[LEAPSECS] Cost: getting rid of GMT & discontinuing leap seconds

Warner Losh imp at bsdimp.com
Sun Oct 24 21:30:27 EDT 2010

> I asked:


> >> What coherent system design process led the engineers to select UTC in spite of the existence of leap seconds?


> Warner Losh asserted:


> > UTC is the only game in town.


> The foundation of civil time is universal time, that is - mean solar

> time. UTC is a widely available realization of universal time.

> Other timescales are all around.

Can you point to one that's broadcast, in real time, around the planet
like UTC is? GPS is broadcast, but nobody does time stamping in GPS,
and it lacks the force of law that UTC has.

> > It is the legal time everywhere, for all practical purposes[*].

> Adding an asterisk does not mitigate all the prior threads

> challenging this assertion.

Yes, actually, it does. The threads have all so far been very
theoretical, with no actual recent court cases to prove to the
contrary. In cases like this, one has to go with common usage.

Put another way, if someone were to say that I was out of compliance
because I delivered something on-time in UTC but not in time in UT1,
they would have a very difficult case to prove.

> > It is the obvious choice because it is the only choice.


> Surely system engineering best practices involve paying as much

> attention to putatively "obvious choices" as to those with known

> options. (Which is to say that system engineering is about "paying

> attention".) That attention might amount to something like "yes,

> the instructions should be in English", but also might amount to

> "hey, there's an internationalization opportunity here".


> That said, UTC is not the only choice now, nor will it be in the

> future.

And what are these other time scales available today?

> > GPS time is weird, as is TAI time.


> GPS time is widely available. NTP, for instance, provides proof of

> concept that GPS is usable (weird or not).

The NTP standard mandates UTC time. NTP recovers UTC time from GPS
time and broadcasts that.

> I personally find it weird that the keepers of the atomic clocks

> apparently want to (also) kill off TAI (International Atomic Time).

Interesting, but not relevant.

> UTC with leap seconds is apparently thought "weird" by many here.

> UTC without leap seconds (thus, no longer a flavor of universal

> time) would certainly be weird.

Also not relevant. UTC is the only game in town, for all practical
purposes. It is the only one broadcast world wide in real time, and
it is the only one that will stand up in court.

> Sexagesimal representations are weird (and imply angles - intervals

> should use decimals).

Also not relevant. Leap seconds make UTC a non-uniform-radix time
notation, which is pretty weird.


More information about the LEAPSECS mailing list