| Question asked
|
One of the main points of this paper is th … One of the main points of this paper is that nested regular expressions can be evaluated efficiently, in linear time w.r.t. the size of the graph and to the length of the expression. From a practical point of view, though, is this really efficient? One of the nice things about traversal-based algorithms is that, depending on the topology of the graph, their performance may be more or less independent the actual size of the graph. The proposed evaluation strategy, in contrast, appears to require a massive global calculation for each step in a path expression, e.g. upwards of one billion operations for a simple one-step path from a node to its immediate neighbors in a large, billion-node graph. neighbors in a large, billion-node graph.
|
