[seqfan] Re: Correcting A002932 (n-step walks on square lattice)

[Joseph S. Myers]

On Mon, 22 Nov 2010, Russ Cox wrote:

> Typically if a sequence is wrong but existed for a long
> time before that was discovered, it gets all the information
> stripped out but leaves a tombstone behind with the
> erroneous sequence data.  Search for keyword:dead
> or erroneous to find examples.

There appear to be two variants around: either put the corrected sequence 
at a new number, or correct the old one (presuming that any links and 
cross-references want the correct version, not the erroneous one) and put 
the erroneous one at a new sequence number for the sake of anyone 
searching using the terms from the original reference.

> That said, I'm skeptical about this one being wrong.
> I'd want to be very sure before asserting that the data
> given (396204) is wrong, since the sequence data
> appears to have gone unchanged since the 1973 book.

That's why I asked on the list for someone else to do a calculation of 
this sequence.  Though since both "formula" comments are unambiguously 
wrong, my suspicion is that no-one actually tried to recalculate/extend 
this one since it appeared in the 1973 book.  (The table of values in the 
1992 reference doesn't go as far as the table in the 1961 reference from 
which the current terms are taken.)

-- 
Joseph S. Myers
jsm at polyomino.org.uk