Sequential Squares -- A005842

[Ed Pegg Jr]

One comment on the new wording.

If a solution exists for ceiling(sqrt(sum(k^2, {k,1,n}))), then that's
a proof.  All entries labeled "solvable" have been proved by
demonstration.

It should be straightforward to prove everything under n=26 is
correct.

n=23 is definitely 66, not 67.  Proof: http://www.mathpuzzle.com/seq23.gif

n=26 and n=28 will be difficult to prove, and likely need long brute-force
computer searches.

n=32, 33, and 34 can likely be proved in some elegant manner, OR they
will have fantastically elegant solutions. 

n=36 has been solved (I think by Friedman). 

I'll make a page about all of this, and will include all of the known 
solutions.

--Ed Pegg Jr.