The current convention is inconsistent across the different types.
I find it confusing that the inference

   fixnum? -->  fixnum+, fixnum<

is correct, but that the inference

   flonum? -->  flonum+, flonum-

is incorrect.  Furthermore, the inference

   exact+, inexact+, fixnum+ --> flonum+

is also incorrect, as is the inference

   fixnum-negative? --> fl-negative?

I don't think brevity is a good reason for inconsistencies that do not
cater as much to those who write fixnum-intensive code as it does to those who
write flintensive code?  In any case, the library proposal allows rebinding of
often-used names for those who really desire brevity.

Erratum: The examples to illustrate div_0 and mod_0 instead show div and mod.

Regards
Andre