Re; The OEIS will be on holiday for the rest of the year!

[Joshua Zucker]

On 1/6/06, N. J. A. Sloane <njas at research.att.com> wrote:
> In any case, let me say that I accept essentially every sequence
> that is submitted!
>
> NJAS

Yes, but I don't think you should -- particularly in the cases where
the author can't be bothered to compute more than a few terms (of  a
sequence which is computationally easy), or in cases where the author
can't be bothered to edit carefully and express the contents clearly.

I find myself wavering on these sequences -- maybe NJAS or someone
else here can give me some advice.  Say someone submits a family of a
few dozen sequences, all of which are with keywords "base" and not
very interesting in my opinion -- but there's only a few terms of
each, so there's "more" given there, and I have a program which will
dump out a bunch more terms in a few milliseconds.

What should I do then?

Submitting more terms might make the author think that I think these
sequences are interesting, thereby encouraging the author to do yet
more of the same, but not submitting terms leaves them as these short
little stubs of sequences ...

I came across them because I skim every once in a while through
sequences with the "more" keyword and see if I can write a program to
generate more terms without very much work on my part.

By the way, in the past I've always used the comment page to submit
more terms of sequences like these, but now that it's a batch of
dozens of them, what's the best way to submit it?  Get the internal
format of each, copy that, add my more terms to the first line, and
email that to NJAS?

Here's how I'd phrase my personal criteria for what should be submitted:
1) if it's published, or it's the answer to a published problem or
puzzle (including from textbooks, contests, the web, or whatever), I
submit it.  This includes times when a puzzle maybe asks for a(2) or
a(127) of some sequence -- that is, I often look for generalizations
of published problems as well, and maybe get a triangle or a sequence
when the original puzzle asked for just one term.  I'll still submit
the whole sequence and reference the original problem.

2) if I worked for a while to figure out a problem, I submit the
answer to the problem whenever it seems naturally to be a sequence.  I
figure someone else might work on the same problem someday and I can
save them some work.  This is especially true if the problem seemed
hard at first but in the end had an easy solution.  A good example of
this is (some parts of) my whole bunch of sequences related to how
many distinct lines through the origin there are in the [0,k]^n cube
of lattice points, where I still don't know good answers to some parts
of the problem, but I found much simpler formulas than I had at first
expected for other parts.

3) my favorite things to add to the OEIS of all: comments on existing
sequences, clarifying them, showing other problems that have the same
answer, things like that.

4) if a student solves a problem that leads to a sequence, I'll
publish it in OEIS (referencing the math club), just because I think
students deserve all the encouragement they can get.  (I teach middle
school and high school math for a living).  If they find something
that's not already in OEIS, I think that's worth publishing!

That pretty much covers it for me.

Of course, other people will have different criteria, and in general I
think that's fine; diversity is a good thing.  But this is NJAS's
project, not a wiki -- he should publish what he likes.  To me,
sequences like (for example)
a(n) = smallest prime formed by taking the digits of a(n-1) and
appending a prime
are not very interesting at all.  I think that's why NJAS is
particularly discouraging sequences with keyword "base", and I agree
-- there are way too many arbitrary things that can be done there.  On
the other hand, some of the other web sites out there already have
lots of stupid prime number facts, and those facts lead to stupid
sequences in the OEIS.  I myself am a culprit on that one -- for
instance, I saw published on a web page that 12 is one of the numbers
with the property that its prime factors, raised to its power, sum to
a prime: 2^12 + 2^12 + 3^12 is prime.  So I figured I'd write a
program to find a few more terms ... but really it's not interesting,
is it.  Oh well.

--Joshua Zucker