Currently there are two mutable_queue implementations in Boost. One in pending (not really official) and another one in the newly released Heap library. Currently, some of our code currently using mutable_queue from pending. Has somebody already switched between both and found performance differences or can I expect them to perform nearly the same? Will mutable_queue in pending be removed since there is now an official library? Cheers, Philipp Moeller GeometryFactory