Re: Timezone database changes

"Trevor Talbot" <quension(at)gmail(dot)com> writes:
> Neither is the birth certificate. The recorded, legal time of the
> birth is the one that was written down. If it doesn't happen to match
> an international notion of current time, that's unfortunate, but it's
> not subject to arbitrary changes later. Even if it does match, it
> still belongs to a specific time zone. That's the key semantic point:
> regurgitating that time as anything other than exactly what it was
> entered as is simply not correct.

I'm not convinced about that. One consideration I think you are failing
to account for is that there is a big difference between past and future
times, at least in terms of what is likely to be the meaning of a
change. The above reasoning might apply to a past time but I think it's
bogus for a future time. If the TZ offset for a future time changes,
it's likely because of a DST law change, and we are in Peter's
what-time-is-the-appointment scenario. A TZ offset for a past time
probably should not change, but if it does, it suggests a retroactive
data correction. Surely you don't intend to prevent people from fixing
bad data?

regards, tom lane