From: Dave Abrahams (abrahams_at_[hidden])
Date: 2000-01-12 19:59:33
> Hence the conditional requirement for member access that is now in the
> Collection concept.
> I do not know of any examples that argue for the more
> flexible position of unconditionally allowing proxies.
> If anyone does, please let me know.
I think the point here is that your graph algorithms require something
stronger than Collection ("2DCollection"?), such that at least begin() and
end() are reflected by the reference type. I'd hate to rule out things that
could reasonably be sorted just because you can't use them for graph
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk