[LEAPSECS] 1.26 microseconds

[Zefram]

Rob Seaman wrote:

>If variation of length-of-year is a significant issue it has not been

>stated previously in this context.


I think the issue there is not about *variation* in the length of
year but about *which* year is referred to.  When referring to time
rather than orbital mechanics, astronomers have traditionally used a
unit called "year" defined as 31557600 s.  (That's based on the length
of the Julian calendar year along with a nominal day of 86400 s, and
of course no one claims that it precisely matches any Terran orbital
parameter.)  Recently there's been some fuss about defining a similar
time unit for use in other sciences, and the IUPAC (chemists) and IUGS
(geologists) have jointly defined a unit called "annus" defined as
31556925.445 s.  (That's based on a measurement of the Terran tropical
year somewhere around the year 2000, I'm hazy on the details.)  I read
about it in New Scientist, <http://www.newscientist.com/article/
dn20423-push-to-define-year-sparks-time-war.html>, and that has a
reference to the IUPAC journal article.

It seems rather silly to me that (a) they would seek such precise
agreement with a physical parameter that is, after all, context
sensitive and slightly variable, and (b) they would ignore the very
similar preexisting unit from astronomy.  That doesn't actually seem to
be causing much controversy, though.  The main controversy seems to be
that geologists object to the notation, where (for example) "1 Ga" now
refers to a duration of 1 gigaannus, whereas preexisting geological usage
has "1 Ga" refer to a *point in time*, a billion years before present.
The geological usage is inconsistent with the principles of SI, and I
think they had to face that sometime, but the symbol clash is moderately
gratuitous.

-zefram