[VA] SRC #012b - Then and Now: Root

[Werner]

(second try..)
Well I do have a result now, my previous one was wrong.
It's on Free42, which appears to be permitted (I understood that wrongly), and while I can make it run on a 42S, it would take too long.
Also, I have no idea if my result is the smallest. I'll have to do some thinking instead of mindless coding ;-)
Werner

- the Free42 solution precalculates the matrix of 10001 coefficients - an easy adaptation of a prime number listing routine I wrote years ago. On a 42S, this is not feasible, so the coefficients would have to be calculated over and over - hence the probably long execution time there. (it takes about 4 hours to generate the coefficients once)
- then, a routine to evaluate a polynomial stored as a row vector, also from my archives (really simple)
- and at last, a 'specific' solver routine, which, I think, was inspired by a post of Valentin himself, long ago. 7 iterations are needed to converge to a result, so, no, I'm not going to run it on my 42S.
When I'm reasonably sure the result is correct I'll post some code.