[LEAPSECS] alternative to smearing

[Zefram]

Preben Norager wrote:
>                                 if leap seconds are abolished, then the
>gregorian calendar is implicitly forced on, because of the implicit
>connection between the international atomic timescale, and the gregorian
>calendar.

There is no such connection.  TAI times are conventionally described by
dividing up the time scale into notional days, which are then labelled
according to the Gregorian calendar.  But this is *only* convention,
and two separate conventions at that.  TAI's actual nature is of a linear
sequence of seconds, which one can label in any regular manner at all.

If the use of UTC (with leap seconds) were abolished for civil time,
in favour of a TI time scale that amounts to TAI plus some fixed offset,
then it would of course be necessary for TI to be divided up into days.
This would be required in order to plug TI into the existing civil
horological mechanisms that have evolved with UT.  It would look much
like the same as how TAI is conventionally represented.  But even
with the division into days having become necessary rather than merely
conventional, there is still a totally free choice as to which calendar
to use to label the days.  Indeed, it is inevitable that each locality
adopting TI would continue to use whatever calendar it had formerly used
with UT.

I am mystified as to how you think the choice of time scale here forces
the choice of calendar.

-zefram