[ntp:hackers] ITU and Leap second elimination

Warner Losh imp at bsdimp.com
Tue Oct 1 21:05:10 UTC 2013


On Oct 1, 2013, at 10:34 AM, Mike S wrote:

> On 10/1/2013 11:14 AM, Warner Losh wrote:
> 
>> So while it is not pedantically speaking correct, there is a
>> non-uniformity around leap seconds, and that non-uniformity is the
>> root of many problems.
> 
> Meh. There's non-uniformity for the number of days in a month. There's non-uniformity in the number of days in a year.

Those have highly regular rules that are well known and basically unchanging on a time scale of centuries. Leap seconds don't have such rules. They happen sometimes, and not other times.

Days/Months/Years have also existed for at least 3k years, while leap seconds barely 1/100th of that, so the rules are known by everybody. Leap second rules are known only after you do a google search and filter out any bad info that comes up.

> Non-uniformity is not the problem, faulty systems which don't accommodate it properly, are. We've had leap seconds for 40 years. They predate POSIX by more than 15 years.

POSIX has known about them and rejected them.

> And no, it's not being pedantic to correct the use of improper terminology, when that terminology is being used as part of an argument. It's not like "monotonic" has some other casual, colloquial meaning.

It actually does. Informally, monotonic means increasing in the way you'd expect it. That's why so many people use it, incorrectly, to describe the leap second non-uniformity. So it's in the semi-technical set this happens.

Warner



More information about the hackers mailing list