[LEAPSECS] ISO Influence

[Warner Losh]

On 12/18/2010 10:49, Finkleman, Dave wrote:

> For everyone to criticize, I have almost convinced the USAF to issue a position statement to OSD and the State Department pleading that UTC not change from the current paradigm.  The rationale is that UTC is called out as the mandatory service for event time tagging and synchronization.   As such, contractors must implement UTC in every Air Force system.  If UTC changes, there will be consequences throughout current and emerging systems.  Any change would mandate extensive assessments and mitigation.  No change being the best alternative for them is a no-brainer.

I know that the high precision time systems will likely work better if 
UTC drops leap seconds.  That's one area I know will be cheaper if leap 
seconds are simply discarded.  As someone who has worked in that field, 
I can tell you that each leap second costed the company I worked for 
between one and two man months to validate and correct problems found in 
the software.  Sure, the software should work, but in every leap second 
simulation I was involved with, there was some problem.  And often there 
was some problem 6 months after the leap second due to unwise 
assumptions that were made to fix the leap second case.

It is far from a 'no brainer' that not changing is the best, most 
cost-effective choice for that segment.  It will be no worse than it is 
today if no change is made, but it likely would be much better if a 
change to eliminate leaps was made.  Other segments may differ, of course.

Warner