So, if you follow my blog you probably remember that I said I would not be at JavaOne this year... But if you follow my twitter feed you'll already know that at the last minute I had an opportunity to go that I could not refuse. So, here I am at Moscone once again.

This year, thankfully, I am doing nothing remotely stressful and can actually just enjoy the show. Oddly enough, I find it a bit hard to do so -- I'm just not that used to "attending" one of these things I guess. :)

Yesterday I dropped in on a talk by Doug Twilleager on MTGame, a "veneer" if you will that Wonderland has placed between them and jME adding some features such as better multiprocessor support, abstraction for collision and picking models, an object model and a few other odds and ends. The talk went smoothly and was interesting, but I actually found the questions at the end to be the best part. After all these years of trying to get the word out about Java, games, and performance, several questions were still along the lines of "Don't you take a big hit by not being in pure C++?!?" Doug even showed a simple demo of 200 teapots spinning at 100FPS and some in the crowd had trouble grasping that this is all client side, no tricks, and yes that's really the framerate. (And honestly, it would have been a dismal frame rate if that were the point of that particular demo - which it wasn't.)

I guess even among Java developers there is a very long road to understanding here. Ah well.

MTGame itself was interesting; I like some of the concepts (perhaps not their specific implementation in mtgame) such as the process/processor model. I think a similar system could make a good replacement for the jme-inherited game tasks and the task queue. I wish that they had worked a little closer with the community instead of waxing over the areas they did not like... but then again, in my experience that seems to be the norm in these situations. (And helping to build Ardor3D could make me part of that crowd I suppose :)