[LEAPSECS] WP7A status and Re: clinical evidence about time and sun

[Zefram]

Rob Seaman wrote:

>historians and long term planners (and yes, some folks do think  

>thousands of years into the past and the future) need a coherent  

>system for tracking clock relationships between countries and centuries?


Historians have to relate whatever time scale was historically in use
to current time scales.  This is a complex job even for pre-industrial
history, given the multitude of calendars historically used.  To mention
but a handful of complications: many societies have labelled years
according to reigns of monarchs; Julian calendar users had widely
differing opinions about when the numbered year starts; some calendars
had irregular intercalation of months.

Obviously historians need to know whatever they can about the time
scales used by the society they're studying.  Future historians looking
at today will be interested in sub-second timing of some of the events
that we record, and for that purpose they'll need to know when the
leap seconds are in UTC.  This is directly analogous to knowing which
years in the Roman civil calendar had a Mercedonius.  Similarly, to
coordinate sub-hour times between countries, historians need the Olson
timezone database.  I don't think we can optimise our time scales to be
easy on future historians, because we don't know what kind of time scale
they'd prefer.  But it is our duty to record the relationships between
UT1, TT, and civil time.

When planning for the future, similar considerations apply, except that
the future civil time scales are all hypothetical.  We must use UT1 to
plan some things and TT to plan others.  Converting these to the civil
time scale of the time in question must be left to future generations.

So overall I don't think these issues form any argument one way or
another about what sort of time scale we ought to use, or plan to use,
for civil purposes.  That's just local convention.  Over long periods of
time we do best thinking in terms of the naturally-occurring timescales,
of both flavours.


>                                  telling, for instance, the  

>historians of the world that they will never again be able to keep  

>track of relative clock time during battles and negotiations and  

>journeys.


I don't see how we'd be telling historians this sort of thing under any of
our scenarios for civil time.  Historians can always perform conversions.
Unless we destroy all the records of what we did.


>There are two kinds of time.  They demand two clocks.  UTC manages to  

>convey both,


UTC only conveys interval time if you have the list of leap seconds.


>will still need two kinds of clocks.


I think the more fundamental issue is that we will, any way round, need
data that relates the two flavours of time.  A single clock reading
can't give both types of time in the absence of such knowledge.

-zefram