[SciPy-Dev] Changes Incurred in the Project Idea

[Joshua Wilson]

Just to second what Nikolay said, there are are lots of interesting
numerical integration things to work on. One thing I've been wanting
to see is more PRs along the lines of

https://github.com/scipy/scipy/pull/4587

which added support for very high order Gauss-Hermite quadrature.
Currently all of the other Gaussian quadrature methods are done by
computing the eigenvalues of a tridiagonal matrix, which works fine
for low orders but becomes inaccurate/slow for very high orders.
There's been recent research on how to improve on that, see for
example

http://epubs.siam.org/doi/abs/10.1137/120889873?journalCode=sjoce3

I think this could make a very nice summer project.

- Josh

On Sat, Feb 24, 2018 at 3:06 AM, Nikolay Mayorov
<nikolay.mayorov at zoho.com> wrote:
> Hi Gagandeep!
>
> I read the PDF you attached. I want to point out to you several points:
>
> 1) You mentioned serval mathematical constructs: curve, surface, volume.
> These don't directly translated to Python, you need to define some
> representation/API for that. And obviously defining any of these is very
> tricky. If you want to continue to develop this idea you need to provide a
> detailed information about that: how do you define curve/surface/volume,
> will it be practical and general enough, will it be convenient for a user.
>
> 2) Next, scipy is a numerical library and we are interested for an
> established and effective numerical methods. Your descriptions doesn't say
> anything about which numerical methods exist for solving problems you
> described. Some problems formulated in a mathematical form are just not
> suitable for numerical methods.
>
> 3) This is rather minor, but don't use double underscores for your methods
> (they indicate special "magic" methods).
>
> Please don't consider this as a discouragement. If you are interested in
> scipy.integrate and numerical integration I'm sure there are enough things
> you can work on, we can discuss and come up with something.
>
> Best,
> Nikolay
>
>
> ---- On Sat, 24 Feb 2018 12:59:10 +0500 GAGANDEEP SINGH <gdp.1807 at gmail.com>
> wrote ----
>
> Dear developers,
>
> After working on feedbacks received from the developers I have made changes
> in my Project Idea for GSoC 2018 for SciPy.
> This time it is based on extending scipy.integrate. You can see the
> attachment for details.
>
> Waiting for feedbacks.
>
> With Regards,
> Gagandeep Singh
>
>
> _______________________________________________
> SciPy-Dev mailing list
> SciPy-Dev at python.org
> https://mail.python.org/mailman/listinfo/scipy-dev
>
>
>
>
> _______________________________________________
> SciPy-Dev mailing list
> SciPy-Dev at python.org
> https://mail.python.org/mailman/listinfo/scipy-dev
>