I was reading up on sailplane turbulators (thanks to piquing my interest by reading stuff posted here) and found this site:
http://www.sinhatech.com/auto.asp
The deturbulator is claimed to work by reducing the drag caused by the big void travelling behind a hatchback or 4wd type profile car. It does this by creating some turbulence at the top rear edge of the roof which breaks the connection between the air being dragged behind the car and that flowing over the roof. This means the air behind the car is pulled along behind the car acting as a kind of virtual tail and not spending valuable energy dragging and mixing in a big way the air travelling over the top of the car.
Renault or Citroen were trialling versions for production vehicles last year with good results.
Obviously it is not perfect (It doesn't affect the turbulence caused by the sides or air coming from underneath) and a better solution is not to be landspeed racing a car with this boxy profile. It is a stop gap solution to an existing profile and not necessarily a method for designing a better lakester or bellytank.
Dirty Dave mentioned paint roughness.
To best maintain laminer flow at the front the smoother the surface the better. As the boundary layer thickens and flow separation occurs the surface roughness becomes less critical.
It makes sense that surface roughness could help slipperiness in the same way that the tapes do (or vortex generators) as all a tape is is a rough surface. What order of magnitude and what pattern of roughness is the big question.
If you have ever patted a shark you would know they are rough as all getout, not squeaky clean like a dolphin and it is one of nature's way of sliperriness (that has inspired Thorpe's swimsuit.)
When you get to this level of detail though there probably are far greater advances that could be made elsewhere on the vehicle for the same amount of research and effort.
If you are reaching for that last 1 mph for the record then it just might come in handy...
Rev. H+