Arnt Gulbrandsen wrote:
It seems safe to assume, then, that the range of time zones will remain
greater than 24 hours in the near future. 199 hours (the range supported
by 2822) is much bigger than the current requirement of 27 hours, but what
should the limit be? 28? 30? 48? 49?
Slightly greater than 24 hours; see the comments in the file
"australasia" from the tar archive mentioned earlier,
particularly at the end of that file -- 26 hours currently,
27 if one includes zones in effect back to 1973.
I believe it suffices to require:
1. that the offset specified correspond to some established
time zone in effect at the specified date-time, leaving
the range subject as it is to the whims of politicians
2. that the "minutes" part be in the range 00 through 59
(2822 seems to permit +9899, for example).
At least such requirements would preclude nonsense that
can't be correlated to a real time zone. Frankly, I don't
think the offset conveys much useful information, anyway