float and double bounded facet value, new 1.0 error not caught by 2E

float and double are not bounded.  This is because NaN compares only to
itself.  E.g., 29 and NaN have no common upper or lower bound.  C.1 should
be corrected.

Given that this parallels the error in 2E that lost the only-partial-order
of float and double, perhaps both can be corrected in 2E.
-- 
Dave Peterson
SGMLWorks!

davep@iit.edu