When I first came across SketchUp 4 I was so impressed by the inferencing it did while you constructed your model that I fell in love with SketchUp.
Today, I was modelling a small, fiddly model for 3d printing (scaled up to avoid any scale snap issues) and found the inferencing in SU2024 mostly didn’t do what I wanted. Really frustrating.
So I went back just to SU2020 and found the inferencing was much better…
Is that possible? Has the inference engine under the hood been changed so much?
I have also noticed occasional awkwardness inferencing in 2024. Purely speculation, but I suspect that the change to the new graphics engine may have affected the inference engine by some subtle changes in how SketchUp detects where things are on the screen. It may have entangled with an attempt to reduce OpenGL’s notorious problems when geometry is too far from the origin. You might try switching to the classic graphics engine to test these theories.
I believe (correct me if I’m wrong) that the use of OpenGL feedback functionality to find picked geometry was an historical artefact from the early days of @Last.
I am presuming (hoping ) they now use an analytic picking implementation - ie calculate where the picking ray intersects the geometry.
So not sure that would explain it. I do have my cursor scaled a bit - perhaps I’ll try with no scaling but my sense thats not It.
not sure who “they” are though.
) they now use an analytic picking implementation - ie calculate where the picking ray intersects the geometry.