collecting terms

...even on an expanded general form expression, t=
here are reasons to
use 'ratcoeff' instead of 'coeff'
 
(%i2) p : 6/((a+b)*x);
(%o2) 6/((b+a)*x)
 
 <=
/DIV>
(%i3) coeff(p,x,-1);
(%o3) 6/(b+a)
 
 
=

(%i4) coeff(expand(p),x,-1);
(%o4) 0
 
 
(%i5) ratcoeff(rat(p,a),x,-1);
(%o5) 6/(b+a)

So don't ta=
ke my suggestion to use 'coeff' instead of 'ratcoeff' too serious=
ly; other
considerations are likely more important.
 
Barton
=