[ntpwg] Further to the timestamping issue

[Rob Seaman]

Hi Dave,

> I don't understand your comment. The IERS can surely announce any  
> future leap epoch,

That was my comment.

> The WWV/H/B and ACTS leap warning indicators apply at the end of the  
> current month, which is how

> the leap warning bits are interpreted in NTP. From past experience,  
> server leap warning bits have had less than perfect reliability and  
> represent a possible vulnerability. This is why the upstratum  
> servers vote on the leap warning bits and are  reluctant to set  
> their own leap warning bits other than on the day of the leap.

As I said in some message, this heuristic seems reasonable to me - but  
moving this responsibility to zoneinfo as Steve Allen suggests might  
reveal even better trade-offs.

Rob