[LEAPSECS] Cost: getting rid of GMT & discontinuing leap seconds

[ashtongj]

My replies are interspersed among Paul's questions.

On 2010-01-14 12:29 PM, p at 2038bug.com wrote:

>

>> it will be necessary to modify all [...]

>

> Why?


Because they will be wrong, civil time will not be based upon the mean 
time an Greenwich, for any meaning of the word Greenwich.


>

>> It will also be necessary to. modify

>> all computer programs which keep civil time [...]

>

> Examples of such programs with a material manifestation of a problem?

>


No such examples exist yet. However, once the difference between UT1 and 
UTC becomes great enough to matter in daily affairs, a person who would 
benefit by interpreting a timestamp applied by some computer, where said 
timestamp contains "GMT", can argue that it MUST mean UT1, and can't 
possibly mean UTC, because the meridian for which UTC is the mean solar 
time does not pass through Greenwich. For example, an advertisement for 
bids states bids must be received by 24:00:00 UTC December 31, 2026. The 
computer that receives such bids marks a bid as having been received at 
00:00:02 January, 2027 GMT. The bidder sues, demanding that the  bid be 
accepted on the basis that GMT is UT1, UT1 runs 6 seconds in advance of 
UTC, so the bid was received at 23:59:56 UTC and met the deadline.

Gerry Ashton