Thorsten Ottosen writes:
Don't worry; I'll propose this as part óf my submission of boost.range.
Please be sure to coordinate with other people who worked/are working on the topic, then. See http://article.gmane.org/gmane.comp.lib.boost.devel/120408/, for instance. This is way too important to allow a false start, IMO.
Well, I'm not proposing new iterator concepts, although David and Dietmar might be.
FWIW, the cited posting is not about "new iterator concepts". It discusses building a proper algorithm dispatch scheme, which is central to any library that introduces overloading of algorithm names.
That said, Ranges pretty much follow whatever iterator categories that exists.
I won't have time to look into this stuff for some time. If you feel its important, go do the work.
Please read more carefully what I said. It's important to get this right, not to get it into standard ASAP. -- Aleksey Gurtovoy MetaCommunications Engineering