General Sequence Submission Advice
Chuck Seggelin
barkeep at plastereddragon.com
Fri Oct 24 03:49:20 CEST 2003
Hello Fellow SeqFans,
I need a little advice before I submit some sequences. Assume you have
found a new sequence you feel is interesting. The sequence involves an
algorithm that divides all the positive integers into two groups:
Set 1: f(x) = y (and assume y is never 0)
Set 2: f(x) gives no result
That there is a set of numbers which have a property such that f(x) cannot
be performed on them seems interesting. Likewise the results of the
algorithm (the y values) are themselves interesting.
QUESTION 1: Is it best to submit this as one sequence or three?
As One Sequence:
A0001: a(n) = f(n) or 0 where f(n) gives no result.
As Three Sequences:
A0001: a(n) = The n'th number m where f(m) gives a result.
A0002: a(n)=f(A0001(n)).
A0003: a(n) = The n'th number m where f(m) gives no result.
The former captures the same information with a smaller footprint, but the
latter is more likely to be found by a lookup. Also the use of zero as a
"signal value" in the former makes me a little nervy because at a glance it
implies f(n)=0 for certain n's upon which f() is not applicable.
QUESTION 2: Assume there are two distinct flavors of f(): f1 and f2. The
number of sets doubles:
Set 1: f1(x) = y1 (y1 is never 0)
Set 2: f2(x) = y2 (y2 is never 0)
Set 3: f1(x) gives no result
Set 4: f2(x) gives no result
And each pair of sets could result in one or three sequences. But would
sequences that compare the sets be useful? In other words:
A0006: x returns a result for both f1(x) and f2(x)
A0007: x returns no result for both f1(x) and f2(x)
Obviously I ask because I am doing some analysis that is yielding results
very much like what I describe here and I want to "package" it in the most
useful way before I submit it to the OEIS. Thanks in advance for your
advice! :)
-- Chuck Seggelin
More information about the SeqFan
mailing list