Hi, in the attached file I ran Rail Revolve with the two curves on the left. The option ‘splitattangents’ is set to yes. The result, on the right, is not split at tangents.split at tangents not working.3dm (101.3 KB)
I think the “problem” is the input curve made by arcs. I tried with DivideAlongCreases and changing the documents tolerances but nothing solved it. at least you can extract some isocurves and split the object by force.
Yep, this looks buggy in both V5 and V6. Also with regular Revolve. If I start the command, the preview seems to indicate that it will split into 5 surfaces -
– but when the command completes it is only 3.
The current WIP also has the same issue.
This is a general behavior of SplitAtTangents. You would see the same thing with ExtrudeCrv, for example, or anything else that makes a nurbs surface from that input curve.
SplitAtTangents is carefully adjusted to not split at transitions from one curved part of a surface to another curved part of the surface unless the curvature reverses - an inflection point.
Right or wrong, that behavior was chosen quite a while ago.
This example has pretty big curvature changes at 4 places, but only two of those are changes from no curvature to some curvature and only those get split.
Since this works the way it was designed, it’s not a coding bug, but it may be a design bug. Those take a little more input to get changed. @pascal may be an advocate for you with that.
Oh wow…that explains so much! So basically, it isn’t designed to split at tangents except sometimes. No wonder I’ve been so confused for years about why it didn’t give me the results that I expected!
So…what’s the point of the option? I mean, why would someone want it to split only if it’s an inflection point? And why is it called split at tangents?
Yeah… we’ve never just gone ahead and tried splitting at all tangents - the one exception that might make sense is contiguous arcs of the same radius - but I think it’s worth a try to see it it messes things up for people - split at all tangents, don’t try to second guess…
-Pascal
Definitely a design bug. So still a bug - no matter how long it has been in existence.
Otherwise, you are going to need to stop calling the option SplitAtTangents - which it doesn’t.
2 Likes
I’d love to see that happen. (And yes, contiguous arcs of same radius not splitting would make sense…) I feel like the command should do what it’s name says, right?
Yep, split at tangents except where the radius is the same on both sides - i.e. split at G1 and not at G2.