equal problem



> ---------------------------------
> I can't find any documentation for signbfloat.

There isn't any. The fact that signbfloat is undocumented is mentioned in 
SF bug 798571.
The signbfloat mechanism is broken: (from the bug report)

(%i30) p : exp(x) - sum(x^k/k!,k,0,15)$

(%i31) p : subst(1/10000001,x,p)$

(%i37) for k : 15 thru 25 do (fpprec : k, print(k,sign(p)));
15 zero
16 pos
17 zero
18 zero
19 neg
20 pos
21 neg
22 neg
23 zero
24 pos
25 zero

Barton