[LEAPSECS] Straw men

Rob Seaman seaman at noao.edu
Mon Jan 9 15:25:26 EST 2012


Ian Batten wrote:


> On 9 Jan 2012, at 1845, Rob Seaman wrote:

>

>> Yes, I expect issues to continue to arise decades hence in currently deployed systems and processes. There are, for instance, ongoing Y2K-related issues.

>

> [Citation Needed]


I was thinking of this: http://catless.ncl.ac.uk/Risks/26.46.html#subj16, but apparently not: http://catless.ncl.ac.uk/Risks/26.47.html#subj11 - will forward citations if I come across any. Absence of handy examples has no implications for the preceding statement.


>> And since time-of-day will fundamentally remain mean solar time,

>

> [Citation Needed]


See http://www.cacr.caltech.edu/futureofutc/preprints/02_AAS_11-661_Seaman.pdf, pages 7 & 8. Also various messages on this list.


> The set of equipment that needs any time other than civil time is vanishingly small. If you need it directly, you have a national LF service. If you need it slightly less directly, you have a GPS receiver and know what the offset is. And if you're normal, NTP will do the job just fine.


NTP does a good job with the abnormal cases, too. I highly recommend the 2nd edition of Dave Mills' book where he adds space applications.


> So long as those all tick the same thing, its relationship to the rotation of the earth is, +/- several hours, irrelevant.


This is an unrelated assertion with which I (obviously) disagree.


> No-one cares what the relationship between their watch/clock/computer and the sun is at anything other than the grossest scale,


Then it would be easy to document this assertion. Rather, others have documented otherwise:

http://futureofutc.org/preprints


> It's different for astronomers. The question is, should the rest of us be obligated to use a time scale which causes us difficulties which we could fix easily were it not for the needs of a small scientific community?


And this inverts the actual situation. Civil time has always been astronomical time, that is Universal Time. The "Draft Revision to ITU-R Recommendation TF.460-6" seeks to change this status quo, not the other way around. There is an implicit and untested assertion that the "rest of us" will be better off in some way. There is an assertion of an obligation to use UTC when TAI and GPS have existed throughout for applications that need/desire them. There is a assertion of an "easy fix" that is not described, certainly not in the ITU proposal. There is some implication that the importance of the scientific use cases (large) somehow scales with the number of scientists (small).


> Indeed, I got six months' salary as a bonus for my contribution to sailing it through it without a single issue in my employer's business.


Apologies for the personal question.


> What were you doing for Y2K?


http://iraf.noao.edu/projects/y2k/y2kplan.html


> You do realise that "The astronomical community started preparing for Y2K in 1996 and barely had enough time" isn't something to be proud of, don't you? We'd _finished_ by 1996, having started the programme in about 1989.


Glad to hear it. I am indeed proud of the astronomical community and its response to Y2K. Would have been nice to have funding commensurate with the urgency. Would be nice not to be faced with an artificial crisis a dozen years later.


> So a little less of the "the astronomical community are the only competent engineers, the rest of you are just charlatans" would be nice.


I don't believe I've said anything similar. I apologize for being pedantic and not an exceptionally good communicator. The fact remains that very little of what has been discussed here by anybody is reflected in the actual "Draft Revision to ITU-R Recommendation TF.460-6".


> When was the _rate_ of UTC such that 86400s == 1 mean solar day?


Interesting question. I will presume you mean "86400 SI-seconds", since by definition (nearly) every day has 24 hours of 60 minutes of 60 seconds. Anyway, the mean value theorem says this occurred at least once. Likely it has toggled back and forth several times over the past couple of centuries. Otherwise see:

http://ucolick.org/~sla/leapsecs/amsci.html

Which shows about 14 instances with a larger number of instantaneous crossings over the past decade. Over the long term there will be an odd number of such instances since the day started shorter and is getting longer, eventually to never occur again.

Rob Seaman
National Optical Astronomy Observatory



More information about the LEAPSECS mailing list