I thought everyone knew that "some" amount of roughness was necesary to interupt the boundry layer to increase flow. The thing that causes it is that the boundry layer is more or less static (ie not moving) and you get huge "shear" along the edges where the fast moving gas trips over/rips away from the stationary gas. By introuducing very small amounts of controlled turbulence you essentially add teeny weeny air ball bearings for the incoming flow to slip against. I've seen the same principle used on the hulls of boats to limit drag (actually used air pumped along the hull). Also that is why golf balls have dimples, and air plane wings have little ridges on them.
The trick is that is takes damn near a Cray super computer to figure out when the turbulence is just enough to help and not slowing things down worse than they were to begin with. The engineering of a simple golf ball is really probably a cumulative effort of 1000's and 1000's of man hours of effort, and if anyone tried to dimple something to improve it's flow they would probably need a CNC machine, a flow bench and lots and lots of spare parts to figure out by trial and error what helped and what hurt.
(Disclaimer: Okay, so I got a C in my fluid dynamics class, and haven't used any of the knowledge in over 10 years - so take it with a grain of salt)
97 SE MTX, KKM, UR UDP, 19# Injectors, SVT TB, SVT UIM, SVT Flywheel, Grizzly Street Clutch, Upgraded motor mounts, mystifed precats, brand new main cat, SVT exhaust sans resonator, 205-55-15 Dunlop SP5000's 154hp / 156 tq before the SVT exhaust, SVT dual honed upper and swapping the mystery main cat for mystery precats. My Profile
| 
