GSoC Physics

Line 9: Line 9:
More physics improvements are of course possible:
More physics improvements are of course possible:
-
* Different kart physics could be implemented, e.g. some karts might turn better, have higher speed, higher mass, don't touch the ground (hovercract, flying carpet), ...
+
* Different kart physics could be implemented, e.g. some karts might turn better, have higher speed, higher mass, don't touch the ground (hovercraft, flying carpet), ...
** This needs additional testing code to make sure the karts are reasonable balanced, so  
** This needs additional testing code to make sure the karts are reasonable balanced, so  
** Also the AI might need to be aware of different kart physics
** Also the AI might need to be aware of different kart physics
Line 15: Line 15:
* Support for real motor bikes (i.e. bikes that actually lean into a curve), perhaps add a wheelie feature.
* Support for real motor bikes (i.e. bikes that actually lean into a curve), perhaps add a wheelie feature.
-
 
+
A proposal for a GSoC project related to the physics does not need to suggest implementing all ideas mentioned here. The most important aspect of a proposal is a realistic schedule.
-
A proposal for a GSoC project related to the physics does not need to suggest all the ideas mentioned here. The most important aspect of a proposal is a realistic schedule.
+
== Hints ==
== Hints ==
-
Due to the larger number of classes that might need to be changed (including bullet) good OO-design skills and quick understanding of existing code would be rather important prerequisites for this task. Also maths and some physics skills are important
+
Have a look at our [http://supertuxkart.sourceforge.net/New_drivelines driveline page] (and the [http://supertuxkart.sourceforge.net/Theoretical_Driveline_Model| theoretical driveline model] page). At this stage STK uses a simple projection of the kart position onto the XZ-plane (by dropping the Y coordinate) to determine where the kart is. This obviously will not work anymore when the 'up' direction can change anytime. A proper implementation of finding where the kart is will probably be the most difficult part of this proposal. Due to the larger number of classes that might need to be changed (including bullet) good OO-design skills and quick understanding of existing code would be rather important prerequisites for this task. Also maths and some physics skills are important.

Revision as of 21:09, 9 February 2014

Physics and Game Play Improvements

Using a corkscrew or loop-the-loop track allows artists to create much more interesting tracks. Unfortunately, this kind of track was not considered with the original game design. Some preliminary work has been done: in our current development version is it possible to change the gravity of a kart depending on its terrain, so karts can drive on the wall or ceiling. But many more features need to be added to make this useable in a game:

Those changes are significant, and it is expected that a significant number of classes need to be improved to handle this.

More physics improvements are of course possible:

A proposal for a GSoC project related to the physics does not need to suggest implementing all ideas mentioned here. The most important aspect of a proposal is a realistic schedule.

Hints

Have a look at our driveline page (and the theoretical driveline model page). At this stage STK uses a simple projection of the kart position onto the XZ-plane (by dropping the Y coordinate) to determine where the kart is. This obviously will not work anymore when the 'up' direction can change anytime. A proper implementation of finding where the kart is will probably be the most difficult part of this proposal. Due to the larger number of classes that might need to be changed (including bullet) good OO-design skills and quick understanding of existing code would be rather important prerequisites for this task. Also maths and some physics skills are important.

Retrieved from "http://supertuxkart.sourceforge.net/GSoC_Physics"

User Tools