tzdata 2003e: imperfections

Paul Eggert eggert at CS.UCLA.EDU
Mon May 17 19:33:58 UTC 2004

Oscar van Vlijmen <ovv at> writes:

> Cyprus is now (May 01, 2004) part of the European Union. In spite of the
> interesting remarks for Cyprus in the tz data file, the Cyprus entry should
> be in the europe file.

This shouldn't matter one way or the other to tz users, since
Asia/Nicosia and Europe/Nicosia are aliases.  Personally I'm a
traditionalist so I'd rather leave it be.

> antarctica file:
> New Zealand: Rules are given, but no Zone information, so no actual time
> information.

The NZAQ rules are defined and used in the antarctica file, so I don't
see a problem here.  (There's a separate copy of the NZ rules here so
that the antarctica file can stand alone.)

> GMT is obsolete. UTC should be used.

Which specific uses of "GMT" are you objecting to here?

Some instances of "GMT" are correct, e.g. in winter Europe/London
users prefer the "date" command to output the string "GMT", not "UTC".
By and large the tz database attempts to support common
English-language usage, even if it's not strictly "correct".

There's another problem I should mention at some point: Calcutta
officially changed its name to Kolkata on 2001-01-01, so the name
"Asia/Calcutta" will become obsolescent at some point.  Altavista
reports 589,792 hits for Calcutta and only 194,498 for Kolkata, so the
name-change hasn't really "taken" yet.  We have similar naming
problems in Greenland and Vietnam so I suppose we can live with one
more such problem for a while.  There are virtues to having stable
names, even if they're "wrong" by some official measure.

More information about the tz mailing list