Personally, I’m fine with the current collision detection method(s). I don’t know how much “weight” it adds to an app to include the physics engine (performance-wise and also compiled app size), but in my experience, it’s worth the slight penalty for what is gained overall.
Perhaps what many Corona devs expect, or assume should be present, is a pixel-based collision system versus a “body-based” system. I think this would be useful in some apps, i.e. if you’re making a “tiled maze game” with tons of individual boundary blocks or something. Using the current system, you need to draw a body for every block (or place complex bodies spanning around many blocks, and they can’t be concave or greater than 8 points, of course!). From a developer’s standpoint, that can be alot of extra work, especially if you have dozens of levels or if levels are randomly generated.
That being said, when Corona Levels (the level designer tool) becomes available, it should help streamline the physics body setup and allow for drawing physics shapes in “real time”.
At this point I can’t say that I feel strongly either way. I haven’t yet needed a “pixel-based” collision system, but maybe in the future I would find that very convenient, and I’d be urging Corona to add that in. 
Brent
[import]uid: 9747 topic_id: 30265 reply_id: 121359[/import]