GSoC: convert Gnobots2 to Clutter

[Diego López León <diega locolandia com ar>]

Hi, I'm attaching the roadmap part of my SoC application looking for
feedback from some mentor.
My plan is to apply for convert Gnobots2 to clutter (as you can read
in the subject :))

<roadmap>
    Since I've dedicated some time to experiment with Clutter, I feel
quite comfortable with the platform to start coding as soon as
possible. The actors involved in this game seem easy to identify; I
will use one for the Human and several for the Robots (with some
possible performance issues that I'll describe below). According to
the characteristics of this game there is no need for the Timeline
concept, because all the events occur when the mouse is clicked.

    I think that by the end of the first month I could replicate the
current game experience using the new library. The main idea in this
stage will be to replace the game_area GtkWidget with the ClutterGtk
widget and change all the interactions with the game_area with
equivalent functions. For this first stage, I propose a simple
approach to visualizing different events like robot collision, winning
the game or being hit by a robot (losing the game).  My idea is to
replace the actors involved with new ones. Of course, this will change
later but this way I can achieve the one month milestone of not using
the ad-hoc canvas draw implementation.

    On the three following weeks I will add eye candy modifications,
like smooth robot and human movement, nice human teleport with at
least a shade, animated explosions a short 'seeking safe place'
animation for safe teleports. The main challenge here will be to keep
the same game experience as the original one. I like playing this game
because is agile and you can use it to really distract your mind
without "dead times" like in games like Nibbles (you have to wait the
snake get its food) or Tetris (wait until the piece hits the bottom).
I want it to stay agile, so I don't want to add long animations, like
big extra animated explosions, or slow teleports. I think I could fine
tune it during the development.

    The last weeks will be dedicated to performance improvements, It
probably will not really consume too many resources, but it's possible
to change things and experiment with features like ClutterClone (since
0.9) to replace the robot instances around. I think that this
experiments will take about two more weeks, leaving two weeks at the
end as backup time to contemplate any delay in this really early
estimations.
</roadmap>

Best regards
-- 
diego