1
= 1
{scope}
All <logarithms> have base two. In part a the <information entropy> differences and information expressions are interpreted where they are well defined; finite <joint entropy> is a sufficient hypothesis. This avoids undefined expressions such as infinity minus infinity.