On 14 Jan 2014 at 17:04, Michael Conlen wrote:

I recently discovered that there seems to be a lack of good Prefix Tries in C++. There are many basic ones and I think a C implementation ran all of 12 lines of code, but something with a nice interface and iterator seemed lacking. (Stop me if I'm missing something here)

GSoC 2013 implemented a Prefix Trie container. I was one of the mentors. AFAIK it was not peer review ready last time I looked.

So I'm interested in developing Prefix (and possibly a Radix) Trie containers in C++. I could see a use for both sequence and map containers; however this would be a learning experience for me. If there's interest in such a creature I'm happy to go through the review process as necessary.

Having said that, is there interest in having such a library and if so what are the pointers on what would be expected (particularly any boost-isms)?

Almost certainly you want to implement it via Boost.Intrusive, and from there into a concrete STL type implementation. I would not underestimate the difficulty of getting an implementation up to peer review quality. Still, if you're game, you'll get a lot of moral support here. You're right that C++ needs that container, it's one of the final classes of STL container C++ is still missing. Niall -- Currently unemployed and looking for work. Work Portfolio: http://careers.stackoverflow.com/nialldouglas/