Leap seconds vs POSIX means something has to give:
Time is monotonically increasing with equal length seconds.
So, you either have to violate the first bit by repeating a second (since
there's no leap second encoding), breaking the encoding (to create an encoding
for leap seconds) or you have to violate the second bit by changing the length
of the second (by introducing a frequency error).
The arrogance of the POSIX committee was such that they decided a
one-size-fits-none approach was best. While it works for many things, it makes
it impossible to work correctly (having a system that's both on-time and
monotonic and standardized encoding).
History has shown different people choose different solutions for this, none of
them pedantically correct, all flawed in some way that matters to some people,
but not to others.
Warner
_______________________________________________
LEAPSECS mailing list
[email protected]
http://six.pairlist.net/mailman/listinfo/leapsecs
