[LEAPSECS] internet drafts about zoneinfo

[Joe Gwinn]

At 1:55 AM +0000 3/7/11, Tony Finch wrote:

>On 5 Mar 2011, at 16:27, Joe Gwinn <joegwinn at comcast.net> wrote:

>>  At 2:27 PM +0000 3/5/11, Tony Finch wrote:

>>>  On 4 Mar 2011, at 22:13, Joe Gwinn <joegwinn at comcast.net> wrote:

>>>>

>>>>  As I have said many times, in theory POSIX time is a form of 

>>>>TAI, being a constant offset from TAI,

>>>

>>>  Wrong in every respect.

>>

>>  Could you be more specific?

>

>There is no "in theory" since POSIX is a practical specification.


The in theory is the POSIX standard.

The practice is that people use GPS and NTP to steer the clock.



>The specification says time_t is based on UTC.


Yes, in a very limited sense - only the timescale origin is defined 
as a particular UTC-defined instant.

The progress rule is not UTC, as there are no leapseconds of any kind.



>POSIX time has never been a constant offset from TAI.


It's true that the POSIX standard does not come out and say this. 
However, the seconds are derived from a hardware clock that tries to 
implement seconds of identical length (to the accuracy of the clock 
hardware), and does not attempt to follow the motions of the planet.

So, we have an instant in time that is forever fixed, and a progress 
rule that counts out uniform seconds (ideally SI seconds the standard 
says in the Rationale).  Which timescale most resembles this 
description?

It's useful to read the actual words. 
<http://pubs.opengroup.org/onlinepubs/009695399/>  See sections 3.149 
(Epoch) and section 4.14 (Seconds Since the Epoch).  Note that the 
statement is the this time is "related to" UTC; this is standards 
bafflegab.  The actual requirement is "As represented in seconds 
since the Epoch, each and every day shall be accounted for by exactly 
86400 seconds."

Also see the rationale sections paralleling these two sections.  In 
particular, the rationale section A.4.14, which comes out ant says 
that POSIX time is not UTC.



>What has been said many times is that it might be nice to change 

>POSIX time to be based on TAI but if you try it you find that this 

>breaks most timekeeping code.


Why?  (Not that the POSIX community wants to be tied to TAI, but 
their objection is not that the code would break.)

Joe Gwinn