I just wanted to finish this up, since right now I'm dying with nothing to write about, even though I did get a few clicks the other day.
Mexico 1985 had a few lessons, right and wrong.
It reinforced the traditional sinusoid, shake table view. The geometric basin amplified by a factor of ten, and the soft soils amplified by another factor of ten. Outside the basin, the ground motion was barely felt. The earthquake destroyed some really crappy mafia-concrete buildings, in a sinusoid manner. The first ground motions cracked all the concrete, and softened the structures, so that they resonated at a very low frequency. This hooked in exactly with the soft soil motions. Destruction!
Any decent structure survived. These buildings did not crack on the first motions, and did not hook into the very low frequencies.
As well, at the epicentre, not a speck of piping was damaged! Piping is never damaged! But do you know how many billions go into seismic piping analysis, and the horrible consequences of that phony analysis? Yes, the old company is full of it.
In the end, the wrong lessons dominated. The piping seismic engineers did not want to put themselves out of work. The structural engineers continued with their shake tables, energy absorbers, modal analysis, etc. The world went on.