Anisotropy literally means "not being the same when turned". It describes an environment in which things do not look the same in all directions.
2004-12-01
Lego started out as a set for building houses. Bricks have a top and bottom. There are knobs on the top part. The knobs are arranged in a square pattern. Stacking bricks vertically leads to a construction that can be described in three orthogonal axes: an x and z axis perpendicular to each other in the horizontal plane and a y axis pointing vertically up.
 |
Along X and Z we see knobs, the distance between knobs is the same in X as in Z: they form a regular grid of squares. However, in the Y axis two things are different: first, there are no knobs; second, the "step" is not the same. The distance between bricks vertically is 9.6mm whereas the difference between knobs in X and Z is 8mm. That makes for better looking bricks, but in a mechanical construction it is a pure nuisance that the vertical pitch is different from the two horizontal ones. It's like having a pixel bitmap with rectangular pixels. |
So in fact we face two problems: not the same quantum steps, and not the same way of attaching things to each other. Both differences produce significant problems when designing models. There is no good solution for the knobs problem, but it is possible to imagine a Lego variant in which the step in all directions would be 8mm.
Fortunately, we are saved by the existence of 1/3 high plates and the fact that
9.6mm = 3 x 3.2mm and 5 x 3.2mm = 16mm = 2 x 8mm
So with the aid of plates, which are 1/3 vertical step, or 3.2mm thick, we can get in step at some multiples of brick heights. See Bracing.
While it is true that Meccano does not have an anisotropy problem (it is the same in all three axes), it has another pretty bad problem: where parts are fixed to each other by nuts and bolts, the thickness of the metal strips adds up. Conceptually Meccano models ignore this thickness, and often it is not more than one or two layers away from the ideal, yet it can lead to severe problems in smaller, complicated models. Yet, look at this...
next problem
