by fzammetti » Mar 27, 2005 @ 12:33am
I don't agree with that. This discussion has been done to death frankly, but still...
The one true benefit of a frame-limiting solution is that you know that your game will run the same on all devices. There is nothing I hate more than a game that runs smooth as silk on one device but runs like a pile of shit on another. Sure, the SPEED is identical, all the TIMING is the same, but the SMOOTHNESS is not. Any time I see a game that gets choppy I walk away from it in a hurry. Frame limiting, WHEN DONE RIGHT, avoids this.
I'd much rather determine the minimum frame rate my game needs, and in fact DESIGN the game to run at some tiven rate on ANY device, and go from there. The trick is to pick a baseline device that no device you will support will perform worse than, and use that as your guide.
I always use my trusty E-125 for that. For those that haven't been around as long, that's a 150MHz MIPS device running the original PocketPC OS. Slow as balls compared to ANY modern device. So, if my game runs consistently at 30FPS on that, I know I'm golden.
Of course, there will always be situations you can't design for... Someone that underclocks their 600MHz PocketPC to 200MHz to save battery is going to make life difficult for any game. I contend you shouldn't be designing for such situations. But, if you target an E-125, or even an original 200MHz iPaq would be safe these days, you will be fine.
Now, to be fair, some games you just CAN'T do this with. I would NEVER use this technique on a 3D first-person shooter. Just wouldn't work right. Many other types of games probably wouldn't work well either. But I dare say that most 2D games probably would be fine.
Aside from the smoothness benefit, you also greatly simplify your code. Well, maybe not *greatly*, but somewhat anyway.
As with most things in development, to say one thing is good or bad, right or wrong in absolute terms is usually a bad thing to do. If you understand the implications and make your decisions accordingly, even the "bad" architectural decisions can be good ones in some cases. Frame limiting is one example of this.
...and so I said to Mr. Gates: "$640 billion should be enough for anyone!"