Continuing from the previous post about how Google Summer of Code started for me...
Since I was a bit intimidated to approach Melange folks, I started out with Django. I had a dream of using Django models on App Engine. At the moment Django models cannot be used on App Engine because Django has database backend support for only traditional relational Databases like MySQL, PostgreSQL, SQLite, Oracle among others. App Engine uses a schema-less backend so it is not possible to use the same Django models. Incidentally it so happened that the same project was listed as a project idea on Django Project ideas page.
Also the reason to apply for Django was that, I was a big fan of Django (and still I am :P). I had started doing some work for Django initially, had read the backend ORM code. I wanted to help Django contributor, David Crammer in completing the implementation of multiple primary key support and other stuff.
Also I was happy to see the project that I had dreamed about. But later as the discussion on the project ideas started, it was told by Django developers, including Malcolm and Jacob Kaplan-Moss that, this project is extremely large and may not be possible to be completed in GSoC timeline. So I gave up that idea after thinking about it a lot, obviously I did not want to take up something that community itself was not sure about. I saw this idea called refactoring/restructuring of Serialization in the project ideas list and developed interest in it. I read about the current implementation of serialization in Django, its code among other things, did a bit of research on it and started discussing about it. Thanks to Russell Keith-Magee and Malcolm Treddinick who made the entire discussion possible and helped me learn a lot of things regarding Django internals and serialization. I also came up with a proposal (Click Here to read the entire proposal) and submitted it.
Update as of April 22nd, 2009
This one did not get thru for GSoC 2009.
My guesses are that, I did not put in as much efforts I put in for Melange proposal and the project itself in general and also may be because Django had better proposals than mine. But still I have no regrets. I am happy with Melange. I am *the happiest*