>>9837

Right, I posted some about a variation of it in this thread >>9585.

I wrote a quick function to enumerate the first row, computing the gcd with c and then a loop to iterate over multiples of c (primes from column 1). You clearly see that the factors exists in different columns of e indexes by polite numbers, but that was when I realized I had already seen this pattern in >>8973 and I was just looking at a part of that.

Once you add more factors to c the sub-factors occurs several more times, in different orders.

Another though I had about the pattern in >>8973 was that, since the squares are rotated on the diagonal maybe we can shift the e's (to straighten out the squares). Like, move the "e-line" to the diagonal. Then the squares would become straight.

>>9850

> Right, I posted some about a variation of it in this thread >>9585.

I suppose it would make sense that there exists other variations as well.

A thing I though of a few days ago was the idea of "virtual columns". It's not really a virtual column, but the fact that every single column can be created by enumerating other columns and combining them.

Take (3, 1). We have two equations we use which is either 2tt - (half of e) or 2t(t+1) - (half of e). But in the "virtual column" you create (3, 1) by merging other columns. I got the idea when I was looking at the change in the columns when following a specific c through the a[t]'s.