[tz] Formal models and the need for storing original source data

Robert Elz kre at munnari.OZ.AU
Sun Jun 21 02:53:36 UTC 2020

    Date:        Sat, 20 Jun 2020 17:05:34 -0700
    From:        Michael Kenniston <tz at michaelkenniston.com>
    Message-ID:  <CAF570XH7PjWAbhEBwN2=kU6TP82WHVUnoi2AAGA3eYr++J9N8g at mail.gmail.com>

  | You can also have data recorded in the past which refers to the future,
  | so even a change that affects only future timestamps can still
  | invalidate existing stored data.

Not if the authoritative original data is what is stored.  What that
means might alter, but the data is still as valid as it ever was.   Of
course, should the data change (which it can, since it has yet to
occur - eg: a meeting originally scheduelled for a future Monday is
moved to the following Wednesday) then the stored data will be invalid
- but I don't see how anything can fix that in advance.

  | When exactly do we *know* anything, I mean really for sure?

I know that the time now us 09:49:43 as I typed that '3' (not
any longer) and that that was 02:49:43 UTC (when I typed it).
Things might change in the future, but that equivalence never
can since it has passed now, and we cannot alter the past.

  | Never say never.

Don't be naive.   (sorry, no way to type an i with two dots...)


More information about the tz mailing list