> So, theoretically speaking, the square doesn't go
 > negative, we just don't get all the bits on a computer
 > doing int maths. When the (truncated) result of an int
 > multiplication is later viewed as an unsigned int, the
 > result may *appear* to be negative, precisely when the
 > highest bit happened to be set to 1 by the
 > multiplication.

    Nice that you agree. :-)

Best Wishes,
Bill.

---