Re: [LEAPSECS] Leap second status?

From: Rob Seaman <seaman_at_NOAO.EDU>
Date: Mon, 2 Dec 2002 12:52:10 -0700

John Cowan says:

> The needs of the astronomy/navigation community are certainly important,
> but the civil day no more needs to be tied to the solar day than the
> civil month needs to be tied to the synodic month for the benefit of the
> Jewish and Muslim communities. Keep your own calendar if you need one.

This is a false analogy. The problem is not that civil time might take
large excursions from local apparent time. The problem is that a small
monotonic secular effect will eventually win out over even large periodic
swings. (Secular in the sense of "accumulating through a long duration
of time". Do folks other than astronomers use this term?)

For many purposes, we could indeed specify civil time as sloppily as we
like. The fact that society manages to handle various daylight saving
schemes is evidence of that. But, besides the question of whether the
precise timing community should be satisfied with such sloppy work, we
would be setting a trap for the future.

There is no possible way that society would overlook some goofy scheme
put forward by us pointy headed poindexters that results in the sun
rising at midnight (yes, yes - in mid-latitudes). That is one
inevitable result of terminating the issuance of leap seconds.

And for many purposes, we wouldn't have to wait very long. Even
amateur astronomers would be able to detect the absence of leap seconds
after just a few years. Could they work around this? Sure - if a
good enough reason were supplied for the change - and if a large
and expensive infrastructure were provided to disseminate the new
multisecond (and later, multiminute) DUTs - and if they were motivated
to install whatever updates and fixes were required to keep their
equipment and software working with a DUT > 0.9s.

Is there a good reason? Have we identified who is going to pay for the
new infrastructure? Do we know what changes will be needed to consumer
electronics and software?

There are a lot of technical issues that have yet to be discussed in
detail by the various professional communities affected by issues of UTC.
But more fundamentally, what do we tell the public about the inevitable
requirement that would be created to reset all the world's clocks at some
future date? That date might be delayed until far in the future, but if
so, the discontinuity in the clocks would be all the greater - as well as
the likelihood for Y2K-like global failures. (Just the potential legal
exposure is awesome to contemplate. Have folks consulted lawyers, and
have those lawyers consulted historians about the legal ramifications
encountered during the transition in the 19th C. from local to standard
time?)

And so, might I humbly suggest that whatever changes may be made to the
UTC standard - now or in the future - that the new standard specify a
solution that is complete indefinitely far into the future? That is the
real problem here, that the current standard was an interim solution.
Apparently, even the extended period of time over which the current
standard is valid - certainly at least the next 1,000 years when we
allow for issuing monthly leap seconds (and maybe several times 1,000
years) - even that extended period is not deemed long enough by some
in the current community.

As far as other replies to my initial message - I fear that some have
fallen into the same trap of focusing on details that besets the
pro-change faction. For instance, I have no personal opinion to make
known about the proposed UTS scheme. Some of the other messages must
require additional context to interpret - context that I fear I don't
possess :-)

In any event, before we analyse *any* solutions, we really ought to
write down a coherent set of use cases and develop detailed requirements
for a wide range of user communities - including just plain citizens.

Has the process to date produced any use cases for UTC or civil time?
Have requirements for any modification to the current standard been
published for discussion? For instance - what is the estimated global
cost for various proposed "solutions"? Have the discussions to date
consisted of much of anything except rather naive political maneuvering?

All I've heard throughout this entire discussion is unfocused fretting
about an issue that won't become significant for decades or centuries.
Why then are we hurrying to force some contentious mandate that meets
some artificial deadline?

Whatever the underlying time systems that are disseminated around the
world, professionals need to implement software, hardware and operating
procedures that use those systems correctly. The current paired UTC and
TAI standards form a solid infrastructure to build upon. Is there a
single example of a precise timing project with substantive "issues" -
issues, that is, that the project team themselves weren't ultimately
responsible for creating? And have we given appropriate weight during
the short circuited deliberations to the many projects, professional
and otherwise, that have *not* suffered significant issues?

Look before you don't leap.

Rob Seaman
National Optical Astronomy Observatory
Received on Mon Dec 02 2002 - 12:00:16 PST