[tz] models for timezones

Steve Summit scs at eskimo.com
Mon Feb 12 19:44:55 UTC 2018

Guy Harris wrote:
> On Feb 11, 2018, at 7:45 PM, Steve Summit <scs at eskimo.com> wrote:
> > In that old model, a time zone has one offset from Greenwich/UTC...
> Which is why I prefer calling the things identified by tzdb identifiers
> "tzdb regions" rather than "time zones" - a tzdb region may, over its
> history, be in more than one of [...] "time zones", and there may be
> more than one tzdb region in a given "time zone".

Well said.

And even though I was just preaching about "things not always
sufficiently appreciated", I'm not sure I've sufficiently
appreciated this point.  I live in America/New_York, and it's
easy to imagine this is just sort of a longhand for EST5EDT,
but of course it's not.  (And if I lived in, say, northwestern
Indiana, I would be less inclined to make this mistake.
Perhaps living in northwestern Indiana for a few years should be
a prerequisite for anyone aspiring to profess any real expertise
on time zone issues.)

More information about the tz mailing list