[LEAPSECS] WP7A status and Re: clinical evidence about time and sun

[John Cowan]

Rob Seaman scripsit:


> Because the past remains with us, and the future requires planning.   


What the past tells us is that prediction is very difficult, *especially*
about the future.  At sea, where no one cares much, the timezone mappings
have remained stable; on land, it's a different story, with multiple
changes for most of the 384 Olson time jurisdictions.

The median number of changes over the century-give-or-take of standard
time (exactly when standard time begins depends on the jurisdiction)
is 3, minimum 1, maximum 17 (!), mean 3.94, standard deviation 2.88.
Even if we ignore all changes before the Unix epoch, we still have 545
changes in 231 jurisdictions.


> By discarding any stationary mapping from local clocks (and calendar)  


You cannot discard what does not exist.


> to an underlying "universal" timescale, historical provenance and long  

> range planning acquire a spatially dependent error term that grows  

> with time.  


Long-range planning does not and cannot extend 3.5 millennia into
the future.


> In short, time has duration not just discrete epochs to worry about.


Local time, even if DST effects are excluded, is unfit to measure
duration: it has too many discontinuities.

-- 
John Cowan    cowan at ccil.org    http://ccil.org/~cowan
        Sound change operates regularly to produce irregularities;
        analogy operates irregularly to produce regularities.
                --E.H. Sturtevant, ca. 1945, probably at Yale