> Meanwhile, is this something that would break AT&T troff if
> groff were to do it right?
The question is, what is "right"? I believe you're looking for
a mapping of integers to booleans, and a "not" operator for
booleans. [ntg]roff already has both (in a way):
* positive numbers are "true", nonpositive numbers are "false"
(in contrast to several other languages which use "zero" and
"nonzero").
* if "n" is a number (integer) interpreted as a boolean, then
"not n" is given by "(1-n)".
Thus:
.nr NC 0
.nr SI 0
.if \n(NC&(1-\n(SI) NC only.
.if (1-\n(NC)&\n(SI SI only.
.if \n(NC&\n(SI NC and SI.
.if (1-\n(NC)&(1-\n(SI) neither.
(tested with OSF1 nroff, which only has 1- and 2-character names).
