Re: [LEAPSECS] An immodest proposal

From: Rob Seaman <seaman_at_noao.edu>
Date: Tue, 14 Feb 2006 20:39:16 -0700

On Feb 14, 2006, at 2:28 PM, M. Warner Losh wrote:

> UTC time stamps in NTP are ambiguous. TAI ones are not.

Requirements should be kept separate from implementation. Whatever
the underlying timescale, certain external global requirements
apply. Whether NTP or some other implementation properly captures
those requirements is a separate issue and should be treated as such.

My understanding of the point you are making is that given an
infrastructure like NTP that (more-or-less) assumes an even interval
time scale, that UTC timestamps are indeterminate (to some level of
precision) in the absence of a table of DTAI. Of course, given an
accurate UTC clock, TAI is indeterminate without that same table of
DTAI values.

> UTC time stamps do not convey enough information to properly
> implement things like intervals, while TAI ones do.

And TAI does not convey Earth orientation. These are all well
traveled issues.

> The NTPNG stuff that I've seen appears to consider these problems
> as worthy of needing a solution and they plan on solving them.

Well, good!

> It isn't rocket science, but one has to divorce ones self from the
> chauvinistic view that UTC is always best.

Chauvinism: Prejudiced belief in the superiority of one's own
gender, group, or kind: “the chauvinism... of making extraterrestrial
life in our own image”. Doesn't really seem that applicable...

As has been repeatedly emphasized, astronomers are serious users of
*both* UTC and TAI. UTC is not always "best", rather different
purposes require different time scales. Many of us happen to think
civil time - on planet Earth - should remain some variant of UTC/
GMT. That current implementations of the UTC standard have
shortcomings is no surprise.

> For time exchange, it is not the best, and has many problems around
> the edges.

Are there any time scales that don't require "time exchange"? The
key issue remains interval time versus Earth orientation. It would
be incorrect to assert that all timekeeping issues in computing
devolve to a need for interval time.

> Doing NTP with TAI (and the implied requirement for DTAI) doesn't
> change what time is displayed for users. It does make it *MUCH*
> easier to get leap seconds right for those users that need them.

Perhaps. Would be prudent to find out. This is, however, only one
aspect of timekeeping.

Leap seconds aren't needed by users - they are required by UTC that
is needed by users for various purposes.

> Anything else is madness.

Now who is chauvinistic?

Surely we could design a series of experimental programs to address
these issues in an unbiased fashion. Is this a controversial statement?

Rob
Received on Tue Feb 14 2006 - 19:39:31 PST