Reply to post:

I think the frustrating thing is that _if_ a company offers email service (supporting, say, joe@foo.com), the incremental cost of adding an additional domain (say, for fred@bar.com) is effectively zero (when amortized over a few dozen users). Sure, the capabilities of the 'new' implementation may not match those of the 'old' one, but that's a separate issue from whether fred@bar.com stops working for incoming mail...