Late last month, I checked in a new version of topoly and
to_poly_solver. Additionally, I contributed user documentation
(share/contrib/topoly-user-doc.texi and
share/contrib/topoly-user-doc.html).
This morning, I appended the to_poly_solver test suite to the share
testsuite. If you all have a chance, run the share_testsuite and let
me know if rtest_to_poly_solve reports problems.
To read a few of my thoughts about topoly and the to_poly_solver, do a
web search on "nary a variety" (with the quotes), or
http://www.unk.edu/uploadedFiles/facstaff/profiles/willisb/solve-talk
(3).pdf
A quick tour:
(%i16) to_poly_solve(sin(x) = 1/2,x);
(%o16) %union([x = 2*%pi*%z2153+%pi/6],[x = 2*%pi*%z2155+5*%pi/6])
(%i17) featurep(%z2153, integer);
(%o17) true
(%i18) to_poly_solve(max(x,4) = x,x);
(%o18) %union(%if(isnonnegative_p(%c2159),[x = %c2159+4],%union()))
/* apply rootscontract to each solution */
(%i19) to_poly_solve(sqrt(x) - x = sqrt(42) - 42,x, 'simpfuncs =
['rootscontract]);
(%o19) %union([x = (-2*sqrt(42)+sqrt(169-4*sqrt(42))+85)/2])
(%i20) float(%);
(%o20) %union([x = 42.0])
Barton