Yes definitely. Removing the old assumptions is just half of the project. I would also love to see things like the ideas discussed at https://github.com/sympy/sympy/pull/1907, and of course in addition to building nice frameworks for new assumptions, actually implementing useful assumptions and inferences on them would be nice too. You could also look at extending the logic module, so that they remain fast. There are also inference features that need implementing, like https://code.google.com/p/sympy/issues/detail?id=3496. You may also want to think about issues with solve, especially relating to https://code.google.com/p/sympy/issues/detail?id=3699 (which is also inherently linked to bigger issues with solve: https://code.google.com/p/sympy/issues/detail?id=3560). If we want to be able to do things like ask(Q.real(log(x)), Q.positive(x)), we have to have more rigor in solve().
Another direction the project could take is to do some refactoring of the core once the assumptions are completely decoupled from it. I'd love to see some of https://code.google.com/p/sympy/issues/detail?id=1941 implemented. But of course, that's not actually related to assumptions, just dependent on it, so perhaps focusing on the above ideas would be better. And while we're on the topic of this project, one thing that really needs to be done well for the new assumptions is documentation (https://code.google.com/p/sympy/issues/detail?id=2196). We need to decide, as a community, the exact definitions we want the core assumptions to have (like "positive", "real", and so on). Currently, we have contradictory facts (https://code.google.com/p/sympy/issues/detail?id=2877), and with the old assumptions, there have been dozens of cases throughout the years where people misused the assumptions in various parts of the code base because it was not clear what exactly they meant, especially with some of the more esoteric ones like finite or comparable (https://code.google.com/p/sympy/issues/detail?id=3684). Aaron Meurer On Mon, Apr 8, 2013 at 6:40 PM, Matthew Rocklin <[email protected]> wrote: > Also, regarding 2, I don't think it's just removing old assumptions. There > is a lot of potentially very constructive work to be done in new > assumptions. It has a decently simple/extensible design and can be taken in > a few interesting directions. > > On Apr 8, 2013 3:38 PM, "Matthew Rocklin" <[email protected]> wrote: >> >> >>> Well this is great to hear. >> >> +1 >> >> I would absolutely love to see (2) done. I think it's very important, >> needs both the dedication of a GSoC student and the experience of a veteran. >> It's a rare combination. I also think it's both a fascinating topic >> intellectually and a real engineering challenge. >> >> That being said I'm confident that your other proposals will also be of >> high quality. I suspect that your preference will be the deciding factor. > > -- > You received this message because you are subscribed to the Google Groups > "sympy" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at http://groups.google.com/group/sympy?hl=en-US. > For more options, visit https://groups.google.com/groups/opt_out. > > -- You received this message because you are subscribed to the Google Groups "sympy" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/sympy?hl=en-US. For more options, visit https://groups.google.com/groups/opt_out.
