Interested in doing a GSoC project - where to get started?

[Esther Robb]

Hello,

My name is Esther, I'm a Computer Science student at Northern Virginia Community College in the US. I want to get involved in developing mathematical software, although I am new to both programming and open source software. I have experience with C++ and Java, I have taken courses in Linear Algebra and Calculus 1 & 2, and I'm currently taking Vector Calculus. I looked over the project page on your website and a few stood out to me, but I wanted to get some guidance on what projects are most needed, what would be appropriate for my skill level, or if there are any small projects or bug fixes that I could start learning on. I'm very interested in mathematical applications of programming and want to do something challenging, but I do have some experience limitations.

I have been programming for less than a year, so I'm new but I want to learn. I have written a program in C++ that converts between color bands on a resistor and the corresponding tolerances and resistances, and have used C++ for some other engineering/science related applications. In Java, I have written some basic math programs, such as approximating e to a power using Taylor Polynomials, encoding and decoding an integer, and tracking the (x,y) coordinates of a point as it moves in a circle over time using parametric equations. I am currently working on a small 2-D Java game using the LibGDX library, and have also written a version of the game Snake which has an AI mode. I am a little more familiar with Java than C++, but am comfortable with both. I have also used Python but only minimally.

The projects that especially interested me were the BLAS wrapper implementation for FLINT, the Gröbner basis project, and the Fast Linear Algebra over Extension Fields project. I'd like to know if any of these are appropriate to my skill level, or if there is any research or smaller project I should start working on! I would love to get involved and gain some awesome math and programming skills.

[Burcin Erocal]

Hi Esther,

thank you for your interest.

As I wrote before, we are still preparing our application as a
mentoring organization for the GSoC. The ideas list is going to change
a little in this process. For example, it still includes some ideas that
were taken up last year. We will also need to take down some ideas
because some people (like me) won't be able to mentor projects this
year.

On Fri, 6 Feb 2015 13:47:31 -0800 (PST)
Esther Robb <estherr...@gmail.com> wrote:

<snip>

> The projects that especially interested me were the BLAS wrapper
> implementation for FLINT, the Gröbner basis project, and the Fast
> Linear Algebra over Extension Fields project. I'd like to know if any
> of these are appropriate to my skill level, or if there is any
> research or smaller project I should start working on! I would love
> to get involved and gain some awesome math and programming skills.

I suggest you start looking at FLINT: Get it working on your system,
make small changes in it and see how it works, try to fix something in
their TODO list, etc. The other projects you mentioned are harder to
get started with.


Cheers,
Burcin

[Esther Robb]

Thank you Burcin, I will get started with FLINT. Is the BLAS project simply less research-intensive, or is it also likely to be easier to implement?

[Burcin Erocal]

I didn't mean to imply that you should concentrate on the BLAS project.
FLINT will have several other project ideas. It is easier to get
started in FLINT as opposed to the other options you mentioned, since
they have a clean code base with a simple build system.

The Groebner bases project, which I'm not sure we will offer this year,
requires looking into a lot of different code bases and trying to
combine their best parts. Linear algebra over small non-prime finite
fields would likely require either designing something from scratch
or figuring out how LinBox is designed.


At this point, instead of concentrating on a specific GSoC idea we
have on the list (which is due to change anyway), it's better to see if
you can handle the code bases of various software projects lmonade is
an umbrella organization for. Then you can pick a specific idea based
on which one you are most comfortable with.


Cheers,
Burcin

[Esther Robb]

Okay makes sense. I'll do that, and thanks again.

Esther