"Lex Spoon" <lex at cc.gatech.edu> wrote:
> Stan Heckman <stan at stanheckman.com> wrote:
> > "Lex Spoon" <lex at cc.gatech.edu> writes:
> > > The epoch is left at Jan 1, 1901.  Does it really matter what we
> > > use? 
> > 
> > It matters because changing the epoch seems likely to break some
> > code. Which means I like your choice to leave it alone. :-) The only
> > (very slight) disadvantage to 1901 is that it makes it more difficult
> > to write method comments.
> 
> :)
> 
> 
> > 
> > If the epoch were after 1972, we could simply describe utcSecondsClock
> > as "return the seconds since <the epoch> minus leap seconds."  But
> > since it is is back in 1901, before the definition of UTC or the
> > official recording of leap seconds, I can't think of a short, clear
> > comment. I imagine unreadable comments like "return 86400 times the
> > number of days since Jan 1, 1901 GMT, plus the number of seconds since
> > UTC midnight today."
> > 
> 
> 
> *Sob*  Actually, I completely forgot about leap seconds.  Drats.  I
> don't suppose there is a time zone that ignores leap seconds?  So that
> it would actually be 86400 times the number of days, plus the number of
> elapsed seconds today?
> 
> 
> Lex

Were it that simple.  Yes, there is a time which doesn't use leapseconds, but
that is UT1.  The catch is that UT1, while continious, has a variable sized
second becuause of the Earth's rotation is non-uniform.

Leap Seconds are added to keep |UTC-UT1| < 1.

One could also use TAI, International Atomic Time (but in french word order)
but that doesn't sync too well with the watches that we all wear.  Currently 
UTC is about 30 secs behind TAI.

More gory details are at:

http://star-www.rl.ac.uk/star/docs/sun67.htx/node217.html

I do work with folks who have to know all of those so I can give it a poke and see what I can come up with.

cheers

bruce