**The first differences of T**

are the terms in odd
position in T

Hello SeqFans,

T = 1,2,1,3,2,5,1,6,3,9,4,11,5,13,3,12,6,17,5,18,7,21,6,22,8,25,10,...

Odd pos.
O = 1 . 1 . 2 . 1 . 3 .
4 . 5 . 3 . 6 . 5 . 7 . 6 . 8 . 10

First dif **D**
= 1 1 2 1 3 4 5 3 6 5 7
6 8 10 ... --> this forms O again

So, generally speaking, we have:

T = a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z, ...

O = a c e
g i k
m o q
s u w y ...

**D** = a c e g i k m o q s u w y ...

To start S we have to be sure that a+a=b; this happens with the “smallest available difference
a” which is 1. This leads to b=2 and we have the first two terms of S:

T = 1,2,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,
...

**D** =
1 c e g i k m o q s u w y ...

The next step consists in extending the **D** line. Should we decide to always take
“the smallest possible difference” would give c=1, third term of S:

T = 1,2,1,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,
...

**D** =
1 1 e g i k m o q s u w y ...

Should we decide again to take “the
smallest possible difference” would give e=1, thus d=2 and g=1:

T = 1,2,1,2,1,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,
...

**D** =
1 1 1 1
i k m o q s u w y ...

... etc. This
way of building S would produce the rather dull (but correct) sequence:

T =
1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1 ...

**D** = 1 1 1 1 1 1 1 1 1 1 1 1 1 ...

... where, indeed,
“the first differences of T are the terms in odd position in T itself”.

So, instead of choosing systematically 1
for the next difference on line D, we have decided here, __when we had the
choice__, to take always “the smallest possible difference __not yet in D
and not leading to a contradiction__”:

T = 1,2,1,3,2,5,1,6,3,9,4,11,5,13,3,12,6,17,5,18,7,21,6,22,8,25,10,...

First dif **D**
= 1 1 2 1 3 4 5 3 6 5 7
6 8 10 ... (those are the terms in odd position in T)

The terms where we had the choice are in cyan
color hereunder (note that when we had the choice we never did repeat an
existing term):

T = 1,2,1,3,2,5,1,6,3,9,4,11,5,13,3,12,6,17,5,18,7,21,6,22,8,25,10,...

First dif **D**
= 1 1 2 1 3 4 5 3 6 5 7 6 8 10 9 6 11 ...

So, this is the “lexicographically first
interesting sequence T, where the absolute first differences of T are the terms
in odd position in T itself”.

If this is of interest, could someone
please compute 100 terms or so – and submit T to the OEIS?

Best,

É.

__________

P.-S.

The “lexicographically first interesting
sequence S, where the absolute first differences of S are the terms in even
position in S itself” is here.
(You will remark that S is exactly the sequence T without its first term.)