GSOC Intro and Proposal Questions

[Tamandeep Singh]

Hello Everyone.

I am Tamandeep, student of Delhi Technological University. 

I did my homework and went through everything. 

I have a few questions regarding Edge connectivity and edge disjoint spanning trees in digraphs

1. We can also implement round robin in such a way so that it augments each path upon discovery , but not done in research paper, should I implement that also or stick to implemention given in research paper.

2. As two algorithms given in the research paper apply on both directed and undirected graphs, but the implementation for undirected graphs is already improved in the issue. Shouldn't we use these algorithms for both undirected and directed graph.

3. As cycle scanning is not valid in round robin, but we can generalize cycle, should we do that also because it achieves the same efficiency. 

4. Round robin is extended to multigraphs also, but should we focus on simply graphs or we should also do for multigraph also? 

[david....@gmail.com]

1. I suggest to first get a valid implementation of the algorithm from the paper. Optimization can be added after
2. We can have several algorithms for directed and undirected graphs. I have added the Roskind-Tarjan algorithm for undirected graphs as it is simple and performs well. It will be useful to check results if another (faster) algorithm is implemented.
3. I'm not sure I get the question
4. The goal is to get at least get a valid and efficient algorithm for simple graphs. Taking multiple edges into account is usually much more complex. So I suggest to focus on simple graphs, unless we have a simple way to deal with multi edges.

Sincerely,

David.