[LEAPSECS] What is GMT?

[Michael Sokolov]

Tony Finch <dot at dotat.at> wrote:


> GMT was discontinued decades ago and has not had a coherent meaning for

> even longer.


For me "GMT" has a very simple meaning: it basically means "the exact
timescale doesn't matter, it can be anything as long as it comes from
someone like Rob Seaman and NOT from someone like PHK".  UT1 would be
perfectly fine, but so would UT2, UT0, UTC-SLS or just about any other
version, OTHER THAN what PHK and the ITU are seeking.

The system time on my non-POSIX UNIX systems is defined to be GMT, not
UTC.  Right now UTC disseminated via NTP (obtained from public
Internet NTP servers) serves as an acceptable realization of GMT, but
this will no longer be the case if the bastards in the ITU have their
way this month.

If the ITU bastards have their way, I'll be forced to drop all my
other projects and *scramble* to build my own non-UTC GMT-generating
apparatus before the first embargoed leap second occurs.  The
apparatus I have in mind is envisioned as being very similar to a
typical stratum 1 NTP server (GPS in, Ethernet out), but with a
special twist: instead of applying the "UTC offset" (or LS count or
whatever it's called) transmitted by GPS, apply a different offset
controlled by ME locally.  I will then have to manually adjust this
offset periodically to steer my timescale to the true GMT, i.e., to
UT1 or whatever else happens to be convenient.  And when this offset
needs to change, it will be done via a rubberization process similar
to UTC-SLS.

MS