Well, getting the next version of the Lunar Lander game out the door for Christmas isn't going to happen.
The LLG plan is to free up CPU cycles from the graphics by switching from 2D to 3D. Which is counter intuitive. In 3D, the game can better access the graphics hardware via Open GL (on Java).
That path is panning out, but 3D is a pain in the neck. Quaternion multiplication is not commutative, (p*q) != (q*p).
I'd describe the situation as vector math being so powerful -- so much is accomplished with such simple expressions (like a*b) -- that the ideas implied must be held quite firmly in mind. But once one has got 3D coordinate spaces and transformations firmly in hand, it's a lot of fun. At least it's a lot of fun just getting there.
The software that I've found in the world for developing 3D applications isn't quite what I'm looking for. So I'll write some. And as "what I don't know" has a possible equivalence with "what unknown future applications will do", I'm just writing tools to open the doors on GL programming for web centric applications for myself.
The LLG game will be a first demo, flying a model. Then there's a list of things to do next.
As I get my head into 3D, all the world of web data applications looks like, "why aren't we doing this in 3D, now?".
I think there's a contribution to be made, here, in what we're looking for and how we're reaching for it.