Yes, there were ways to avoid it (shared vertex calculations), but the effect seen was mostly where there are two separate vertices that are supposed to be at t

Yes, there were ways to avoid it (shared vertex calculations), but the effect seen was mostly where there are two separate vertices that are supposed to be at the same position, aren't, because each vertex was positioned with different transformation matrices.