hello everyone,
if the goal is to carry out a numerical integration, then simpson and
other quadrature functions of the newton-cotes variety (of which there
are several implementations in maxima) are all superseded by the
gauss-kronrod rules implemented by the quadpack functions (quad_qags, etc).
? quad_ shows info about all the quadpack functions.
> I realize this is inconsistent with the conventions in other cases
> (e.g. plot2d). At some point, we really ought to try to decide on a
> convention and apply it globally.
well, i'm not so sure we need to make a choice here.
plot2d / plot3d and quad_* all construct their plottand (ha ha) or
integrand via COERCE-FLOAT-FUN, which can (in 5.9.3)
handle a function name, a general expression, or a lambda.
other functions in need of a similar construct should also
go through COERCE-FLOAT-FUN; no need to reinvent that
particular wheel. i don't know if there are any such functions,
but i haven't looked either.
all the best,
robert dodier