[seqfan] Re: OEIS Wiki: Offsets
Joerg Arndt
arndt at jjj.de
Tue Apr 14 20:12:12 CEST 2015
* Benoît Jubin <benoit.jubin at gmail.com> [Apr 14. 2015 19:47]:
> >> "Think long and hard whether a(0) is naturally defined. If yes, add it."
>
> Among the first nine OEIS sequences, A000001(0)=0 and A000003(0)=0 should
> therefore be added.
>
> >> Back to the original subject, I think it is worth while mentioning the
> example of sets & lists, and also insist on the example of general
> functions defined from some given value on (which may be more often 0 than
> 1, but also something else).
>
> I like to think that there are two big purveyors of integer sequences:
> combinatorics and number theory. Sequences coming from combinatorics are
> often defined for all natural numbers (for instance, the number of group
> structures up to isomorphism on a set of size n) whereas sequences coming
> from number theory are often defined (or at least, interesting) for nonzero
> arguments, mainly because 0 is absorbing.
>
> Benoit
>
Another approach/guideline to "1 or 0?" is "multiplicative or additive by nature?".
Indeed we do have "multiplicative by nature" sequences with a term a(0)
that should not be there IMHO.
My own rule "give a(0) whenever you can" (which I did not mean to go
into the Wiki page) comes from the unfortunate omission of a(0) or
column-0 in many sequences about combinatorial objects.
Another telltale sign for a(0)=1 appears as
G.f.: (some natural expression) - 1.
Now if anybody challenges me to define "natural"/"by nature" I'll send
black helicopters in your general direction.
Best regards, jj
>
> [...]
More information about the SeqFan
mailing list