[LEAPSECS] Leap smear

[Rob Seaman]

On Sep 19, 2011, at 11:23 PM, Warner Losh wrote:


> There are many more steps possible, up to and including just going with atomic time and forgetting the sun.


Either you are asserting:

1) the SI-second is "close enough" (for the foreseeable future), or

2) we could pick any random period of time and call it a "day".

If #1, the thing that it is close enough to is actual time-of-day and the proposal is to shave fractions of a second from the actual time-of-day.  Thus you are acknowledging that there is an actual definition of the word day different from the pretend day that the ITU is attempting to substitute.

Or if #2, demonstrate how the word "day" could mean 50,000 SI-seconds or 100,000 SI-seconds or anything other than a number very close to 86,400+epsilon.

I do not say the ITU won't succeed in passing this insipid measure or that they can't cheat some of the people some of the time.  What I am asserting is that there are implications from such an action.  Those implications are there precisely because there is a requirement that would not be satisfied by the proposal.  That requirement (description of the problem space) is that civil time-of-day is mean solar time.

Some of those implications are a large expense to astronomers.  Others are that at some point in time an action will be needed by some civil authority to accommodate the resulting divergence of the promulgated definition of the word "day" from the actual definition of the word day.  A coherent engineering proposal would address the implications of adopting the proposal.

Rob