[LEAPSECS] 25000 years, or my money back?

Magnus Danielson magnus at rubidium.dyndns.org
Sat Mar 7 14:13:50 EST 2009

Steve Allen skrev:

> Another "atomic clock" advert; good for 1 second in 25000 years.

> http://news.thomasnet.com/fullstory/556476


> If it fails by one second before then should I sue? Who?

> NIST? Control Company of Texas?


> How exactly does one get ISO 9001 certs that extend 25 millennia?


> Or maybe they just noted that the historical variations in the rate of

> what we now call TAI are just about that big.

I don't think they actually means that TDEV(25000y) + MTIE(25000y) < 1 s
or that expected lifetime of the clocks, NIST transmission service or
the user to be 25000 years either. It's all about some product people
trying to make our incomprehensible jibberish understandable to the lay
person while trying to say that this is a bargain deal. Scientific
analysis usually breaks before you look at the title of such a thing.

So in the end all you can do is to laugh at it and go about the usual

I expect the LCD to fade long before 25000y for instance.

Still, it can be a usefull clock for many not being time-nuts but just
want a fair feel of time of day without reset hazzle and for about 10
years or so. Should fit the need to know when to run for the bus etc.


More information about the LEAPSECS mailing list