[time-nuts] Re: UTC - A Cautionary Tale - My last word

Bill Hawkins bill at iaxs.net
Sun Jul 17 19:30:30 EDT 2005


My apologies to those who are offended by a common-sense
application of UTC as civil time. I apologize for not
taking your argument seriously but adding it to the medical
irritations in my life. I have no right to comment on the
amount of traffic in this list.

I measure caliber by correctness, completeness and consistency
of people's contributions. None of the high-caliber people on
this group have consistently contributed to this thread.
Would that I had been so wise.

I have spent 40 years in process control, not all of it in
programming. My major work was done in FORTRAN, about the
time that UTC was born. It was all control algorithms, not
time code, but I have read some interesting routines written
by high-caliber people. I'm finishing a book on process control
for the only association of control people in the US, and I've
contributed to some international standards on process control.

The adjustable rate time system was added because customers
wanted monotonically increasing time. It was a matter of
economics, not purity.

I also wrote part of the specification for the network that
connects field sensors and actuators so that they can run
distributed applications. Time synchronization is imperative.
Data link time, scheduling time and application time all run
from a synchronized oscillator in each device. Only scheduling
time is monotonically increasing. NTP is a fine way to sync
all of the oscillators using leaping UTC. Application time is
user wall clock time, so it has correction factors applied to
the display of its counter.

I believe that I have grasped the timing issues that apply to
the field of process control very well. I am unable to grasp
the esoteric point that underlies this discussion. It certainly
is not about civil time. Perhaps it is about cryptography. The
issue is clearly the abolition of leap seconds. The reason for
doing that has not been explained. Adjusting UTC to stay with
astronomical time is a feature, not a bug. I have no idea why
six months is inadequate notice for a leap second. My request
for information did not receive a reply.

But then, I didn't start reading this thread until the number
of messages grew large. The initial message asserts that leap
seconds are harmful. The argument remains unfocused because the
nature of the harm has not been specified. Unfocused arguments
die of old age after many messages.

I don't use kill files. People are entitled to their opinions,
as they are entitled to change their minds on further discussion.
The next discussion could be quite different.

Regards,
Bill Hawkins





More information about the time-nuts mailing list