# [seqfan] Re: A159559

Benoît Jubin benoit.jubin at gmail.com
Tue Jul 28 21:50:28 CEST 2009

```On Tue, Jul 28, 2009 at 9:10 PM, Ray Chandler<rayjchandler at sbcglobal.net> wrote:
> You don't need a special case for a(1)=2 if you define as:
>
> a(n) is non-composite iff n is non-composite.
>
> Or replace "prime" with "non-composite" in your definition.

I'm not sure I understand: your definition would give the sequence
a(n)=n.  In a way or another, we need an initialization.
But this makes me think that the current comment that n is prime iff
a(n) is should be changed to n is composite iff a(n) is (so that it
still holds for a(1)=2).

Benoit

>
> Ray
>
>> I agree that setting a(1)=2 is natural (taking a(1)=1 gives
>> the positive integers, of course).
>> What do you think of this definition:
>> a(1)=2 and for n>1, if n is prime (resp. composite) then a(n)
>> is the smallest prime (resp. composite) greater than a(n-1).
>>
>> Benoit
>
>
>
> _______________________________________________
>
> Seqfan Mailing list - http://list.seqfan.eu/
>

```