[LEAPSECS] Degrees of Accommodating Time Based on Earth Rotation

Poul-Henning Kamp phk at phk.freebsd.dk
Tue Nov 2 14:04:10 EDT 2010

In message <3B33E89C51D2DE44BE2F0C757C656C88099AAC8C at mail02.stk.com>, "Finklema
n, Dave" writes:

>It is a hierarchy each level of which is sufficient for a range of

>applications. Every time we solve dynamical equations, we are defining

>a unique time scale and time interval based on things such as analytical

>discretization and computational architecture.

That is a good summation and an interesting way to look at it.

I would put leap days on top of the heap of approximations, even
though that is a different physical rotation, simply so they appear
on the list and people don't ask stupid questions.

So here is an instructive little table:

Finagle Subject to
Factor Unit/Resolution politics since
Leap days 86400 sec. 2000 bc.
Timezones 3600 sec. 1884
Leap seconds 1 sec. 1958/1972
DUT1 (Bul. A) 1 microsec. Not yet

Politization of these finagle factors have always been
motions toward higher predictability:

We went from priests announcing Easter based on astronomical
observations, to putting them on a mathematical formula that could
predict the dates for future, as far as we can see it.

We went from random local timescales to UT[C] with 24 standardized

We went from rubber seconds to identical seconds.

In that light, it seems awfully logical that we would want to
nail the length of the days down also...

Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG | TCP/IP since RFC 956
FreeBSD committer | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.

More information about the LEAPSECS mailing list