EDITED A104732

[Maximilian Hasler]

This was "uned", with definition "Triangle, row sums are A001924" not
clearly indicating how values are to be computed.
The %F provides a method, but I don't think it is possible to
summarize it in a sound definition not exceeding 2 lines.
Interpreting the data rather as square array seems more natural and
allows a simple, constructive definition.
Neil, please confirm me again if the line breaks (as received in
"internal format") need to be removed.

%I A104732
%S A104732 1,2,1,3,3,1,4,5,4,1,5,7,8,5,1,6,9,12,12,6,1,7,11,16,20,17,7,1,8,13,20,
%T A104732 28,32,23,8,1
%N A104732 Square array T[i,j]=T[i-1,j]+T[i-1,j-1], T[1,j]=j,
T[i,1]=1; read by antidiagonals.
%C A104732 Original definition was "Triangle, row sums are A001924".
Reading the rows of the triangle as antidiagonals of a square array
allows a precise, yet simple definition and method to compute the
terms. - M. F. Hasler, Apr 26 2008
%C A104732 When formatted as a triangle, row sums are A001924: 1, 3,
7, 14, 26...(apply the partial sum operator twice to the Fibonacci
sequence).
%F A104732 The triangle is extracted from A * B; A = [1; 2, 1; 3, 2,
1;...] B = [1; 0, 1; 0, 1, 1; 0, 0, 2, 1;...]; both infinite lower
triangular matrices with the rest of the terms zeros. The sequence in
"B" (1, 0, 1, 0, 1, 1, 0, 0, 2, 1...) = A026729.
%e A104732 The first few rows of the triangle (= raising diagonals of
the square array) are:
%e A104732 1;
%e A104732 2, 1;
%e A104732 3, 3, 1;
%e A104732 4, 5, 4, 1;
%e A104732 5, 7, 8, 5, 1;
%e A104732 6, 9, 12, 12, 6, 1;
%e A104732 ...
%Y A104732 Cf. A001924, A026729.
%Y A104732 Adjacent sequences: A104729 A104730 A104731 this_sequence
A104733 A104734 A104735
%Y A104732 Sequence in context: A115131 A117895 A128139 this_sequence
A132108 A125175 A073020
%K A104732 nonn,tabl
%O A104732 1,2
%A A104732 Gary W. Adamson (qntmpkt(AT)yahoo.com), Mar 20 2005
%E A104732 Edited by M. F. Hasler (www.univ-ag.fr/~mhasler), Apr 26 2008