[LEAPSECS] New time scale name

p at 2038bug.com p at 2038bug.com
Fri Aug 13 05:00:57 EDT 2010



>

> UTC cannot exist without [...]



All this talk of GMT/UTC/Legislature makes me ask how the world
currently syncs their time.

In the case of SMS networks - they use NTP or nothing.
In the case of in-house business systems - they use NTP or nothing.
In the case of Internet hosts - DNS/SMTP servers - they use NTP or nothing.
Then there are many embedded devices that use the radio time signal.
Then there are a few atomic time clocks for specialized cases not
part of the NTP network.

What else?

The point is that there is *small* number of time syncronization
"networks" used by *everybody*.

When you talk about changing this or that standard, it is meaningless
unless you ask also how these systems are going to deal with it.

This is analogous to the IPv6 upgrade debate.

There is no point defining new standards unless you also have a:

1. VIABLE UPGRADE PATH
2. REASON TO UPGRADE
3. COMPATIBILITY BETWEEN SYSTEMS THROUGH AND AFTER THE UPGRADE

These other conversations would have made sense before the age of the
Internet.


From all I've heard it seems best to make UTC and UT1 identical,
and to start broadcasting UT1 over the radio and NTP networks.
And drop further leap seconds of course. I have tried, but can't
find a system that would break because of this.

2nd option is to keep leap seconds but upgrade the NTP and radio
protocols to give more warning - not 10 years warning, just *more*
warning.

3rd (worst) option is to just stop announcing leap seconds.

-paul


More information about the LEAPSECS mailing list