[LEAPSECS] Leap Sec vs Y2K

[Rob Seaman]

On Dec 10, 2010, at 11:42 AM, Poul-Henning Kamp wrote:


> In message <3B33E89C51D2DE44BE2F0C757C656C8809CDA7EB at mail02.stk.com>, "Finklema

> n, Dave" writes:

> 

>> We can just let things be as they have been for nearly 40 years.  

> 

> Sure, no argument from here: Please shut down your Internet connection

> and any cell-phones you might have, and don't use them ever again :-)



(Passing over the obvious response that the Internet and cell phones have been demonstrated to function with the current definition of UTC.)

This was taken out of context:

On Dec 10, 2010, at 11:22 AM, Finkleman, Dave wrote:


> The Y2K effort was necessary.  Everyone knew that we could not just

> watch what might happen and catch up afterwards.  In the case of leap

> seconds, no one knows what the real consequences might be if we changed,

> and change is not necessary as it was for years that stopped at 99.   We

> can just let things be as they have been for nearly 40 years.


Which is to say that the schedule for Y2K remediation was forced.  We get to choose when to address a possible redefinition of civil timekeeping.  It is clear that consensus does not yet exist.

PHK continues:


> The crucial change in exactly the last 40 years, is that computers

> of all sizes are communicating and the applications we want them 

> to run for us, very much need to know and agree what time it is.


Not disputed.  What is the concept of operations for the pertaining timescale(s)?  What are the requirements?  Innumerable clocks based on different types of interval timekeeping as well as on earth orientation exist.  Pretending time doesn't come in different flavors is not going to work.  System engineering provides tools to reach consensus on the problem before entertaining possible solutions.  Why not use those tools?

Rob