[time-nuts] GMT vs. UTC

[Magnus Danielson]

From: "Poul-Henning Kamp" <phk at phk.freebsd.dk>
Subject: Re: [time-nuts] GMT vs. UTC
Date: Tue, 21 Nov 2006 17:14:11 +0000
Message-ID: <2494.1164129251 at critter.freebsd.dk>

> In message <20061121.092927.58461219.imp at bsdimp.com>, "M. Warner Losh" writes:
> 
> >GMT means many different things over time.  Once upon a time it did
> >mean [...]
> 
> Just to be clear:  GMT has always meant "Greenwich Mean Time".

Indeed.

> What time that was on the other hand, has changed a lot.

Certainly, what time the timescale showed has changed over time.

> >Now it is much
> >closer to UTC in definition, but it can be hard to find an exact
> >definition because it keeps changing.  The differences are < 1s (thank
> >you leap seconds, or curse you depending on your stripe).
> 
> The important point here is that GMT is a political timezone, not a
> scientific one.  You get the wrong minister in whitehall and GMT
> does something silly.

One has to understand the reason for my question, which stems back into the
legalities of things. Can we state a time event in GMT and relate that back to
our "non-political timescales" of TAI and UTC (UTC has politics around it for
sure, but those are on the leapsecond aspect) which we methods for transporting
and realizing, and then say with certainty that this time is _known_ in a legal
sense. Especially, can we do this in some other country than GB?

I would feel more comfortable if I knew that the official time (and those the
legal time) would be some 1h multiple of UTC and that this 1h multiple is
precisely defined for normal and summer time. However, if I start to ask
questions like "can I have calibrated and traceable time?" people look
confused. The concept seems to be missing. Your clock may be unstable and
wrongly set, but if I have the right calibration and traceability track for it,
I might get the corrections needed to know within some suitable certainty that
a certain event having legal meaning did occur at some time in the legal
timescale. This is just the same as with other representations of measures,
such as elapsed time, frequency, length, weigth etc. Whenever it does matter,
how is it traceable back to what have been defined as the legal representation.

Time is an important property which we use for all kinds of things in our
modern society, but getting it clearly defined to have the meaning we now
assign to it have been lacking.

For instance, in GB there is a very elaborate system to monitor streets with
cameras, especially in London. It has internal synchronisation of time.
However, Scotland Yard was unaware that NPL was holding the national time and
their system was thus not connected to the NPL time. If you want traceability
you must be dreaming. If you want time to actually matter in a legal sense,
then there is a list of things to do before you can expect time to be traceable
to a common source within each country. Now, with internationalization, how do
you make time measures traceable to the same source so you can coordinate time
events? If you can't even get the national regulations right, you can give up
all hope right there. Strange that it would have to take a couple of swedes to
tell Scotland Yard where to fetch the time, it was not our intention to
interfere with the national affairs of GB. :P

Infact, there isn't an agreement on when summertime occurs, it differs from
country to country by up to 0.9s. Maybe not the biggest of issues, but it just
shows the difficulties involved.

Cheers,
Magnus