Error-finding program

[T. D. Noe]

At 11:57 PM -0400 10/24/06, David Wilson wrote:
>T.D. Noe just found an error in A057886, to wit:
>
>%U A057886 825,836,946,957,1078,10,90,1222,1234,1378,1391,1547,1560,1729,1743
>
>The element 1090 was split into two elements 10,90 by a typo or inaccurate
>submission.
>
>In this case, looking at the graph immediately raises suspicion, because a
>general trend is punctuated by two very small values.
>
>Some very common errors in OEIS sequences are:
>
>- Lexically splitting elements
>- Lexically joining adjacent elements
>- Omitting elements
>- Transposing elements
>
>In sequences that have smooth growth, any of these errors potentially
>induces a suspicious disruption in that growth.
>
>Might it be possible to create an AI that would detect and flag such
>suspicious elements?


I just finished writing a simple program to find these types of errors.  My
first hit is A000708, in which a(20) is split.  I'll submit the corrections
as comments as I find them.

Tony