[LEAPSECS] Straw men

[Rob Seaman]

Warner Losh wrote:


> On Jan 9, 2012, at 11:17 AM, Rob Seaman wrote:

> 

>> Warner Losh wrote:

>> 

>>> Actually it is a fair engineering question: Why pay the cost of leap seconds when we can keep civil time aligned to the sun with time zones once every N generations.

>> 

>> First off - what is that cost?  There is an absence of costing efforts for either leap seconds as they are, or redefined UTC as is proposed.

>> 

>> Also, an "engineering question" is not the same as a "question uttered somewhere in the vicinity of an engineer".

> 

> You keep saying that Rob, but you keep not making sense.

> 

> You keep trying to cook the process so that it only winds up with your solution.

> 

> That's not engineering.


What I have said on the issue is:

	http://www.cacr.caltech.edu/futureofutc/preprints/02_AAS_11-661_Seaman.pdf

I am well aware that my attempt to explicate the difference between engineering requirements and specifications has not been entirely successful.  I am not trying to "cook the process".  The assertion of the notion that timezones will magically fix "the problem" is rather a confirmation of my point.  There would be no problem, no engineering requirement demanding a fix, if time-of-day weren't the same thing as mean solar time.

That is a requirement.  It is a fact about the problem space.  How that requirement is met is in the solution space.  We can certainly entertain different ways of meeting the requirements, different approximations to time-of-day.  But the draft revision being voted on next week discusses precisely nothing.  It converts UTC to TAI (+offset) and ceases the currently required transmission of DUT1.  Period.  That is an example of cooking a process.


> And this question is highly insulting.  Knock that off.


The only question above is:  "What is the cost?"  I guess you mean that my parsing of your own question is a personal insult. I apologize.  It certainly wasn't intended to be.  Let's try again:


>>> Actually it is a fair engineering question: Why pay the cost of leap seconds when we can keep civil time aligned to the sun with time zones once every N generations.


1) What is the cost of leap seconds?

2) There appears to be an assumption (that astronomers themselves might find insulting) that redefining UTC itself has no cost.  Actually it will very clearly cost many millions of dollars across the astronomy community.  And estimates from China and from the aerospace community are each in the hundreds of millions of dollars.

3) There is an assumption - without benefit of any documentation whatsoever - that timezone adjustments can indeed serve this stated purpose.  If this is obvious (I don't find it such), then it should be easy to write a description of how this would work.

	a) Please address the northern/southern hemisphere issue.
	b) Only a small fraction of the world observes daylight savings, please address what the others might do.
	c) Please indicate how this addresses the problem of acquiring Universal Time in between adjustments.
	d) And what is the cost of implementing this?

Currently the zone system is tied worldwide to an underlying mean solar time standard.  The notion is to fragment this such that different localities will separately realize whatever synchronization they deem necessary.  That is, a single common civil timescale is being replaced by a myriad of separate realizations.  Please discuss the resulting issues, e.g., record-keeping, long-term scientific planning (astronomers are not the only ones who looks centuries or millennia forward and backward), etc.

Perhaps I was a bit snarky about the meaning of an "engineering question".  An engineering question would be something addressed using an engineering process.  The "Draft Revision to ITU-R Recommendation TF.460-6" shows precious little evidence of such.

Rob