My primary objective with these three investigations is visibility. By using depth you can eliminate views through the walls at steep angles. With the cylinders, each piece of the wall can be angled so that it is more favorable to one direction of sight.
With the cone wall, every tile can be varied differently (independently) than the neighboring one. This will allow for varying view angles throughout the wall. I tried to make this more clear in the pdf attached to the end of the post.
The triangular tile idea was similar. It was my intention to make double sided triangles, and the visibility would be dependent on the position of the point that extends from the wall. The further the point, the less willing the wall would be to let you see through it at a steep angle. Yet, if all the point are extruded as well as shifted in direction, they can have the same effect as the cone wall, but the primary viewing aperture would be the spaces between the tiles, where the cones would utilize the hole (or absence of a peak)