>>>>> "Stavros" == Stavros Macrakis <stavros.macrakis@verizon.net> writes:
[nice test cases snipped]
Stavros> But let's not throw out the baby with the bathwater. The
Stavros> purpose of all the complexity around %e is to prevent
Stavros> strictly symbolic expressions from becoming gratuitously
Stavros> numerical; who wants to see exp(%i*x) => 2.718^(%i*x)?
Stavros> This currently never happens; neither does x^3 => x^3.0
Stavros> or 1/x => x^-1.0. On the other hand, sin(%pi*x) does
Stavros> gratuitously become sin(3.142*x), except in float().
I don't actually mind sin(%pi*x) becoming sin(3.142*x) when I asked
for numeric results. I did, after all, ask for numeric results and
maxima did the best it could. I find it much more annoying when I ask
for numeric results and don't get them when it should have.
I do agree that the meanings of float, numer, and bfloat and their
interactions is quite a mess.
Ray