Thank you everyone who has answered so far.
A year has gone by. Has anyone found a general solution?
How do you robustly sweep profiles along rails in Rhino?
Generally speaking, the problem arises when a (planar) rail has tight inner corners and the (planar) profile is wide. The problem gets worse as the inner corners get tighter, the profiles get wider, and the parts get more complex. The first post in this thread explains the problem in some detail. This problem still remains. Both sweep1 and sweep2 are affected.
Please note that the image in the first post illustrates merely one case.
Our question is not just about the illustrated case, but how to solve the general case.
As to the individual posters:
@davide76 - Yes, there is something to this. But let us be kind to the McNeel folks, they are doing their best. That said, it is good thing to let them know what we users prioritise.
@cadmaster - Yes indeed, we like to work sometimes with meshes, sometimes with nurbs, and always ending up with solids. Sketchup certainly is not âgarbageâ, nor is Rhino. They are different tools for different jobs. You can create a masterpiece with a piece of paper and a turd if youâve got the most important thing - talent.
@mcramblet - Price tag is not the main issue here. The price of the software will not explain why certain programs can handle these cases, while certain others cannot. Once you understand the problem, the code required to make sweep work in this instance, at least for cases where both rails and profile are planar, is probably a matter not of weeks, but of days.
@skysurfer - Thank you. Your method seems to work in the example in the first postâs image, but please note that this image is provided only as a pedagogic example. In the general case, both rails and profiles may have subparts in the hundreds including curves. So although offsetting and extruding may help in the particular case illustrated, I fail to see how it works in the general case. In the general case, more complex forms need sweep or equivalent. And then we are again back where we started.
@rabbit - My example invoking the millions of historical plaster cornices out there was made to bring attention to the ubiquitousness of swept forms similar to these. They really are everywhere. A more robust sweep function assumes no âmagicâ. Why? Please look at the first image in the first post. Few people would argue that Rhinoâs actual Sweep1 result is preferred above the intended result. There seems to be general agreement that sweep should solve these cases as well. Not only that, there is general agreement on what the result should be.
@Sabino - Yes, Rhino would improve if McNeelâs developers gave certain basic tools more attention. For instance - no pun intended - the handling of blocks remains somewhat convoluted for such a basic feature. But that is another question. Let us agree that sweep needs some attention.
@wim - I agree that we users need to clarify to McNeel what we need. Still, âsweep1â is more fundamental than, say, âflow along surfaceâ. âLineâ is more fundamental than âsweep1â, and so on. In most workflows, 3D forms are constructed cascading historically from simple to complex. Therefore, they are hierarchically dependent on each other. So there is a case for commands generating more fundamental geometry to be at least as robust as commands further down the general workflow.
@nz_sales - You are right in principle. However, Sketchup, and allegedly Geomagic, too, solve sweeps like this in one single go. So some convenient solution is more than possible - it is actual.
@Stratosfear - If true, this is good news. Please describe one of the âeasy workaroundsâ you mention, step by step, for the general case where profiles and rails both have straight and curved parts, both in the hundreds.
@2DCube - Good to know. An excellent consolation for destitute 3D modelers:
âIf it feels like there should be a simpler solution, there usually isâ. Well, let us go there.
@carvecream - Thank you for your kind words. Exactly my point.
@cdordoni - Yes, in a way. Better tools have been indispensable to the history of economic and scientific progress. I have read that the oldest known tool was a stone pebble which was chipped to have a sharp edge. This tool was manufactured in the same fashion for quite some time. After about 500000 years, however, some primate simply had enough. He improved the pebble by adding a second edge, making each tool last twice as long. The next improvement, however, was much quicker. This time around, it took only about 250000 years to arrive. Lesson learned?
We humans need to become better skilled at specifying what tools we need, and why we need them.
This sweep issue remains unresolved.
In the meantime, we have used other 3D modelers.
How do you robustly sweep profiles in Rhino?
Hoping for good news.