Although trained in geophysics and rock mechanics, gf had to tackle earthquakes early in his career. He had created a computer program that simulated underground dynamic motion, but he needed inputs. As well, he realized that the only serious damage that a seismic wave could inflict on a tunnel was induced rockbursts. By playing with this model, he could simulate rockbursts by manipulating a small number of variables. The law of scale invariance led to the simulation of earthquakes. This was important, since he needed to nail down the amount of stress disturbance that an earthquake could generate. This, in turn, gave him the degree of sensitivity required to trigger a rock failure in an underground excavation.
The search for physical limits on these key variables led him to research the effects of major earthquakes around the world. He, then, formulated his thoughts in a number of blog posts.
Seismic Boring Story - his first post on seismic waves
Fault rupture 1 - starting with the actual grain adhesion properties, a key variable
Fault rupture 2
Fault mechanics 1 - going into dynamic friction and self-similarity
Fault mechanics 2 - stressing the importance of a continuous mechanism, or everything locks up. Most people think these things lock up.
In summary, east and west earthquakes differ only in the type of external force. The plate margin earthquakes have a following force of plate movement, which always rejuvinates the mechanism, and eventually forms large slip, or a mountain range. The interior earthquakes run on stress relief (from a highly stress-charged continent). The interior zones rely on water (stress corrosion), start small, and eventually grow to considerable size.
The latest twist to the earthquake story is the 'Fist of God' earthquake, also known as a super-quake. This happens when the stress regime is fairly homogeneous, the critical displacement is low, and difference between dynamic and static friction is high. At that point, a wave can travel up the fault plane at the speed of a P-wave, but with all the PGV of a shear body wave. For both strike-slip, and thrust faults, this results in a directed wave with a strong impact.
Recently, some of his ideas have gone into practice. They started injecting in Arkansas, and formed a large earthquake zone. That eventually stopped, but the 2011 Virginia Earthquake was most likely induced by a dammed lake. Too many posts were written on this, and are best found by searching.