[tz] Moving more zones to 'backzone'

Robert Elz kre at munnari.OZ.AU
Mon Aug 15 22:14:45 UTC 2022

    Date:        Mon, 15 Aug 2022 13:45:20 -0700
    From:        Paul Eggert <eggert at cs.ucla.edu>
    Message-ID:  <d841baed-8aa1-708f-f653-3e409ddb93f8 at cs.ucla.edu>

  | Then perhaps you could come up with some other form for changes to the 
  | backzone data, and someone else who cares about backzone data could 
  | volunteer to put those changes into git format-patch form.

That might be possible, but I really don't understand, other than
zealotry, why it is important.   patch can handle all kinds of input.
I cannot see that it should matter which form is used (though one with
context is obviously better than one without).   diff -u is our most
common form.

If you want git format to get the change log message, then from me, that's
unlikely to help you - first because in non-compiled data (data not subject
to automated checking) I tend to make all kinds of typos, which would
need correcting, and second, because the kind of comments I would supply
are unlikely to be the kind you'd be willing to commit.

  | That being said, it's certainly fine if you don't want to bother. 
  | 'backzone' is a mess, it's unimportant to almost everybody, maintenance 
  | would be a thankless and time-consuming job,

I don't actually see many changes to data that's in backzone, and most
of what does appears from people who have spent quite a bit of time
researching the changes.

  | and everybody's better off 
  | spending their limited time doing other things.

I disagree.   The README file that was (still, and unchanged) in that
tar file you made available (at least known) in the previous message

This database of historical local time information has several goals:

 * Provide a compendium of data about the history of civil time that
   is useful even if not 100% accurate.

 * Give an idea of the variety of local time rules that have existed
   in the past and thus may be expected in the future.

 * Test the generality of the local time rule description system.

Those I think all remain important.    Eg: as now being distributed
Cambodia claims to have used UTC+0700 since 1921.   But that's not
what the old Asia/Phnom_Penh zone said, with timezone changes in the
WWII period.   Now the precise details of the changes listed might or
might not be correct, but pretending that no zone change happened at
all is much less correct (unless you have some evidence that's true,
which I doubt).

  | For example, despite recent events 
  | in Kosovo we haven't had to worry about creating a name like 
  | Europe/Pristina, and this is a win.

I don't recall seeing anyone request one, which is one reason for
not creating it - but if they did, I believe you should.  I certainly
don't see failing to create it as any kind of win.

I understand the argument - the guidelines say there is no need for
a new zone as the time has been the same as ... since at least 1970.
And you can claim that you're not making a political decision because
of that - though people in Kosovo might (one day) start complain that
you're a Russian/Serbian puppet, and are trying to pretend Kosovo doesn't

Now understand that I'm not claiming that my version of the guidelines
would make much difference in this area - but as Kosovo clearly has a
government, who are capable of choosing what timezone they're in (and as
not EU members, or not currently anyway, they're not even restricted by
EU policies), so under my proposal, they'd get a zone if they asked for one.

Of course, to their opponents, that would mean that you're adopting the
Kosovo side of their dispute, you're just a US puppet, and proclaiming
that everything is biased against them.

That is, there is no avoiding of political controversy here, act, don't
act, you're always going to anger the people who wanted the other decision.

Or in other words, "avoiding political issues" simply doesn't work.

What you're really saving is 15 minutes editing and testing now, and
deferring it to sometime later, when they decide to change their zone
rules, and a new zone is necessary, but at this point it becomes a
harder transition for everyone (everyone else anyway) as people who would
logically have been using Europe/Pristina (and could have switched to that
from whatever they have been using, whenever they felt inclined, since
both would have generated the same results - at least post 1970) would
then all have to switch in a hurry, to get the modified zone in time for
the change to take effect.   Not nice to inflict on people for no good reason.

This is exactly the same as when you said that we won't tell people in
Sweden to switch to using Europe/Berlin (though that's effectively what
you are doing to them) so things are easier if they Sweden and Germany
separate their zones (like one keeping summer time, and the other not).
That's fine for the Swedes (though dropping their historical data is not),
but why are the people in Kosovo not afforded the same courtesy?


More information about the tz mailing list