greetings all, long time user and lurker, low probability poster... i've been mulling over an SoC proposal for Boost.Graph based on work i've been doing over the past year related to measures on random graphs. specifically, i've been building up a set of graph algorithms using boost graph to compute (for example) degree distributions, clustering coefficient, mean geodesic distance and diameter, clique counts, etc. i had also noticed on the wiki that there were some proposals for fixing documentation, examples, and even a boost graph 2 library. part of my intended proposal was to spend some time working on these issues like moving the documentation to Boost.Book (if it hasn't been done) and improving it, making the examples a little bit more readable and commented, basic housekeeping type stuff. one of the other ideas i've been kicking around was how to rewrite Boost.Graph using Boost.MPL. i had intended part of the deliverables to be a proposed redesign and evaluation. this would address issues like basic graph wrappers (directed, undirected), hypergraphs (edges with more than 2 or more endpoints), etc. quick background... i'm a phd student at kent state doing research in software engineering. the reason i'm proposing work on Boost.Graph is because i've been using it for class projects to measure features of massive software graphs derived from type references. so... if anybody thinks this project is worthwhile and would be interested in mentoring, i'd like to know. thanks, andrew sutton asutton@cs.kent.edu