[LEAPSECS] L-format. Re: Crunching Bulletin B numbers (POSIX time)

Gerard Ashton ashtongj at comcast.net
Mon Feb 21 11:25:52 EST 2011


On 2/21/2011 10:49 AM, Paul Sheer wrote, in part:

>

> No, it's my own idea.

>

> Hereby released to the world:

>

> "2010-02-21 09:40:27 -0600 L0024"

>

> I hereby name it "timestamp L-format".

>

> It solves the problem of absolutely specifying a future time where you

> don't know how many leap seconds there will be between now and then.

>

> With the L included it means that the timestamp can be absolutely

> converted to TAI and represents a precise caesium second on the TAI

> scale.

>

> Without the L it means a regular Posix time relative to the start of the

> day - so called "appointment time".

>

> In this way it is backwardly compatible.

>

> However if leap seconds are published ten years in advance, it would

> mostly obviate a need for these new timestamps.

>

> -paul

>

So if we suppose that Paul had actually invented and used this format
long ago, and is
just now telling us about it, I would interpret the timestamp as
follows, neglecting the date
for conciseness:

09:40:27 -0600 = 15:40:27 UT

At the time Paul wrote down the time stamp he believed TAI - UTC would be
24 s as of the time and date stated.
He therefore was specifying 15:40:51 TAI.
However, this prediction turned out to be wrong, TAI - UTC = 34 s in Feb. 2010.
So if I was searching through records labeled with UTC, I would look for
an event at 15:40:17 UTC. Have I got that right?

Gerry Ashton






More information about the LEAPSECS mailing list