Re: how do computer people want their time clocked?

From: Paul Eggert <eggert_at_twinsun.com>
Date: Mon, 28 May 2001 22:51:46 -0700 (PDT)

> Date: Mon, 28 May 2001 16:23:54 +0200
> From: "Deckers, Michael" <Michael.Deckers_at_FUJITSU-SIEMENS.COM>
>
> leap seconds in UTC would not trouble them at all
> as long as they were all "negative", that is, if UTC were consistently
> slower in rate than UT1.

Alas, UT1 is slowing down, not speeding up, so to make them happy
indefinitely we'd have to permanently lengthen the UTC second every
now and then. And if we were willing to do that, we might as well
revert to the UTC regime before 1972, which lengthened the UTC second
periodically to keep it in sync with UT1 without discontinuities.

> Is monotonicity really good enough for those who
> cannot afford the access to a smooth timescale such as TAI or
> GPS time, or do those people need a timescale that is continuous
> as well (one without jumps), such as the proposed UTS?

For many applications, monotonicity plus continuity is better than
monotonicity alone. For example, a one-second jump forward can cause
problems with applications that need to sleep for a few milliseconds
only.

I suppose monotonicity alone is better than nothing; but if we're
going to introduce a new monotonic time scale like UTS (or revert to
pre-1972 UTC) we might as well have continuity too.
Received on Mon May 28 2001 - 22:51:56 PDT