It's an interesting problem, isn't it? For the record, on my Axiom
3*1.3^2 produces 5.07
but
3*1.3^3 produces 6.5909999999 999999999
Hmmm....Maybe what we need in Maxima is a user switch so that all finite
decimals (such as 1.4) are internally converted to rationals, and all
subsequent arithmetic done exactly.
It seems to me that there is a dichotomy between the exact results produced
by arbitrary precision integer arithmetic, and approximate results produced
by floating point arithmetic. Perhaps all users of systems such as Maxima
need to get used to - and understand - the rounding effects of floating
point arithmetic?
For the record, Maple produces finite decimals (or appropriate zeros) for
all of 1.96-1.4^2, 3*1.4^2, 3*1,3^2, 3*1.3^3, etc.
-Alasdair