Keith Peters put up a comprehensive blog post on why he disagrees with generic keyboard managers. It’s well worth reading. I have to admit, i couldn’t even conceive of input events directly altering the model without first passing through an input testing step in the game loop. It’s practically game dev heresy!
It apparently boils down to an optimization issue. I’ll say that if you need those extra tiny cycles, then sure, Keith’s approach is a fair optimization. But, call me conservative, i would personally never think to base my game development workflow on such an approach. Writing custom code for a very abstract step of the game framework on a game by game basis seems like an unnecessary speed bump. In other words, i’d go mad writing boring code before i got to do any of the fun stuff.
Iain Lobb is considering publishing his gamepad class, and i do believe i’ll do the same soon.