This experimental data is small scale. We have always more or less known that "rough" edges (as I recall we termed them asperities or something similar?) could cause differential movement along faults. Years ago, I read that the subduction of seamounts might have influence on major quakes. The influence (as Chris says below) of temperature, water content, density is also a concern. I think that the headlines regarding this for quake prediction are a bit overdone. But all knowledge is useful!