As a related issue, would it make sense to also review the hash_map/hash_set implementations at the same time (assuming Daniel et al can finish it in time)?
Sorry it took me a while to answer this. I haven't done much on these recently. I can get the implementation and tests ready in time, but I haven't done much in the way of documentation. I suppose if this is a part of Boost.TR1 that won't matter so much. I think this should be up to John? It seems a little late to be adding things.
I don't mind the hash tables being reviewed at the same time, but without documentation (which is what most folks review), it'll be harder for people to get to grips with it. Do you have a separate review scheduled yet? I don't particularly see any problem with extending TR1 to adapt to new Boost submissions that come along in the future. John.