[tz] proposed changes: past Altai/Tomsk time zone shifts

[Robert Elz]

    Date:        Sat, 5 May 2012 04:11:55 +0200
    From:        Tobias Conradi <tobias.conradi at gmail.com>
    Message-ID:  <CAAGevbVsY5SEs=3NGyau7T_aLW+W0cV18tofP3d_aDfjTw1M0w at mail.gmail.com>

  | Putting something into the wrong timezone is by definition having a
  | bug in the TIMEZONE database.

No, once again, it is not, or not in any way that is important.   We
make the various timezones, we do not assign them to locations, other
than for our convenience in making sure that we have all of the zones
defined that we should have.

Certainly, if we have those assignments incorrect, they can get fixed - just
as we would fix an erroneous comment in the source code of the C files
if there was such an erroneous comment, but neither of those is a bug
in any real sense.

  | If that would be true, the project wouldn't exist in that form.
  | Because anyone could write down all offsets with full seconds from 0
  | second to e.g. 30 * 60 * 60, assign names to them and would be
  | finished.

No, that's approximately what existed (in some implementations anyway) when
we started - the timezone isn't just the offset from UTC, it is a list of
changes of that offset (when summer time started and ended in various years,
when the zone offset altered, ...)   That complete set of data is a timezone
for our purposes, and getting all of those defined (correctly) is what we're
attempting to do.   Specific locations are only relevant first as picking one
to provide a name for the timezone (and the theory file explains how that's
done, it isn't relevant to the current discussion) and second, as a way of
verifying that we aren't missing a zone definition.

  | > If the comments in the zone file are wrong,
  | > we should certainly fix them sometime,
  | Now you change your opinion?

No, not at all, we fix erroneous comments, but erroneous comments are
not a bug, as they don't affect the results, they're just a minor
insignificant error.   Like any error, such things should get fixed,
but like other irrelevant errors, fixing them isn't urgent.

  | Any prove that all the view at
  | http://stats.grok.se/en/201204/List_of_tz_database_time_zones
  | are from list members?

List members can do whatever they like, and some of the people on
this list (and associated with this project) certainly do attempt to
make maps of timezones in one way or another.   That's all fine, but
it is not this project.

  | >(no-one else is really very likely to
  | > read through all of that stuff - zic certainly doesn't care.)
  | I read the comments before I was a list member.

Sure, there are people who look at this stuff, but not very many
compared with the number of people who use the real product of this
project - which is the zone database.

kre