[asterisk-dev] GSoC 2010 - Calling for Project Ideas and Mentors

[Mark Michelson]

Russell Bryant wrote:
> Greetings,
> 
> The period for organizations to apply for participation in the Google 
> Summer of Code program for 2010 is next week.  That means that by the 
> end of next week, we need to put together a couple of things:
> 
> 1) I need to know who would be willing to mentor a student.  This is not 
> an extremely large commitment.  You are there to provide some technical 
> guidance, as well as general guidance for how to interact with the 
> Asterisk development community.  If you have questions about being a 
> mentor, there are a number of people on this list that have done this 
> before and can provide feedback.  Feel free to also contact me directly 
> on this topic.
> 

I would certainly be willing to do this again.

> 2) We need a suggested projects list.  Historically, I think our project 
> ideas have been a bit too big.  We need to aim for coming up with 
> projects more concretely achievable in a summer for a college student 
> not necessarily familiar with the Asterisk code base.  I think coming up 
> with ideas for specific tests to implement is a really good area to 
> think about.  Please post any project ideas that you have so that we can 
> discuss them and get them on the list.

I have a couple of ideas for student projects that are based on old projects 
that I initially pursued but then abandoned either due to a lack of concrete 
requirements or due to other limitations I encountered.

One idea that I had for a project, and which I pursued myself at one point was 
to expand ao2 containers to support more containers than just a hashtable. In my 
previous attempt at this, my ideas were to employ a linked list and skiplist. 
The biggest problem I had was that I did no planning; I just dived right in and 
began coding. When I hit hurdles, I would have to make huge changes in order to 
correct the holes I dug myself into. Eventually, the code was such a mess that I 
just decided to move on to greener pastures.

This on its own is probably too large for an SoC project. However, if someone 
were to set up an astobj2 container framework beforehand and convert the current 
hashtable implementation to use that method, then there would be a template in 
place for a student to follow. With that, the student would have a lot less work 
ahead and could just implement various container types. This, I think, is better 
scoped to an SoC project.

Another idea that I pursued briefly was an attempt to have all manager events 
trigger ast events as well. Once again, the main reason I abandoned this project 
was because I jumped into coding prior to actually planning or discussing anything.

Once again, this is probably too large (and more importantly, too nebulous) a 
project for a student to try to do as an SoC project. However, if the proper 
preparation were done by the mentor (i.e. defining requirements, splitting the 
project into achievable milestones), then the actual student project could be to 
achieve one or more of the milestones set forth.

I will try to think of other project ideas as well and will present them in this 
thread as I think of them.

Mark Michelson

> 
> I will be gathering project ideas and saving them here:
> 
> http://svn.digium.com/svn/asterisk/team/group/gsoc-2010/ideas.txt
> 
> Thanks,
>