genfact, fix, and entier



To convert genfact to gamma form, Maxima uses entier
(floor), where Macsyma used ceiling:

(%i1) makegamma(genfact(a,b,c));
(%o1) (gamma(a/c+1)*c^entier(b))/gamma(a/c-entier(b)+1)

Question: Does anybody know why Macsyma changed floor to ceiling
in this conversion?

Question: Maxima has a few calls to 'fix' and 'entier' in the source. 
Would it be OK to 

  (defun $fix (x) (simplify `(($floor) ,x)) 

and similarly for $entier? I suppose fix could be a speedy alternative
to floor when it's known that the input is a number. I don't favor 
this idea.

Barton