as we say in milwaukee: library that shit!
you'll get better flow with slight roughness, in a turbulent regime, because the boundary layer will be thicker, the key is to avoid waviness which will set up conflicting eddy currents and disrupt the flow. If you've ever been water skiing, you'll know what i'm talking about, going outside the wake on smooth water, you feel a lot of drag on your skiis, but being inside the wake on the rough water, it feels much 'harder' and you feel less drag. now imagine your ski is the wall of the intake and the water is air. Its not a great comparason, but its the only thing i can think of that demonstrates it on a personal level.
having the inside edge of a turn neck in, can help because it prevents a flow seperation, where slow air on the inside of the bend drags on the fast air to the outside.
bottom line, two strokes have such sporadic flows from either the reeds shutting or the piston port closing, that it really doesn't matter. small flow differences are going to make an almost infinitessimal change in actual throughput of the system, because the sonic dynamics are so much more powerful.
if you really want to 'tune' an intake manifold, look into 'hemholz resonant circuit' i think bell might have something in his book about it, otherwise Heywood's 'internal combustion engine fundamentals' does a good job of explaining it... as does blair, right before he tells you its basically crap and then proceeds to blow your mind. but thats a much longer post.