[LEAPSECS] An example

Warner Losh imp at bsdimp.com
Tue Nov 2 16:19:37 EDT 2010

On 11/02/2010 12:26, Finkleman, Dave wrote:

> My colleague, Sal Alfano, included in distribution observes with

> relevance that the state of Indiana in the U.S., legislated that Pi=3

> for convenience. Eliminating leap seconds for convenience is similar.

> In the former case, it meant that no circle would ever close, lacking

> (pi/3-1) fraction of the circumference. So it will be with time.


Is this analogy really apt? the former was a mathematical fact that
could be shown to be flawed, and the degree to which it is flawed could
be understood. The latter is a matter of how to synchronize two
different clocks, and the magnitude to which DUT1 can grow. While it
was convenient for many communities to bake DUT1 coarsely into a widely
propagated time scale, it was inconvenient for others. As the relative
sizes of these communities change, it make sense to reevaluate what we
bake into the widely propagated time scale.

There's nothing magical about leap seconds. They are not an inherent
property of time. They are a means to the end of relating the number of
SI seconds that have elapsed to the angular orientation of the earth.
One could make the exact opposite argument about leap seconds: They
create a number of small holes in time and therefore distort time in a
way that is all about convenience to some. They make one falsely think
that the SI second is somehow the same, or close enough to the same, as
one second of angular motion by the Earth. Is that also a lie, just
like pi being 3 is a lie?


More information about the LEAPSECS mailing list