[tombert]

I'm not in game development, and I'm a pretty mediocre-at-best artist, but I personally have found it considerably easier to draw pixel art simply because you limit yourself to like a 16x16 grid for a sprite. There's a lot less room to screw stuff up, and you get a lot of implied detail while also being pretty easy to be very precise.

If you took a vector drawing and rasterized it to 16x16, I think that might kind of look like crap, because it'd be hard to know ahead of time how the rasterizing will work. How will it handle diagonal lines? How will it figure out which pixels to shade and which ones not to? I'm not saying you couldn't make a vector renderer that handles these things in a way that you'd like, but I am saying that you'd most likely not be terribly happy with the default results.