People have limited abilities to remember information. Hence breadth-first search is difficult without some external aid (e.g annotating a search tree). Depth-first search doesn't have this problem and it is a regular search strategy - hence cheap to do. However, in research on the way people actually do perform search the evidence points to a more opportunistic search pattern. People look for ways of cutting out useless parts of the search - a property machines have some difficulty in managing. Looking at typical human search patterns, they can often be seen as a kind of depth-first search with a tendency a) to explore only some of the options at a particular place in the search before pursuing the depth-first approach, b) to occasionally move ``sideways'', abandoning a particular subtree completely, c) not to go too deeply down any path, before exploring other options. Human search is somewhat opportunistic.