modeling improvements are so left behind. Not everybody can really sculpt or afford a drawing tablet. I used to like nurbs workflow in the other old apps but seeings it faded and lack of updates I backed off from playing with them. Fusion 360 is not for artists.
a bit offtopic There are so many cases in which Pablo dobaro work on the sculpting branch could be far more useful in the forgoten edit mode polygon modeling.
There are not only artists.
And anyway, by system, by its nature blender will tend to cover all areas of design, it is the nature of the opensource system that imposes it.
When the other design compartments are solid, there will be even less demand on them, because the functions will satisfy the needs of the large mass of users, and developers will necessarily have to find new fresh air in which to invest ⌠and the makers area, of the engineers, of the architects is well mature, they are just waiting to be freed from the âcobwebs of spidersâ.
Perhaps only one reasonable approach for this project to succeed is to find a middle ground between the best of both worlds. Nurbs based modeling is superior in terms of speed and precision. However poly modeling is superior in terms of topology control.
From what I have seen in various other videos of pure nurbs-based modeling software, there are lots of decision and upfront planning mostly when it comes to how the surfaces can be joined and blended. And not to mention that there are lots of extremely advanced tools and features for these purposes that most of the times are derived from top-notch scientific papers. And despite the fact having these tools, these designers always double check their surfaces with zebra-matcaps just in case the result was not correct. I a make the point here, that pure nurbs-based modeling is sometimes too specialized with humongous backend development, and other times still too ambiguous because it makes surfaces too much high level that in the end the designer has no idea how they are composed of.
With this mindset, I can find the best other alternative choice is to discard any nurbs-based approach and go for a hybrid nurbs+mesh approach. The user can work only with nurbs while the backend engine is based on mesh patches. You might know this technique as spline-patch modeling, it is very efficient and useful. The designer will have only to consider that the surfaces form quad-spline-cages and then once this is ensured the backend system can figure out to construct the plane-patches accordingly.
I donât know exactly if this post sticks to nurbs discussion at all. It might be a topic for the spline-patch-modeling tools. If you think that this is a viable plan let us to consider think about it in a more serious manner. 
2022 01 18 13 29 46 - YouTube
attempt merge this branch against the newest master code.
Too bad the Rhino logo surfaces are not coming over well 
Are you still fully doing this with C++ OpenNURBS SDK? In case you werenât aware yet: https://github.com/mcneel/rhino3dm may be useful if you want to try work in Python.
I just want to bring the complete NURBS backend with domain Boolean to blender. File IO is only a part of it.
Yes, I understand you want to bring complete functionality. I was talking about the Python OpenNURBS wrapper rhino3dm in case you want to do less compiling and more typing.
OK, it must be very convenient to use this library to test after getting the internal data of blender correct. Thank you.
Feel free to ask questions about rhino3dm or OpenNURBS. If I donât know the answer I can always ask the other Rhino devs.
I came back, after working for a BIM software company for five months. I got into a graduate school of civil engineering in a university. So, there may be some time to get back to these problems.
For implementing improvements to surfaces, I would suggest this as the approach: T100127: New surface data-block