Dome forum.3dm (9.7 MB)
DOME FORUM.gh (11.6 KB)
i already have the dome shape but i can´t get the curvature that goes around the opening. My logic was to intersect with the mesh wich is the shape of the opening that i want and then connect the points but after that i can´t really think how to move foward with the script
I gave it a quick try, let me know what you think. I wrote some notes in the script, so you can follow my thoughts 
romtecmax-dome-forum.gh (21.6 KB)
Broken, but the basic idea is there.
DOME FORUM _ VR 01.gh (49.2 KB)
Here’s a method that lays the rib lines out on the ground, tweened from side to side, then SrfMorphs them onto the “bulb” and makes RuleSrfs between them. Nice sharp edges and corners. Door dims are set as fractions of whole surface for ease, but absolute numbers could be calculated with some more work. To further refine it, you could limit the tweening to a handful of lines either side of the door and have the rest be perfectly straight. I notice the “elbow” in the original image fades away quite rapidly.
parayurt.gh (25.5 KB)
Thanks for the answers guys, it helped me understand better my code, one thing that i needed to do was this curves that followed the opening but in my model the opening is way bigger than the model it was based of, where i´m stuck now is getting the extrusion of the lines it may have to be different widths but they get all buggy.
Dome forum.3dm (9.9 MB)
DOME FORUM.gh (22.3 KB)
Thanks for the reply! The elbow indeed fades away rapidly, in my model the opening influences much more on it, not making such a rapidly fade away.
Nice definition. That is how I would normally do it; work it out in the plane and then abuse tranformation tools… I have been familiarizing myself with the Pufferfish components and wanted to try using one of the Tween Curve on Surface tools. The problem with this is when the tween passes across the seam of the surface. It begins to view the curves as on the “wrong side” of the seam. The surface seam seems to be a problem in your script as well. I wonder if it can’t be fixed…
Yup. It isn't even that hard:
Dumpling House VR 01.gh (64.5 KB)
I’d rebuild the tween curves though. They look silly with a sharp corner.
Starting from @Tom_Newsom 's superior script, I worked on a definition that creates an elliptic opening:
Dumpling House VR 03.gh (75.2 KB)
I’m very proud of the basic math I did to get the partial ellipse. However, when the ellipse is projected onto the inner surface it goes from a very simple curve to a heavily convoluted one. I am hoping somone out there could show me how to simplify this because it is causing problems down the line in creating the ruled surfaces. I am not yet sure yet of my prefered method of creating the ribs that lie on the offsetted outer shell.
I’m digging through this and haven’t grokked it fully yet, but my first intuition is that you went to “bulb space” too early. Is it so super critical that the opening be a perfect ellipse? You go to great trouble making an ellipse of exact dimensions (of which you should be proud - it’s nice little algorithm), then distort it by projection onto the bulb, then distort it by sporph onto the plane, then distort it by sporph back onto the bulb. Poor curve has been through the wringer so many times it barely remembers being an ellipse!
The power of the lay-everything-out-on-the-ground-then-sporph-it method is that all your geometry remains 2D and easily graspable right up till the very end.. By mixing things up, you introduce more opportunities for error (of the human and computing kind).
I think you could achieve a very similar result by modifying the arc-based opening in my original such that the “head” of the door has a horizontal tangent, and changing the aspect ratio of the sporph source to make it more or less elliptic (in the flat plane anyway).
Thank you for your input. You say you haven’t fully gokked the script, yet by your description you essentially have. Instead of working directly from the planar formulation, the extra steps of projecting the opening profiles on the 3D surface and mapping these to the plane are prepended. This is done to achieve dimensional accuracy. Working strictly from the plane to the final 3D form, you don’t know what dimensions the shapes you are constucting in the plane eventually become. So, to the question:
…yes, if this is to become a constructed object. While the original ellipse becomes deformed as it is projected onto the shell surface, it maintains a shape that congruent with the planar insert that is the entrance facade:
This is to the advantage to the window dresser, who, if not given a rough approximation of the opening shape and size, will be found on-site cursing profusely as he attempts to obtain measurements to this end.
I suppose you could argue that you could make a guess-timation of the opening’s shape in the plane and then derive the actual shape of the planar insert from the 3D form with dimensions as a plan. But this won’t be as nice as the clean ellipse the window maker would otherwise be working with, and the added steps of obtaining the planar insert would eventually add up to the same amount of work in the script as what I currently have.
My hope was that at this point in the script:
the curves could be cleaned up to maintain their approximate shape while reducing the number of control points to the minimum necessary and to an equal number between curves. The final output of the current script is functional, but very poorly formulated from a surface modelling point of view:
P.S. I’m not trying to preclude a different approach as I think mine may be wrong. I was just trying to detail my thought process here.