil - Brasília, Rio de Janeiro, São Paulo, etc." and leave the
precise enumeration of states to the shapefile project.

It's also worth noting that timezones in the timezone database don't
generally correspond one-to-one with civil timezones, but subdivide them
based on regions within a timezone that used to be in a different civil
timezone before the present day but after 1970, or have or had different
rules for daylight saving time. For example, Mato Grosso and Mato Grosso
do Sol have different timezones (America/Cuiaba and
America/Campo_Grande, respectively), because the former did not observe
daylight saving time in 2003-2004. In all, there are sixteen timezones
applying to various parts of Brazil, and it's simply good fortune that
there is (so far as I can tell) only one that applies to the whole
section of UTC-3:00 that presently observes daylight saving time.

Out of curiosity, are the northern parts of UTC-3:00 that do not observe
daylight saving time (Bahia, Fortaleza, etc) also considered "Brasilia
Time"? Or for only part of the year, and the rest it is "Brasilia time
-1", when they're not the ones whose clocks have changed [and then
presumably Fernando de Noronha, whose clocks also do not change, is
"Brasilia time" for part of the year]? How should this be differentiated
if this name is used?

More information about the tz mailing list