On roundoff error distributions in floating point and logarithmic arithmetic

Abstract
No abstract available