[seqfan] Recaman's transform of its own digits

[Eric Angelini]

A "Recaman's transform" I like, of course, is this one,
where the succession of the sequence digits is the succession
of the absolute first differences:

1 2 4 8 16 15 9 8 3 12 4 1 0 2 6 5 5 3 9 4 9 6 15 11 2 8 7 2 1 0 2 10 3 1 0 0 2 1 1 4 3 3 ...
 1 2 4 8  1  6 1 5 9  8 3 1 2 4 1 0 2 6 5 5 3 9  4  9 6 1 5 1 1 2 8  7 2 1 0 2 1 0 3 1 0 ...

Accepting "0" in the sequence is not interesting at a too early stage:

1 0 0 0 0 0 ...
 1 0 0 0 0 0 ...

1 2 4 8 0 0 0 0 0 ...
 1 2 4 8 0 0 0 0 0 ...

If we don't want any 0 in the sequence, we might impose:

1) Monotonically increasing terms:
1 2 4 8 16 17 23 24 31
 1 2 4 8  1  6  1  7    --> but this gives https://oeis.org/A100787

2) Zeroes are simply forbidden in the sequence:
1 2 4 8 16 15 9 8 3 12 4 1 2 4 8 7 5 1 9 2 7 6 15 13 6 12 11 6 5 2 8 7 5 ...
 1 2 4 8  1  6 1 5 9  8 3 1 2 4 1 2 4 8 7 5 1 9  2  7 6  1  5 1 3 6 1 2 ...

Best,
É.