On Thu, Aug 20, 2009 at 12:13 AM, Chris Wj<address@hidden
> What do you guys think about the idea of chaining operations that leverage
> common calculations. Say, for example, that you had 2 algorithms that
> required shortest paths to be calculated before continuing, if you run them
> both from scratch, they both calculate shortest paths first, then thier own
> calculations. So shortest paths gets computed twice (and that is an
> expensive calculation!). So what if you could either chain operations, or
> hold on to the intermediate calculation until the graph structure actually
> changed ("dirty bit").
> Also, what algorithms are similar enough to one another that we could avoid
> duplicate work by performing them simultaneously?