Sure, I can fix it manually to a certain degree - but the question is: Is there a more accurate or “correct” way - if possible some kind of “automatic” way to avoid or correct these bumps? Even a Grasshopper / Kangaroo solution would be ok. The more automatic and “correct” the better.

Here is the file used in this demo problem.3dm (170.8 KB)

Any suggestions welcome - also to already existing forum discussions which I missed while searching.

@theoutside
I am not completely convinced on this solution - as it depends on this position / z-height.
And with have a “wild” combination of N5-Gon, Starpoint, Triangular Face.

@inju Thank you very much for your help and even creating a video. I really appreciate that.

However I have problems understanding how you set the two new vertex points - as also mentioned by @Tom_P

Is that at 25 % of the edge length (half of half the length) or is it just placed somewhere in between center and end of that edge? Sorry for being so picky on that - but as mentioned, I am searching for a solution that is “correct” as much as possible.

Sorry for my poor English.
Just a reminder, SubD is not for precision modeling.
The answer to your question, showed one of the methods
of adjusting the point by height, at random took 1/4 of
the length of the edge, for a more accurate adjustment
of the height of the point, select the position of the inserted point on the edge.

You are right, for a more or less accurate size,
the position of this point on the edge must be picked up.
The position of this point depends on the shape of the model itself.
It’s just a method, not an axiom.

Assume that the purpose of SubD is the beauty of form, smoothness, not accuracy.
The fact that I showed you this is one of the methods, I do not pretend to be true.
Quotation:
The arbitrary-looking barycenter formula was chosen by Catmull and Clark
based on the aesthetic appearance of the resulting surfaces rather than
on a mathematical derivation, although they do go to great lengths to
rigorously show that the method converges to bicubic B-spline surfaces.
In general, approximating schemes have greater smoothness,
but the user has less overall control of the outcome.

Maybe an extensive (20-40 minutes) video about “Getting more precision in SubD” ( by @BrianJ ?) would be a great idea. I am thinking of using construction lines and arcs or regular shapes like hexagons and octagons on well constructed planes to snap SubD vertices to. And also an explanation of iso curve flow in SubD and some tricks on typical problems in SubD and how to avoid or solve them.

That should give more correct results (which in many cases also means greater aesthetic appearance) than guessing and fiddling and hoping for the best. And a YT video on the official channel would be seen by more Rhino users.

On this subject there is good educational material, authored by William Vaughan, a series of illustrated books - The Pushing Points Topology Workbook, I in my early posts gave links to these books, and other materials on the correct topology of SubD.

Great tip - thanks, I’ve put that into my wishlist.

However I think that - apart from correcting and optimizing topology which is probably the best first step - I assume this book comes more from the artist/designer side modeling nice looking organic shapes. These offer a lot of artistic freedom and personal taste. I am not saying that this is easier or requires less skills. It is just different.

What I was initially hoping to get was more from the engineering side constructing technical hard surfaces in the best possible quality where there is not much personal taste involved - like the surface of a shiny faucet. But maybe that’s what NURBS surfaces are meant for.