On Tue, 19 Sep 2000, Rick Thomas - Richard wrote:
> The problem is that there does not exist a relevant POSIX committee
> any more to accept that proposal and change the standard.
Proposals should go to pasc-time-study_at_opengroup.org. (However, there is
no consensus in that group on the correct form of the changes needed,
because of the contradictory requirements that (a) the conversion of an OS
timestamp to a (UTC or local time) display timestamp should not change
over time or with updates of leap second tables (such changes have caused
problems with FTP site mirroring in the past) and (b) it should be
possible to use ordinary C arithmetic operators to manipulate timestamps
and compute time intervals between them. If one considers (b) more
important, TAI second or nanosecond counts are preferred; if (a),
structure types or fixed point types that can represent all valid leap
seconds are preferred.)
--
Joseph S. Myers
jsm28_at_cam.ac.uk
Received on Tue Sep 19 2000 - 15:01:44 PDT