Hi All, I have a newbie Boost user question. I need to store a large graph and then repeatedly break it up into sub-graphs based on an algorithm. I also need to store properties about each vertex (name) and edge (how many times the edge was traversed in each direction, hence a directed graph). Currently, I am creating a directed graph to store the full data set using custom and built-in internal property maps to store the data. In addition, I created a undirected graph to store subgraphs. It too has internal property maps using vertex_index_t and edge_index_t, but they will simply reference vertexes and edges in the first (full) graph. Since there will be many of these sub-graphs, I didn't think it was a good idea to be copying data over and over. Anyone have any suggestions or see anything wrong with this approach? Thanks, -d