[LEAPSECS] Leap smear

[Ian Batten]

On 24 Sep 2011, at 1652, Nero Imhard wrote:


> 

> On 2011-09-20, at 09:11, Ian Batten wrote:

> 

>> On 19 Sep 2011, at 2350, Rob Seaman wrote:

>>> IThat requirement (description of the problem space) is that civil time-of-day is mean solar time.

>> 

>> So astronomers say.  No-one else cares, and if they should, astronomers are making an incredibly bad job of explaining why.

> 

> Well, it has been that way for ages, and it's the very reason why UTC is defined as it is: to stay near UT within a small margin. Second-guessing who cares, or having an opinion on who should care is rather condescending and also irrelevant to the (quite serious) issue of reliability and trustworthiness of standards.

> 

> If I, for whatever reason, chose UTC for use as a time standard, do I have to explain why I don't want its definition to change? Really? I think not. And I won't.


Well, it's changed several times already.  The original UTC as defined didn't even have seconds that were the same length as those in TAI, and the steering target when UTC was stepped in 50ms lumps was 0.1s.  The people who used those versions of UTC could rightly complain that the version we're using now is ten times less precise an estimate of UT1 than it used to be.  Had UTC remained the same, inviolate, since inception the argument that it should be left alone would have more weight; given it's changed both its rate and its offset mechanisms relative to TAI twice since its inception, the "you shouldn't change running standards" argument has already been lost.

ian