GSOC Application
Tom Bachmann
I would like to express my interest in working no the symbolic
integration capabilities of sympy as part of a GSOC project.
My name is Tom Bachmann and I study mathematics (second year) at the
university of cambridge, england. Here is an overview of my computer
programming experience: I have previously worked on the Hurd project
(in C), I did a project that started the port of the kaXen/afterburner
pre-virtualisation environment to amd64 (in C++), and I have extended
the wikireader codebase to handle ebooks from project gutenberg
(mostly python). I can supply more details and/or references if you
wish. I have also created some "fun projects" on my own, the most
relevant here being probably what I call "fz" [1], a program to plot
various special functions in the complex plane and on riemann surfaces
(in C++). Finally here in cambridge there are so-called "CATAM" [2]
(computer-aided teaching of all of mathematics) projects on which I
got excellent results; this may or may not be meaningful to you.
With this background settled, let me say that I find both of the
proposed approaches to symbolic integration (resdiue theorem and Mejer
functions) very interesting. I believe I do understand well the
mathematics behind both. Depending on what you perceive to be more
important, I would be happy to work on either, with possibly a slight
preference for the residue method. What is the state of any existing
implementation in sympy?
If there is any specific other information that you want me to supply,
please don't hesitate to let me know.
Thanks,
Tom
[1] https://bitbucket.org/ness/fz/overview
[2] http://www.maths.cam.ac.uk/undergrad/catam/
Aaron S. Meurer
Currently, SymPy is getting pretty strong with indefinite integration, thanks to my project to implement the Risch Algorithm. However, there is really no progress with definite integration. The integrator basically uses the fundamental theorem of calculus (integrate and evaluate at the end points) to evaluate definite integrals, and that is it. So there is definitely much room for improvement in this area.
Now, the Meijer G function project would be much more general (i.e., powerful) than the residue one, but I think it would also be more difficult. If you are interested in that one, you might look at the paper by Kelly Roach [1].
The residue project would probably be easier, but again, much less powerful.
By the way, the integration algorithms are my particular interest in SymPy, so I would like to discuss this more with you. If you want, we can talk on IRC. Our channel is #sympy on Freenode.
Finally, I want to remind you that we require all student applicants to submit at least one patch to the project that gets reviewed and pushed in. See https://github.com/sympy/sympy/wiki/development-workflow for a guide on how to submit a patch. Some easy to fix issues that can get you started are labeled EasyToFix at our issue tracker. See http://code.google.com/p/sympy/issues/list?q=label:EasyToFix.
Aaron Meurer
[1] - K. Roach. Meijer g function representations. In ISSAC ’97: Proceedings of the 1997 international symposium on Symbolic and algebraic computation, pages 205–211, New York, NY, USA, 1997. ACM.
Tom Bachmann
suggested: https://github.com/sympy/sympy/wiki/GSoC%202011%20Tom%20Bachmann:%20Definite%20Integration
> By the way, the integration algorithms are my particular interest in SymPy, so I would like to discuss this more with you. If you want, we can talk on IRC. Our channel is #sympy on Freenode.
>
thingy works properly (I can't seem to get my chat client running at
my current [holiday] location).
Tom Bachmann
I'm just bumping this up to let potential mentors and/or application
reviewers (except for Aaron and Ronan, thanks a lot again for your
input!) know that my application is at [1]. If there is anything
unclear to you, or if there is anything you would like me to add or
expand on, please don't hesitate to let me know.
Thanks,
Tom
[1] http://www.google-melange.com/gsoc/proposal/review/google/gsoc2011/ness/1