[tz] Antarctic research base Vostok local time

Robert Elz kre at munnari.OZ.AU
Fri Dec 22 00:19:38 UTC 2023

    Date:        Thu, 21 Dec 2023 12:08:27 -0800
    From:        Paul Eggert via tz <tz at iana.org>
    Message-ID:  <c9670e1d-31cf-4449-a57e-60c48b112a3c at cs.ucla.edu>

  | Unfortunately, it doesn't suffice to know just the current UTC offset. 
  | We would need to know the past UTC offsets back to at least 1970,

Not need, desire - we have had incorrect information in the past,
we undoubtably will continue to do so (as we obviously have had
recently).   Refusing to install anything because it isn't perfect
is only slightly better than refusing to take any information at all.

We should be accepting the best that's available - we can guess at
what we don't know based upon what seems likely, include comments to
indicate that some of the information is unverified and likely wrong,
and wait for someone to do better (even an anecdote "my grandfather
said he visited XXX in 1982 and the time then was 5 hours different
from where he came from...") which might provide something more
accurate than we had before, and just keep on improving as better
info appears (which is actually quite rare - most people are happy
to accept whatever they're told, and if that is reasonable, they
don't care, and nor should we.)

  | and also to know when the base was and wasn't inhabited during that period. 

We definitely don't need to know that.   That's just an internal rule
we invented that says that time doesn't matter if there's no human to
observe it.   Pre networking and computers everywhere, that was kind
of reasonable.   Now it isn't - for places no-one has ever been, it
remains OK to fail to speculate upon what the time might be there, if
only there was someone to observe (or define) it.   But once someone
has set the time at a place, that time should be assumed to continue to
run, until someone else changes it.   Even if there has been no human
at the location in a long time - they may have left running experiments
that are logging events at the current idea of the time (they can be
using GPS to get reasonable accuracy) which need a frame of reference
if they're not just using UTC.

  | And we need citations to sources for this information.

No, we don't need those either.   We prefer to have them, as something
with which to show others from where we got the data, but "personal
communication from Mr X on 2023-12-xx" is good enough really.   Better
sources are always better to have, but that are not a requirement, or
should not be.


More information about the tz mailing list