Brushing off the dust

Following up on the Open Wonderland meeting; I grabbed the last version of the IMI avatars project from its google code repository and started integrating it into the WL codebase. This newer avatars code base (newer than the version that is currently in WL) is actually composed of three projects: MTGame, EngineCore and Avatars.

• MTGame - multi threading, component based architecture, rendering framework
• EngineCore - scene graph, collada loader and more
• Avatars - human avatars

(more details: http://code.google.com/p/cosmic-engine/wiki/SoftwareStack)

My first step was to switch the Avatars jar that was used by WL avatarbase module (the WL module that is responsible for integration with the avatar system features) with the newer version of the Avatars jar and include the EngineCore jar. Some modifications were done in MTGame so I'll have to address that as well.

In the old Avatars version, the rendering of the avatars was done with a JMonkey root node acting as an adapter, in the newer version there is a new renderer that is managed by MTGame, it performs culling, flattens transforms, calculates scene graph bounds etc. The new version doesn't use any JMonkey nodes, but it does use the JMonkey geometry classes with an MTGame compatible renderer. While I was getting everything to compile again I disabled name tags - the implementation of the name tags used the JMonkey root node of an avatar. There is an alternate implementation for name tags in the new system so once I get some basic tests running I'll put that back.