Ahh yes - the product and sum of quadratic roots that is expected to be used for trivial factorisation problems.
OP - this will come in useful if you are looking for roots of the form a + bi , or k + y (where a + bi denotes an imaginary number) and k and y are two numbers that are not known.
However when it is clear (test some numbers - if you don't spot the difference of two squares, whack in say, x = 1, x = 2, xx= 3 - if you get zero, you have a root). You know that if (foor example 3 is a root), then you have a bracket that yields x = 3. This will only occur if the equation equals zero, and if you have (x - 3)(ax + b) = 0. [for this problem, of course with higher powers you could have (2x - 6)(ax^n-1 + bx^n-2 ... + z) = 0).