Hi Bruno,

re:
> I re-built them and then the sweep was much more cleaner. That Is the reason why I am
> confused by the ''- Sweep does not use the control point structure of the rail curves directly"

Maybe the dense curves also had some slight oscillations which can be difficult to see, and the rebuilt ones are smoother.

Or it could also be a matter of what's called "parameterization" of the curve, which has to do with how evaluating the curve might have spots where marching along the curve in the curve's parameter space may be either stretched out or compressed together instead of producing evenly spaced points.

Anyway, just to demonstrate it try sweeping the 2 cases in the attached file and then look at the control points of the rails.

Dense control points do give more potential for a curve to possibly have little tiny wiggles in it though so it can be good to be suspicious of them but it's not necessarily automatically bad.

- Michael