On 3-mei-04, at 7:23, Keith Moore wrote:
I'm not arguing that the RFC822 date format is sensible and good.
nor, really, am I. but if we're going to change it, we should change
it to something that will actually interoperate better - not just
something which will produce a different set of interoperability
problems.
Can't disagree with that. And I think the way to accomplish this is to
have a rule that given a date/time/timezone, there must be only a
single valid representation of that time in the format we end up
selecting.
(note that if we change the date format we'll get interop problems
from the conversion between old and new formats no matter what - so
the change isn't justified unless we drastically reduce the interop
problems associated with the new format as compared with the old.)
Interoperation with legacy email solutions is off-topic on this list.
:-)