[reply at bottom -- rich] Quoting Bill Gosper <billgosper@gmail.com>:

Quoting Bill Gosper <billgosper@gmail.com>:

<clip>

On Tue, Dec 17, 2013 at 6:20 PM, Bill Gosper <billgosper@gmail.com> wrote:

Neil[B] and I recently noticed

(2^(2*k + 1) - 1)^2 + 8 ==

(2^(2*k + 1) + 2^(k + 2) + 3) * (2^(2*k + 1) - 2^(k + 2) + 3) Old? --rwg

This is pretty dull, since Factor[...] can simply do it. Slightly more

interesting is In[182]:= Factor[(2^(k + 1) - 3)^2 - 8]

Out[182]= 1 - 3 2^(2 + k) + 2^(2 + 2 k) (Fails) but both the "odd" and "even" subcases work:

In[185]:= Factor[(2^(2*k + 1) - 3)^2 - 8]

Out[185]= (1 - 2^(2 + k) + 2^(1 + 2 k)) (1 + 2^(2 + k) + 2^(1 + 2 k))

In[186]:= Factor[(2^(2*k) - 3)^2 - 8]

Out[186]= (-1 + 2^(2 k) - 2^(1 + k)) (-1 + 2^(2 k) + 2^(1 + k)) --rwg

In the even case, we can generalize 2^_ to x^_:

In[208]:= Factor[(x^(2*k) - 3)^2 - 8]

Out[208]= (-1 - 2 x^k + x^(2 k)) (-1 + 2 x^k + x^(2 k))

In[210]:= Expand[%]

Out[210]= 1 - 6 x^(2 k) + x^(4 k)

But

In[209]:= Factor[(x^(2*k + 1) - 3)^2 - 8]

Out[209]= 1 - 6 x^(1 + 2 k) + x^(2 + 4 k) In fact,

In[210]:= PrimeQ[% /. x -> 6 /. k -> 16]

Out[210]= True --rwg

rcs> Out209 can be morphed to Y^4 - 6 Y^2 + 1 = (Y^2 - 1)^2 - 4 Y^2 = (Y^2 - 2 Y - 1) * (Y^2 + 2 Y - 1).

This is similar to the canonical Aurif factorizations for 2^(4K+2)+1 and 3^(6K+3)+1, which can be morphed to factorizations of 4 Y^4 + 1 = (2Y^2+1)^2 - 4Y^2 and 27 Y^6 + 1 = (3Y^2+1) (9Y^4 - 3Y^2 + 1), with the final factor splitting as (3Y^2+1)^2 - 9Y^2. I like to think of these as [1 -6 1] = [1 -2 1] - [4 0] and [4 0 1] = [4 4 1] - [4 0] and [9 -3 1] = [9 6 1] - [9 0], followed by making the subtrahends into squares.

Rich ------

Rich, then how would you explain the primality of

(6^33-3)^2-8? Your "morphing" requires

Factor[_,Extension->√2] .

--rwg

I overlooked your interesting method of indexing. My remarks apply not to Out209, but to the earlier couplets In208,Out208 and the first In210,Out210 pair. Rich