Actually that kind of "fuse together tangent rail segments" thing that the Rhino sweeper does tends to cause a lot of problems with 1 rail sweeps, people have been requesting in the Rhino newsgroup for a long long time for Rhino not to do that.

But I think it makes sense to do it for 2-rail sweeps which can kind of get more chaotic with trying to handle multiple segments on an segment-by-segment basis.


Here's an example of when MoI's "treat segments separately" approach for 1 rail sweep yields a better result.

Here is a multi-segment curve made up of lines and arcs:
(curves attached as sweep_1rail_fuse_test.3dm)



If you do a 1-rail sweep of this in MoI, you get this result:



Each segment in the rail curve created one surface in the result object, you can then do stuff like Edit/Separate to break out one piece and do some work on it like trimming it, etc...

Here is what this MoI geometry looks like in Rhino so you can get an idea of the surface structure:




Now if you use Rhino to do that 1-rail sweep, you will get this result instead:



There are several problems - the object is one great big surface and can't be easily exploded into individual parts. Also you can see from the wireframe that the surface is more complex at the joint areas, it doesn't come to a kind of crisp clean end at each joint, they kind of blur together. There is some asymmetry to it, the surface does not have the same kind of uniform density throughout even though the originating rail structure was very uniform.


So it was to avoid this problem that I made MoI's sweeper work on a kind of separate segment-by-segment basis instead of trying to fuse things together.

But in your 2-rail sweep case, fusing things together is what makes a better result there, it is just not as straight-forward for 2-rail sweep to try and treat it on a segment-by-segment basis, I think I will try to do the fusing for 2-rail sweep but leave it as the non-fused method for 1-rail sweep.

- Michael