18 comments

[ 2.4 ms ] story [ 32.5 ms ] thread
OEIS is such a wonderful reference. I've had occasions where software I was building needed to compute certain sequences, but I hadn't yet figured out the underlying math. I popped the sequence into OEIS and found the closed form solution. It was a huge productivity boost.
For me it was a favorite place to visit every so often. I also really enjoyed mathworld.wolfram.com a few decades ago. (A true shame that he went insane)
Coding exercise: write a function

    boolean isInSequence(n):
that decides whether the given integer is part of that sequence or not. However, pre-storing the sequence and only performing a lookup is not allowed.
Compute the sequence until you get n or m > n?
How about the following Haskell program?

    rec ((x:xs),p) = (filter (/= p+x) xs,p+x)
    sequ = map snd $ iterate rec ([2..],1)
sequ is an infinite list of terms of the sequence A005228.
I don’t know but I think I could probably implement IsInSequenceOrFirstDifferences(n)
Recursive (n choose 2) is my favorite.

https://oeis.org/A086714

If you think about it, it quantifies emergence of harmonic interference in the superposition of 4 distinct waveforms. If those waveforms happen to have irrational wavelengths (wrt. each other), their combination will never be in the same state twice.

This obviously has implications for pseudorandomness, etc.

Is there a sequence where the sequence and all its differences contain each positive integer once?

Something like

    1 3 9   26  66
     2 6  17  40
      4 11  23
       7  12
        5
Oh, here it is: https://oeis.org/A035313
(comment deleted)
> Oh, here it is: https://oeis.org/A035313

That sequence is not known to match what you asked for:

>> Conjecturally, every positive integer occurs in the sequence or one of its n-th differences, which would imply that the sequence and its n-th differences partition the positive integers.

For an intuition of why this might be hard to prove, note that you had to insert 7 into your structure before you inserted 5. In the general case, there might be a long waiting period before you're able to place some particular integer n. It might be infinitely long.

Can someone please explain this to me? I tried to make sense but couldn’t.
The sequence union the differences span all integer values.
The initial sequence is 1, 3, 7, 12, 18, 26, 35, etc. The difference between each term in that sequence produces a second sequence: 2, 4, 5, 6, 8, 9, 10, etc. If you merge those two sequences together in sorted order, you get 1, 2, 3, 4, 5, 6, 7, etc. Each whole number appears in the result exactly once.
I wonder why the title of the sequence isn't set to "Hofstadter's sequence" since that seems to be what it's called according to A030124 when it refers back to this one
Hofstadter introduces several sequences in GEB, [0] may be an interesting submission on its own but I was especially captivated by this self-referencing one. Plus a title including both Hofstadter's sequence and a description is too long for HN and I preferred the descriptive one

[0] https://en.wikipedia.org/wiki/Hofstadter_sequence

I meant the title as it appears in OEIS, not as it appears on HN :)