Hi Bill,

Thanks for the correction. You may have noticed that your revised
"zero point" of 63.829 seconds matches that for 2000 from the table I
posted. This isn't a coincidence. To get that linear formula, all
I did was to take Delta-T for 2004 and 2000, subtract the latter from
the former, and say, "It's increasing at that rate every four years,
i.e., .1862 seconds/year." I then meant to use the 2000 value as the
zero point, and accidentally put in the 2004 one.

John, your suggestion of putting the 32.5 seconds/century^2 term
in seems reasonable to me. It matches the long-term behavior of Delta-T
over millennia, and we can hope that this is passably consistent.
I certainly can't see any reason it would be _worse_ than the linear
methods proposed.

-- Bill