[seqfan] Re: A008477 errors

[Hans Havermann]

Jaroslav's flagged indices are those for which a(n) <= 0. Using the  
sequence's Mathematica program, in addition to these 48, I find n =  
6728 is also at variance. If that is so, then a replacement b-file is  
here:

http://chesswanks.com/seq/b008477.txt

On May 20, 2012, at 1:53 PM, David Applegate wrote:

> Jaroslav Krizek has pointed out:
>
>> In table of sequence A008477 are erroneous values a(n) minimally  
>> for n = 961, 1369, 1681, 1849, 1922, 2209, 2738, 2809, 2883, 3362,  
>> 3364, 3481, 3698, 3721, 3844, 4107, 4418, 4489, 4805, 5041, 5043,  
>> 5329, 5476, 5547, 5618, 5766, 6241, 6627, 6724, 6727, 6845, 6889,  
>> 6962, 7396, 7442, 7569, 7688, 7921, 8214, 8405, 8427, 8649, 8836,  
>> 8978, 9245, 9409, 9583, 9610.
>> Best regards,
>> Jaroslav Krizek, jaroslav.krizek at atlas.cz.
>
> In particular, it appears that 32-bit arithmetic was used for
> computing the entries in the b-file, resulting in overflow errors.
>
> Perhaps someone has time to compute corrected values?