Some modelling mysteries to solve

I have wondered quite long time but is it better to sunk one piece of model tiny bit into another or just move really really close to each other? Model will be joined eventually. I know rendering do not really care but how UE reacts better to this. I have considered deleting unneeded inner faces too that will not be visible and do not get any light.

Meeting flush is generally the best way to go. Lightmaps will have bleeding errors if meshes don’t meet up perfectly.

So best way to go is to sunk it little bit in and what about deleting that face?

Nooooo. Do not sink it in a bit, have it meet up flush. There’s no reason to delete the face, it wont improve performance.

Sorry for being dumb but what is “flush”? Not native English speaker here.

Flush means completely level or even with another surface.

Thanks, then i probably have to snap it somehow to surface if it is important to Lightmass. Blender 2.79rc i use.

Just curious does this not affect something if faces are snapped?