Hi Jason and Luke, On Mon, Feb 25, 2013 at 11:00 AM, Jason Moore <moorepa...@gmail.com> wrote: > Aaron, > > Just to reiterate Luke some: We've been working on a vision of a domain > specific tool chain based primarily in the scientific Python ecosystem that > will allow us to do multibody dynamics and control work. One of the most > important pieces is the core that we've developed in SymPy with you all's > help (the mechanics package). We plan to continue to develop that and work > on core pieces of SymPy to improve package and we hope that the work that > branches into the core of SymPy (code gen, cse, plotting, etc) will be > beneficial to all of SymPy. But we will soon be developing some parts of > this tool chain that rely more on other software or that don't fit into > SymPy's mission (we'll pitch these as the time comes to see where the best > home is for them). > > So, we're hoping to help potential GSoC applicants that are interested in > our vision submit applications to both SymPy and other relevant projects > this year. We may also apply to be a mentoring organization ourselves so > that we can better facilitate that, but I'm not sure of the likelihood that > we will be accepted. Never-the-less SymPy will have our help in development > work and mentorship roles in the far foreseeable future. We would surely not > be in the position to broaden our vision if SymPy hadn't taken us under > their wing. > > Let us know if you have any thoughts on this plan.
This is a great plan. Speaking for myself, I am *very* happy to see this development in pydy. When Luke and I did our first plans in 2009 as I described in my blogpost then: http://ondrejcertik.blogspot.com/2009/03/newtonian-mechanics-with-sympy.html I was hoping that exactly this would happen. Since then Luke has done a lot of hard work and then you joined and other people at UC Davis, which is really great. Pydy is an awesome application of sympy and I really believe in sympy as a symbolic library with batteries included. So that all of us can make sure that things work out of the box, in a good default way. As we discussed before on this list I think, it used to be that mostly Luke were the only one contributing to the pydy codebase (as time permitted), usually over the summer. Then it made great sense to keep it in sympy, as it got "free" maintenance with sympy itself. Maintenance is very important, that involves making sure all tests pass as sympy gets updated, that it works in all supported Pythons (especially Python 3) and so on. Now when there is a team around pydy, e.g. Luke, Jason, Stefen, Gilbert, who contribute all year long and then more people just over the summer (I got this information by looking at people+dates in "git log sympy/physics/mechanics/"), it might make sense to develop pydy separately if you guys want, but even better --- as you said Jason --- would be to keep the symbolic parts in sympy, and more pydy related things outside. The exact line what should go into sympy and what should be outside will naturally establish itself as time goes on I believe. There is a high value in providing a good end user experience for the whole stack, which beings with sympy, then pydy, probably ipython notebook, code generation, numerical solvers etc. I would be very happy if you keep developing some parts in sympy as then sympy can benefit from your work, and you can benefit from sympy as well. Ondrej -- 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 sympy+unsubscr...@googlegroups.com. To post to this group, send email to sympy@googlegroups.com. Visit this group at http://groups.google.com/group/sympy?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
