On 10/30/06, Jason Hise <0xchaos@gmail.com> wrote:
On 10/30/06, Michael Fawcett <michael.fawcett@gmail.com> wrote:
How about ra_sequence?
Random access sequence sounds like the definition of vector more than a new container type. Actually, I wonder if it is even valid to say that this container supports random access... although it supports the syntax for it, isn't random access defined as being constant time?
I think I read that somewhere too.
Perhaps it would be best to leave the ultimate decision up to the library author, who came up with the idea and may want the privlege of choosing a name for it. If he would prefer to leave the decision to group concensus, we could put the name to a vote.
It seemed like the author was asking for suggestions. I can't speak definitively for the others, but I don't think anyone was presuming to name the library for him.
Oh, sorry. I was actually asking for help in the naming issue (too).
[..] focusing on reading through and commenting on his implementation is probably a higher priority, [..]
Yes, I'd love to read your comments about both, interface and implementation. I still have not written a separate documentation, but you will find _lots_ of comments (really). So, what should I change in first place? ;-) Back to the naming issue, and more important: another precedent of the idea in boost, "order-statistics trees" together with "ranked indices" are the names used in the "future work" section of multi_index documentation: http://www.boost.org/libs/multi_index/doc/future_work.html#ranked_indices
From there: http://pine.cs.yale.edu/pinewiki/OrderStatisticsTree
Though, I can't see how is it related to "Order statistic" (I'm not a mathematician): http://en.wikipedia.org/wiki/Order_statistics#Computing_order_statistics Thanks in advance, Martin