[CF-metadata] New standard name: datetime_iso8601

From: Jonathan Gregory <j.m.gregory>
Date: Tue, 19 Mar 2013 11:06:25 +0000

Dear all

While I agree that "35246 hours since 1970-01-01" is not helpful to humans,
I agree with Steve that we should not allow an alternative representation of
time in CF, and that we should depend on software to do the encoding and
decoding into more easily understable strings (just as we depend on software
to read and write the binary files). Allowing a string representation would
not avoid the complexity of calendars, which differ in regard to what dates
are legal.

I agree with John that mistakes can easily be made, because this is a complex
problem. Are there any other ways in which we could reduce the mistakes? For
instance, would it be useful to take the simple step of providing various
examples of encoded and decoded dates in various calendars, against which
software could be checked?

Best wishes

Jonathan
Received on Tue Mar 19 2013 - 05:06:25 GMT