Exposition Linden: 1) Create an algorithm and OpenGL demo for converting SL geometry into 3D imposters. This is different than traditional 2D imposters. A 3D imposter is a reduced geoemetric representation that can still be lit.
Exposition Linden: A good place to start might be here:
http://www.cs.berkeley.edu/b-cam/Pa...en-2004-IAI.pdfExposition Linden: So, for example, a large building (this is a simple case) could be represented by a single, large cube. Imposters need not be constrained to SL primitives and should instead be created from polygons (albeit, a small number of polygons).
Exposition Linden: 2) Build an Open GL application that allows you to build flow charts, where output from specific elements act as inputs to other elements, with the ability to make different shapes and proper path finding for the connections.
Exposition Linden: Used as a UI element, this could allow for visual script building, material definitions, etc.
Exposition Linden: 3) Create an optimizer for transforming arbitrary 3D mesh geometry into SL primitives. The optimizer would need to be able to tradeoff processing time, error, and number of created primitives.
Exposition Linden: This would be an incredibly useful addition because it would be the major step in allowing data from applications like Max or Maya to be imported into SL.
Exposition Linden: 4) High quality machinima. Given the flexibility of SL, especially with private islands, it seems possible to create some truly amazing machinima.
Exposition Linden: 5) A cross-platform speaker dependent speach recognizer. There are a host of features that this would allow, but sadly most of the products in this area are either not cross platform or are speaker independent, resulting in much lower recognition rates.
Exposition Linden: 6) A high-quality text-to-speech system (again, cross platform) that allows dynamic blending of speech fonts or truly synthesized speech. Again, many interesting options appear if this technology was around.
Exposition Linden: 7) Demonstrate high-quality procedural locomotion cycles for different combinations of legs. More importantly, demonstrate methods for blending those cycles with non-kinematic and prebuilt animations.