[tz] What data should TZDB offer?

Stephen Colebourne scolebourne at joda.org
Mon Jun 7 15:28:10 UTC 2021

On Mon, 7 Jun 2021 at 10:08, Clive D.W. Feather via tz <tz at iana.org> wrote:
> Stephen Colebourne via tz said:
> > 1) LMT
> I thought you were the one objecting violently to the idea that a zone
> might contain fake data; wasn't that the whole point of your Stockholm
> argument? So why do you want to create fake data now?

I'm recording the issues that I've seen users have with tzdb data, and
proposing a possible solution. LMT as currently defined causes issues
and I believe the proposal would be less surprising to non-expert

> > That TZDB shall adopt the principle that the main geographic files
> > (africa to southamerica) shall contain data with full history for
> > locations where zone history has differed since 1970 subject to the
> > minimum requirement that there is at least one full zone with history
> > defined for each independent country as defined by ISO-3166-1.
> I disagree with this. There is no need to create zones just to have one per
> country.

TZDB does not live in an abstract idealised world. The vast majority
of the world's population associates strongly with the country they
are in.

> I also disagree with this. If it's justified to have separate zones for
> countries, why not for dependent territories? And why should the distance
> matter?
> Oh, and why on earth "1/24th" instead of "15 degrees" like everyone is used
> to?

Meh, of course 15 degrees is a better way to put it. The distance is a
general guidance (LMT location to LMT location) to separate "local"
from "far away". eg. Aruba is far away from the Netherlands, but close
to Bonaire - that is the distinction that I tried to capture.

Dependent territories are of course entitled to their own zone
providing that data has differed since 1970. The only reason for
including the dependent territories part at all is to permit rare
cases of merging where the locations are local and aligned by
sovereignty. If there is a general view that the dependent territories
part of the proposal is overly complex, it can be dropped at the
expense of creating more zones.

> Oh, please define "dependent".

Listed at Wikipedia, based on 3 sources. The official ISO data
indicates whether the code is "independent" which is also useful.

> Why are you happy for Taiwan to be excluded under these rules but not
> Sweden? Answer: politics, which is what we are trying to avoid.
> Better would be to ignore politics entirely and say that TZDB would not
> include a zone for Kosovo until its time differs from wherever is used now.

Taiwan has an ISO-3166-1 code, just like Sweden.

The only case of interest that I can find is Kosovo, which does not
have either an ISO or TZDB code. It is *very* easy to say "TZDB will
add a zone representing time in Kosovo as soon as ISO-3166-1 includes
it". End of politics.

> For the record, I OBJECT to this proposal.

For the record, I OBJECT to the decimation of TZDB data over the last few years.

If you object then feel free to provide a counter proposal. (One that
seeks to address the issues at hand).

As a reminder, the ISO-3166-1 rule is a *minimum standard". Nothing
would change about creating or merging time zones within a country.


More information about the tz mailing list