Hi. I'm trying to add an optional named parameter to non_distributed_brandes_betweenness_centrality. I've managed before to do it in the brandes_betweenness_centrality function of the sequential BGL, but here the named parameters seem to be managed a little differently. Currently there is only one optional named parameter, weight_map, but I've seen that here, unlike the sequential version, it is not actually optional, as I cannot compile a call to the named parameter version of the function without providing that parameter. I'm attaching a piece of code as an example of what does and what does not compile. Is this behavior wrong? I suppose it is, but named parameters are kinda obscure for me, so any insight about it would be welcome. My intention is to copy the named parameter collection code from the sequential version, but I wanted to know any advice or comments anyone could provide. Thanks. -- Javier Dehesa