I wrote just a minute ago,

> Similarly, a negative zero (max negative number) makes a good NULL for I, J,
> K, and R datatypes (with a max negative being especially good for J
> datatypes because it too is an illegal bit pattern).

I need to correct an error before Stan or Bruce tells me I screwed up. For the
K datatype, the best NULL value wouldn't be a 1000000.... type representation.
Indeed, that value would be right in the middle of its range. The best and
most obvious NULL representation for K datatypes would have to be
111111111.... (the max positive number).

Wirt Atmar