I’ve encountered a situation where it would seem to make sense to break the rule of keeping all geometry on Layer0. Other than being branded a Sketchup outlaw, I’m wondering what terrible things will happen if I do this. Here’s the story:
I use Sketchup to design cabinetry and interior millwork. My rooms are just context - they don’t have to be very detailed, but they do have to be accurately sized. Typically I start from an architect’s plan, and later revise based on actual site measurements. Because I need to control visibility of each separately, I put the walls, floors, and ceilings on separate layers. I draw the walls on Layer0, group them, and assign it to the WALL layer. Same with floor and ceiling.
Because of this, whenever I revise a drawing to reflect actual dimensions, I have to make each and every adjustment three times, to the walls, floor, and ceiling. This seems inefficient and prone to error. But if I were to put the walls, floor, and ceiling all together in one group, and assign the walls, floor and ceiling geometry to the appropriate layers within the group, I could use Sketchup’s inherent “sticky geometry” to my advantage. Whenever I moved a wall 1", the floor and ceiling would automatically stretch with it. That would mean less work and fewer errors. And I’d still have the same control over visibility.
Draw a simple cube.
Assign the selected top face [no edges] to layer CLNG.
Then the bottom face to layer FLOR.
And the other 4 vertical faces and their 4 edges assigned to layer WALL.
Now switch those 3 layers on/off to see the changes.
So far so good…
Now with just LAyer0 and the WALL layer visible, erase a top and bottom edge of one of the wall’s faces - the face vanishes.
Now switch on layers CLNG and FLOR.
There are no faces on those layers, because the supporting edges are now missing.
Let’s alternatively assume you erase the top edge of one wall face, so it vanishes.
You redraw that edge and the face reappears.
That edge and face are on the current-layer - probably Layer0 !
And no longer controllable using the WALL layer.
Also a face will have appeared for the ceiling - but rather than be associated with CLNG it’s also assigned to the current-layer [Layer0 ?] !
So you now have to reassign some layers again…
Had the face planes/edges been inside groups or components with all geometry with Layer0, and other layers assigned to the ‘containers’, there is less layer reassignment needed after editing…
Provided that you accept all of these pitfalls, then do as you want - after all it’s your model and your sanity…
Those are certainly the sort of pitfalls that convince me to keep my geometry on Layer 0.
The main problem I’ve run into with separate groups is that mistakes are hard to spot. If I move a wall 1/2" but forget to move the floor the same amount, the misalignment is quite small relative to the size of the room. Small enough that when I export to Layout a half inch to the foot plan or elevation and dimension it, I may think I am snapping a dimension to the adjusted wall, but actually it is the edge of the unadjusted floor, which at that scale is indistinguishable from the wall. This will result in an incorrect dimension, which I have little chance of catching unless I already know what the dimension is supposed to be.
For this particular purpose, I rarely erase or create any lines when adjusting after site measure. Rather, I just move things a bit this way or that. I do occasionally need to add or remove a wall, and no doubt could easily end up with some geometry on the wrong layers, which would mess up my scenes. But that problem would at least be easier to spot, and wouldn’t ever (I don’t think) result in a cabinet being made 1/2" too small.