Uncertain sequences (was Re: TR : Re : The OEIS will be on holiday for the rest of the year!)

[Jaap Spies]

N. J. A. Sloane wrote:

> 
>>And why are at least no less interessant sequences than supra not accepted?
>>(ex: 1903, 2303, 2307, 2311, 2315... "Exceptional years of 23rd december's
>>winter solstice since the beginning of the Gregorian Calendar in 1582" is no
>>less interessant than a "word" sequence).
> 
> 
> 
> OK, I can see that I would wrong to exclude that - please send it again!
> 
> 

May I utter a word of warning. The sequence mentioned is very much
unsure. Calculating solstices (and other phenomena) in the far future
is uncertain because of the unpredictable earth rotation.
First of all you have to define 23rd december (in which timezone?),
let it be in UT. What is the DELTA T, the difference between
terrestial time (TT) en UT?
UT is defined by the rotation of the Earth. DELTA T was negative
from 1871 to 1901, but increasing now. We had an extra second last
year before at Jan 1th 00.00h we could say "happy new year".
All calculations depend on an extrapolating model. DELTA T now is
approximately 65 s, but what will it be in the year 3000?

There is another uncertain factor in the calculations and that is
the planetary model. Some calculations use the VSOP 87 model and
others depend on the JPL ephemerides from NASA.

What policy do you have with sequences that may be changing in time?

Jaap Spies

(OK, I have to update my pynovas software).