From: John Stanley (stanley@oce.orst.edu)
Date: Thu Sep 11 1997 - 13:57:29 CDT
Pete Resnick wrote:
> What gave use the Year 2000 problem was that people thought that they
> didn't need to plan for "forever." In the DRUMS rewrite of RFC 822, the
> Date: fields do last forever because the year can have an infinite number
> of digits. I will repeat:
>
> It's simple to form UIDs that last forever.
IF you assume that nothing in the furture will ever be different than
what has happened in the last two days, yes, it is very simple to assume
you can form message id's that last forever.
Once you realize that you can't predict what will happen in the future,
you realize that trying to claim that something will be "forever" is
silly.
One example. You keep talking about the year being included in the
message id. You do realize, I hope, that the year that you and I think
it is now is not the same as the Jewish year or the Chinese year, and
that it is quite possible that the enumeration of the year will change
at some time in the future.
Another? I gave one already: the next revision keeps compatibility with
this one, but the one after that does not. "Forever" is a lofty goal,
but givin that you can't predict the future, a little silly to claim.
> Do NOT put any limit on the length of time UIDs must last. It's a silly
> thing to do.
Do not pretend you can predict "forever", it is a silly thing to do.