A notable difference between these versions and previous ones is that the CGPM's UTC shall be interpreted by the Secretary of the Navy as well as NIST.
So the final determination of the policy issue (yea or nay on leap seconds) shall be whether the US military industrial complex believes there are more risks in continuing leap seconds (tested over 35 years) or in a boundless DUT1 (never tested, heretofore negligible for many applications). Presumably a Y2K-like inventory of DOD systems, embedded and otherwise, is called for. Since implicit timekeeping requirements would be made explicit in the deferred-leap world, all DOD systems are potentially affected.
On the other hand, one can be confident the final determination of the actual timekeeping issue (whether civil time ultimately is reconfirmed as a flavor of mean solar time) will be decided by the actual requirements placed on our civilization's many instrumentalities by the fact that we live on a rotating planet in orbit about a star that illuminates our days.
My assertion that there are many such requirements to be revealed may yet prove wrong - although confidence in this result would only build over decades. But then, the implicit - and therefore completely unsubstantiated - assertion of the "leap second die, Die, DIE!" crowd that not a single human activity - outside of astronomy and traditional sextant navigation - could possibly depend on mean solar time may itself also prove wrong. That the requirements and budgets of astronomers and TSNs are considered unworthy of any consideration also seems unfortunate.
Note that even the opponents of "leap scheduling with a granularity of one second" acknowledge that these embargoed leap seconds must be released with a granularity no larger than one hour. Nobody is recommending getting rid of leap seconds, rather merely placing them in a "lock box" for the benefit of our grandchildren's grandchildren several hundred years hence.
In any event, one can reject the pertinence of brain-dead computing standards such as POSIX to either decision. One way or another software architects will have to design library interfaces that model the actual behavior of the physical world, and applications programmers will no longer be able to rely on naive thinking about timekeeping issues.
Rob Seaman National Optical Astronomy Observatory
_______________________________________________ LEAPSECS mailing list [email protected] http://six.pairlist.net/mailman/listinfo/leapsecs
