UTC Questionnaire response
Joseph S. Myers
jsm28 at cam.ac.uk
Sat Dec 11 14:06:46 UTC 1999
This is a personal response to the UTC Questionnaire, as an individual
contributor to the public domain UNIX timezone database at
I had originally not intended to respond to this questionnaire, since my
experience is in the study of historical timezone changes and summer-time
rules, and my views on the technical issues considered on their own are
generally in accordance with those expressed in the response by Markus
Kuhn, but on further consideration I decided that historical timezone
research shows up potential practical, civil and legal issues that could
arise with the decoupling of UTC from UT1, and of which it might be
desirable to present a summary.
In general, the time used for everyday purposes in most parts of the
developed world (where time signals are readily available) is based on
UTC, as distributed by time signals, with an offset that is normally an
integral number of hours depending on the local timezone and summer-time
arrangements. However, the legal definition of the time to be used may
vary from country to country; for example, in the United Kingdom the time
has been based on Greenwich Mean Time since 1880 (except that the whole of
Ireland used Dublin Mean Time until 1916, then a single period of summer
time at DMT plus one hour, then time based on Greenwich); while it has
been proposed to change to a UTC-based definition, this sort of matter has
not been a legislative priority. Although GMT has been replaced as a
designation of a timescale by more precise definitions of variants of UT,
it seems most likely that GMT would be taken, if UTC were decoupled from
UT1, to be UT1 or very close to it, rather than UTC. In France, the time
is now based on UTC, but from 1911 until 1978 the definition was based on
Paris Mean Time, retarded by 9 minutes 21 seconds. The European Union
directives defining the start and end points of summer time in the EU
still state the relevant points in terms of Greenwich Mean Time. The time
used in Liberia did not change to a Greenwich-based zone until 1972.
>From this, it seems likely that many countries have definitions of legal
time based on GMT rather than UTC; and, while the difference is
insignificant for the purposes of everyday life, changing it has been of
no great priority; a second here or there is not generally of great
significance in the timing of everyday events. However, if UTC and UT1
are decoupled, it would seem likely that the legal and practical times
would for the most part be allowed to drift, at least until this caused
notable problems and quite likely for some time thereafter. Examples of
the problems arising, taken from the 1847-1880 period when legal time in
Great Britain was local time but for most purposes Greenwich time, as used
by the railways, was used, may be found in Howse's book (Greenwich Time
and the Longitude). In 1858 the case of Curtis v. March (where a court
wrongly disposed of a case because the defendant had not appeared on time
by the court clock, on Greenwich Time, but appeared on time by local time)
lead to the case law that legal time was local time (though almost all
clocks showed Greenwich time). Despite the confusion so caused, and the
lack of ready availability of local time, the matter was not legislated
upon until 1880. It seems plausible that similar disputes, about whether
or not something was done within specified times, could arise if UTC and
UT1 were substantially decoupled.
While it would be clearly desirable for legislation in all countries to
establish the legal time as being based on the readily available time
scale, UTC, rather than GMT, I do not expect this state to be approached
for several decades if it becomes practically important, or much longer if
it does not.
> A. If the appropriate international bodies decide to eliminate the
> insertion of new leap seconds, would you foresee any practical
> problems for your institution/instrument/observations?
I would expect the problems described above to arise somewhere for some
people, to cause confusion in areas where exact timing is legally
important, and so to provide a source of undeserved extra income to
lawyers. I would not expect to be involved in any such problems myself.
It would be appropriate to study what legal definitions of time scales are
used around the world, and whether eliminating leap seconds would cause
any change in such definitions.
> B. Would you be in favor of such a proposal?
My preference would be for Markus Kuhn's proposal of leap seconds
announced several decades in advance.
Joseph S. Myers
jsm28 at cam.ac.uk
More information about the tz