[seqfan] Re: Prime signature of 1, and second signature

Marc LeBrun mlb at well.com
Sun Jun 10 20:42:12 CEST 2012

>="David Wilson" <davidwwilson at comcast.net>
> ...it is my observation that it is better to omit elements entirely than
> kludge them. Better to start at index 2 than to treat 1 inconsistently.

I second this advice, although it took me a while to overcome the natural
mathematical impulse to try to address the most general cases.  In the OEIS
it's actually best to resist the temptation to extend  sequences "backwards"
when the interpretation becomes in any way problematic (controversial,
ambiguous, kludgey, arguable, etc).  For one thing doing this can thereupon
create a temptation for someone to later submit an otherwise identical
sequence that differs only by that term, which then becomes a maintenance
burden (eg which do you cross reference, etc).  It also might cause
automation like superseeker to miss a legitimate "hit" on the sequence,
because, for example, some transform it tries might not naturally produce
that "artificial" value, or not even try because it isn't applicable to the
thus-extended range (eg Mobius transform, since mu(n) only applies to n>0).

More information about the SeqFan mailing list