GSoC 2022 Introduction

[Jyothsna M S]

Hey Everyone! 

I am Jyothsna M S. I am new to open source and very much eager to contribute to open source. I am capable of coding in Python and C. I am looking forward to contributing to this organization through GSoC 2022. It would be of great help if I could get some assistance as to where to start contributing. I'd also like to know if there is any discord or slack server I could join.

[david....@gmail.com]

Thank you for your interest in GSoc2022.

Before applying for GSoC with Sagemath, you should start getting familiar with Sagemath (how to play with graphs, do basic algorithms, etc.) and with the developer guide http://doc.sagemath.org/html/en/developer/index.html (what’s the programing standard, how to use git, etc.). Then, you can start contributing Sagemath with smalls patchs (implementation of a new algorithm, improvement of an existing algorithm, improvement of the documentation).

You can see the current activity at https://trac.sagemath.org/

Sincerely,

David.

[Pedro Orlando]

Hello everyone,

I'm Pedro Orlando and I am currently doing a double degree between the State University of Campinas (Unicamp), Brazil and Télécom Paris, France. At Unicamp I was in the last year of my BSc in Applied Mathematics, and now I'm specialising in Applied Algebra and Theoretical Computer Science at Télécom Paris during the double degree.

The Applied Maths syllabus at Unicamp is quite open-ended and, as such, I dedicated about half of my graduation to pure mathematics, and the other half to algorithmics.
At Télécom, I've had, most notably, courses dedicated to the implementation of graph algorithms and even an entire semester dedicated to using Sage for algebraic applications, and with it I gained some experience with its uses.

Albeit never having worked with open-source per se, I'm very passionate about it. In terms of experience, however, I have worked for two years as a full-stack developer using Python, C++ and other languages (SQL, javascript, shellscript, ...), even acting as code maintainer for a while, and thus I am quite accustomed to the git workflow of big projects; I have even already started getting acquainted with your development flow by making a small contribution to a documentation ticket #33271 in Sage.

From the ideas list, the two topics which have interested me the most were:
- Rewrite exterior algebra and implement Gröbner bases
- Edge connectivity and edge disjoint spanning trees in digraphs

The first idea is very closely related to the subjects I've studied in Applied Algebra, with special mention to Gröbner bases which were often used during courses (I might need a refresher though). In this topic, I'd love further clarification on ticket #32369 for rewriting exterior algebra (namely, the starting point: is the problem in the ExteriorAlgebra class in algebras/clifford_algebra.py ?). This project might require me to understand Cython more deeply, as well as find out how one would go about implementing the bonus "ambitious" goal, but I'm up for the task of learning these!

As for the second one, I've always been very interested in graph theory and their algorithms, this being one of my favourite topics, and it seems to be very well documented, discussed and explained in Sage's trac, even containing a paper with the related algorithm, so I'm ready to dive head-first into all of the discussions relating to this project. Should this be of interest to anyone, I have a small project on GitHub implementing an algorithm for an approximation to the densest subgraph problem in linear time.

Cheers and all the best,
Pedro

[tcscrims]

Hi Pedro,

   The exterior algebra is a subclass of the Clifford algebra (as it is a special case), and the problem is the basis is indexed by subsets of {1, ..., n}, which is a bit wasteful in terms of memory and manipulations. These are much more compactly written as integers with n bits. So the first step would be to change the implementation to use these. One of the fundamental principles of that project would be to make the exterior algebras very fast over a variety of rings in order to compute quotients and modules. As mentioned in the description, another related project would be combine the polynomial ring implementation with the exterior algebra to compute natively graded-commutative algebras.

Best,

Travis

[david....@gmail.com]

Concerning packing arborescences (edge disjoint spanning trees), I added a more recent paper on the page of projects proposals. It improves the time complexity, but I don't know how difficult it is to implement. Having a proper implementation of Gabow's algorithm would already be a nice contribution.

David.