[LEAPSECS] LEAPSECS Digest, Vol 71, Issue 4

[Warner Losh]

On Sep 4, 2012, at 9:16 PM, Rob Seaman wrote:


> On Sep 4, 2012, at 11:06 AM, Warner Losh wrote:

> 

>> The ntp announcement thing is just one more brick in this rather large wall.

> 

> Nah, just more spray paint.  It appears that the "ntp thing" is nothing new; if there are bugs fix them.


Pervasive bugs that last years across multiple implementations suggest a flaw in the underlying spec. NTP isn't the only program/protocol to have this problem with leap seconds.


> Redefining UTC to be something other than Universal Time would not provide additional access to leap less time scales; we already have TAI and GPS.  Abolishing Universal Time would rather simply deny access to solar time.


Leap seconds, as implemented today, are fragile.  There's no denying that.  Many folks get them wrong, despite years of scolding and efforts.  The "what to do about it" is a separate issue though. Dropping leap seconds, or changing how they are done to have a longer-term average that's right wouldn't deny access to solar time: it would just change the magnitude of the adjustment that's needed today.  Might break other things, but it wouldn't deny access to solar time. But that wasn't the point I was trying to make.  The point is simply leap seconds are fragile.


> Every wall has two sides.



True enough.

Warner