[CF-metadata] mixed Julian-Gregorian calendar

Jonathan Gregory j.m.gregory at reading.ac.uk
Tue Dec 18 09:11:59 MST 2012

Dear Steve

I agree with all you say about the messiness of it, certainly. I also agree
that we should change the default for datasets created with the next version
of CF (as I have voted in Cecelia's poll). I think the choice of the default
is the main issue we are trying to agree on, isn't it.

>From your last posting I understand that you think we should remove support
entirely from CF for the mixed Julian-Gregorian calendar. I don't agree with
that still. Despite its messiness, especially that it changed at different
times in different countries, it is possible to define the "real world"
calendar, by making a particular choice (as udunits and cal(1) do), and anyone
dealing with historical data archives has to do that as well. Hence it is 
usable, if inconvenient. Even if no-one replies on this email list to say that
they are using it, I don't think we should remove it from CF. CF has been a
standard for more than a decade and the readers of this email list do not speak
for all existing data and applications which make use of CF. Since udunits and
cal and no doubt many other softwares support the real-world calendar, despite
its messiness, I think it would be excessive for CF not to do so. By changing
the default, however, we can greatly reduce the risk of errors being made, and
it is important to do that.

Best wishes


More information about the CF-metadata mailing list