-----Rupert wrote: -----
>There might be another way of getting this result using the format
>library, but that's somewhat unmaintained if I remember correctly.
Using format, the best I could do was
(%i124) load(format)$
(%i128) format(e, %poly(A));
(%o128) (1/E+1/D)/A^2+1/(A*B)+1/AC
If you are picky about the appearance of expressions in publications, there is no way around doing
some hand tweaking, I think. And Rupert is correct: the format package doesn't have a regression test,
its user documentation isn't a part of the main documentation (look for it in share/contrib/format), and
it's somewhat unmaintained (or maybe format is bug-free ;)
--Barton