This should be starting point of discussions about proper indexed_set naming. indexed_set name was picked long before the library gained much more functionality. Instead of changing name immediatelly it was decided to wait on opinion of others (any change requires quite a lot of work). Here are my suggestion of names, with rationale: 1. the data structure should be named "multi_container" and should reside (or be lifted) in boost namespace. Rationale: it is superset of std::set/multiset/list and possibly slist and deque). Therefore the "multi". The structure is important and will be probably unique - therefore the boost namespace. 2. library specific names that won't fit into boost namespace should be in "multiindex" namespace. Rationale: it will be used mainly for index related names. The name is aso reasonably short. 3. indexes should be named: set_like, multiset_like, list_like, slist_like, deque_like. Rationale: it gives idea to user what it is about (more than e.g. sequenced). 4. Since indexed_set::unique may clash with std::unique I suggest to use words "distinctive", "non_distinctive". Rationale: I found it in a dictionary as the only one word synonum to unique. English native comment welcomed. /Pavel