FillSrf - Challenges - support the development

FillSrf Challenges

looks like @menno is developing a great tool _fillSrf
to allow similar functionality like Xnurbs or parasolids Fill command.
let s collect challenges to support this development in this topic.

here is a first challenge

watchmaker-tool-challenge

people will cry “subd” i know - but it is super technical / geometrical and therefore more a nurbs modelling.


the transition at the “head” is a nice challenge - without additional details.

uhrmacher_rh9_fillsrf.3dm (3.2 MB)

FillSrf does a nice job (violettt) with G1 but needs a inner curve (blue) to do it.

I moved my wishes / feature Request to this topic

looking forward to see this command being developed and fine-tuned. thanks a lot - kind regards. tom

3 Likes

@Tom_P can you edit this post and make feature requests a separate post, or better, topic?

1 Like

5-side-hole - guitar neck

taken from this topic

FillSrf


there is some minimal bumbyness - it can also been seen when pulling a curve to the _fillSrf

onshape / parasolid

that s the winning result - also in attached file

i think it s a nice example to finetune some of the crazy advanced pramaeters

guitarNeck5side_00_fillsrf.3dm (4.6 MB)

looking forward to the next updates / developments - cheers -tom

Looks like the bumpiness in your FillSrf result can be improved on by using DegreeU=DegreeV=3 and Bending=0 in the Advanced options. This is food for thought, thank you.

1 Like

nice result.
those advanced options are mysterious and look like alchemy to me.
Maybe a first step to handle them would be a collection of successful fine-tunes (like the example above ?)
Maybe this collection will point towards some rule of thumb for the number of convex/concave /s-shaped edges.

Alchemy - got it. I’ll soon be turning out gold nuggets from each succesful FillSrf command :wink:
But joking aside, these options need to be explained better if they are mysterious to you. Or, even better, not shown at all to reduce cognitive load.

I am a big fan of having both

  • a reduced, simple approach with preferences or a single value
  • advanced options.

(the mesh command is a nice example, ToNurbs (G1, G1x, G1xx) … is a compromise.)
I think it s a question of how this command is developing / is used.
…or even this is something that can be finetuned for Version 10 in a few years, based on user-statistics.
All my students will love this command, but only 10% (including me) will be interested in the technical background and dig into advanced options.

so maybe a few predefinde settings

  • default best practice
  • convex/concave only
  • bumby-s-challenge-simple
  • bumby-s-Challenge-extreme
    and a custom/advanced setting with all the exciting parameters.

and yes - would be nice to have a bit explanation / documentation on the parameters.
at least if they are 0…1 or 0…10000
and if they increase / decrease computation (or stability)
… i already managed to crash rhino with them.

atomic bomb

the ultimate brutal unprofessional solution for details - if it s 23h and you have to start the 3d printer for tomorrows presentation…

not nice, not bad…

atomic_bomb_fillet_demo_rh9.3dm (894.4 KB)

… of course trimming with a sphere is very brutal and unsophisticated.

cheers - tom

@menno @Gijs Please do not eliminate the ability for the user to set degrees. For most situations I have a strong preference for degree 3, while recognizing other users prefer higher degree surfaces. To “reduce cognotive loac” provide the option for the degrees to be set automatically by the user setting the degrees to 0, similar to how some mesh settings, etc work.

1 Like

no UI will be hidden, but the idea is to have the advanced settings that you have right now separated from the rest.
Here’s a current mockup (but it is still far from finished)


8 Likes

Does this look good for the challenge?
It’s from a file mabe by one of my student. They always find a way to surprise me.

20250218_CrazyCorner.3dm (2.7 MB)

2 Likes

that model is quite a mess, quite a challenge indeed then… one edge does not even intersect, the other overlaps… maybe they clean that up before they make it a challenge :smiley:

the surface on the left requires at least one G0 edge:


20250218_CrazyCorner_FillSrf.3dm (190.3 KB)

4 Likes