Todd, yes, you can have faults up to 100s of km-long move in earthquakes. But, the M7 1/2 1992 Landers quake broke about 5 different faults, jumping from one to the other (although generally linked directly I think). On some very small scale there would be more irregularity (I think). It also would depend on the rock type and type of fault and sense of movement. I've seen publications that faults with smaller total displacement (total movement since the fault was formed) will tend to be more irregular and smooth out with increased displacement.

Some of the dangerous segments of Califoria fault are locked, so may have very little seismicity until they break in big quakes. The southern Hayward (and Calaveras) fault creep, so should be illuminated by little quakes. The Hayward is dangerous because only part of the motion is taken up as creep. Not sure how that works physically...seems like there would be a problem with it sliding but also building up stress and strain.

Chris