[LEAPSECS] Crunching Bulletin B numbers (POSIX time)

Rob Seaman seaman at noao.edu
Mon Feb 21 13:14:46 EST 2011


Ask Bjørn Hansen wrote:


> Making timekeeping more complex for the general use case[...] is DOOMED.


Rather, what is complex (and will remain so) are the timekeeping requirements. We cannot make the problem of timekeeping "more complex" (or less complex, for that matter). The question is how to meet the requirements in a sufficiently simple fashion.

We would likely all share a level of skepticism about the tender of elaborately over-engineered proposed-solutions. We should be similarly skeptical about under-engineered proposed-solutions (like the ITU draft does such a poor job of describing). The hidden complexity that is being swept under the rug will inevitably pop up somewhere else.

Actual solutions don't seek to ignore the system requirements. Politics is bad engineering.

Rob


More information about the LEAPSECS mailing list