> > It seems that this has finally been pushed in (including, unfortunately, > a large mess of meaningless merges).
Large mess? There shouldn't be so many since it was cleaned up with the help of Vinzent and Aaron. I think there are still a few > things that should be addressed before the next release: > * The static compilation of known_facts_dict is a problem > I'm on it: - http://code.google.com/p/sympy/issues/detail?id=2016 <http://code.google.com/p/sympy/issues/detail?id=2016&can=3>- http://code.google.com/p/sympy/issues/detail?id=2017 Note that in the spirit of knowledge compilation, these facts are quite suited for static compilation. They don't change with new assumptions, but when someone wants to redefine the notion that things can be both odd and even, or want to introduce some entirely new predicate that is assumable. * The speed of its generation can and must be improved a lot (in > particular, to_int_repr is very inefficient) > Seeing as it should only be done once, I don't see this speed really being an issue (still a matter of seconds on my machine). I can look into a different form of inference rather than calling the SAT solver for every pair of facts, but I don't want to resort to simple sufficient conditions (we want every knowable fact to be compiled offline so the online inference doesn't have to recompute it). * There's no point in keeping the old DPLL implementation. > S'pose so. Multiple DPLL algorithms are possible now, however. The old implementation serves as a simpler example of how to interface a new SAT solver into the mix. > Also, where did the _subs_atoms methods come from?? > ... -- You received this message because you are subscribed to the Google Groups "sympy-patches" group. To post to this group, send email to sympy-patc...@googlegroups.com. To unsubscribe from this group, send email to sympy-patches+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/sympy-patches?hl=en.
