Information foraging models of browsers for very large document spaces