[seqfan] Re: A169902: all ones(?)

[Andrew Weimholt]

On Fri, Jul 9, 2010 at 7:37 PM, zak seidov <zakseidov at yahoo.com> wrote:
> May be it'd be more clearer:
>
> %N A169900 Earliest sequence such that a(i)*a(k) divides
>  a((a(i)+a(k)) for all i>=1, k>=1
>
> but still this does not preclude from "all ones" case
> a(i)=1 for all i>=1
>
> Zak
>

That's a completely different definition.

a(i)*a(k) | a(a(i)+a(k))  for all i>=1, k>=1

IS NOT THE SAME AS

i*k | a(i+k)  for all i>=1, k>=1

The first definition (yours) allows for a sequence of all ones.
The second definition (mine) does not allow for a sequence of all ones.

i*k | 1 fails for i>1 or k>1

Andrew