Just as a note, the R user community is several orders of magnitude
larger than that of Maxima (or Sage, for that matter), and their help
list (about a hundred messages a day, give or take) constantly has
people asking for symbolic manipulations (and other things usually
performed more traditionally by Maple/Mma/Maxima/Sage/even
Matlab/Octave..). So this is not a stupid question at all.
There is a R port of both sympy and Yacas, and it is possible (though
not always syntactically easy) to mix R and Sage/Maxima/Ginac symbolic
and other things in Sage. I'm sure there are other packages which do
various symbolics, including some calculus stuff in the new
(pedagogically-oriented) mosaic package, the NSF grant of which should
be of independent interest to Maxima folks (see
http://mosaic-web.org/).
So I don't see why it wouldn't be effective to make "hooks" in Maxima
(or R) to use the other one directly, and there should be interest.
Rewriting R, on the other hand, makes no sense; the sheer number of
*quality* user-contributed packages to R is one of its strongest
suits, one that could never be replicated. (In fact, it's why so many
people are ditching the proprietary guys for R. There are at least
FOUR companies which intend to eventually make a profit supporting or
adding on to R in some fashion.)
Karl-Dieter
PS to rjf: and no, it doesn't really make sense to rewrite Maxima in
Python or Cython (nor does Sage do so). Just forestalling that
inevitable comment :)