Joaquín Mª López Muñoz wrote:
Hi Fernando,[snip...]
Besides, no one except you seems to be interested in multiindex_type :( If only some boosters jump into this discussion I'm sure many interesting insights would be provided.
I'm interested in this discussion but generally keep my mouth shut on the boost list since I'm still very much a newbie to boost stuff. I know that in the past I've needed a map with two indexes into it. I recognize the potential of your multi-index set for both my past application and also possibly for creating an SQL (or similar) interface.
From your descriptions in e-mail, I like the concept. I have not reveiewed the code in any form. I think that the data structure kind of blurs the distinction between sets and maps. Effectively, access by any given index is like either a map or multimap based on if the index points to unique entries or not. I think requiring that all index's point to unique entries will probably be problematic. I don't believe that is the case though... I would also be interested in ensuring that operations on a single index are efficient and don't result in massive overhead due to the presence of the other indecies.
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost