Often, I overlap two meshes together to make some desired geometry. In this case, there are always invisible faces.
When I was just beginner, I made some bad topology sometimes, and this makes bad UV results. Maybe that's why I try to fill every faces even they are invisible, because I had bad experience with empty faces in invisible geometry.
(I think those problems were coming from duplicated vertices, maybe)
The problem is that when I'm making textures, invisible area still takes texture spaces even they are invisible, and sometimes it takes a lot, depending on size of the faces.
So theoretically, I can get more spaces to make more high resolution texture by riding off those invisible faces from UV.
Does it matter that remove invisible faces? Should I fill every face include invisible area?
Any advice will very appreciate it.