Giovanni,
'numer' is applied when an expression is simplified. Once an
expression is simplified, Maxima won't normally resimplify it, even if
some of the original assumptions it made have changed.
You can, however, resimplify with the current environment. Here are
some examples to show how this work, including some of the
idiosyncracies.
(%i1) display2d:false$ /* show output
on one line */
(%i2) exprlist: [ abs(x), sin(%pi/23), x*a ]$
(%i3) resimplify(expr):= expand(expr,1,1)$ /* define resimplify */
(%i4) resimplify(exprlist);
(%o4) [abs(x),sin(%pi/23),a*x]
/* no changes to environment, so no change when resimplified */
(%i5) assume(x<0)$
(%i6) resimplify(exprlist);
(%o6) [-x,sin(%pi/23),a*x]
/* takes advantage of assumption */
(%i7) numer:true$
(%i8) resimplify(exprlist);
(%o8) [-x,sin(0.043478260869565*%pi),a*x]
/* setting numer makes 1/23 into a float, but not %pi */
(%i9) float:true$
(%i10) resimplify(exprlist);
(%o10) [-x,0.13616664909625,a*x]
/* float flag controls %pi -> float */
(%i11) assume(equal(a,0))$
(%i12) resimplify(exprlist);
(%o12) [-x,sin(%pi/23),a*x]
/* assuming equality does *not* feed into simplification here */
(%i13) a: 23$
(%i14) resimplify(exprlist);
(%o14) [-x,sin(%pi/23),a*x]
/* resimplifying does not reevaluate */
(%i15) ev(exprlist); /* reevaluate */
(%o18) [-x,sin(%pi/23),3*x]
Hope these examples are helpful and not too confusing....
-s
> You example is a almost what i need, but don't work if the variable
> is pre declared:
> (%i1) numer;
> (%o1) false
> (%i2) x:tan((2*%pi)/9);
> 2 %pi
> (%o2) tan(-----)
> 9
> (%i3) numer:true;
> (%o3) true
> (%i4) x;
> 2 %pi
> (%o4) tan(-----)
> 9
...
> It's possible to have a variable recalculated (or just displayed) in
> a numerical form?