Predicting earthquakes along the major plate tectonic boundaries in the Pacific

In an article in the last issue of the Earthquake Information Bulletin ("Earthquakes and Plate Tectonics," by Henry Spall), we saw how 90 percent of the world's earthquakes occur at the margins of the Earth's major crustal plates. however, when we look at the distribution of earthquakes in detail, we see that a number of nearly aseismic regions, or seismic gaps, can be found along the present-day plate boundaries. Why is this? And can we regard these areas as being more likely to be the sites for future larger earthquakes than those segments of the plate boundaries that have ruptured recently.