On Wed, 30 Jul 1997, John C Klensin wrote:
I think you are right, and that it is probably worth giving
up the availability "+" for anything else to get where we
need to be. But, logically speaking, most of your argument
evaporates if either of two courses is taken:
I'm much less convinced. I use subaddresses for non-mailing list things.
I believe they're generally useful regardless of alternative solutions to
the list subscription/restriction problems.
If this logic is correct, can you make the case that it is
worth a possibly-significant risk to the infrastructure
(yes, there really are addresses -- especially in mappings
to and from strange gateways to LAN mail systems -- in
which "+" is a delimiter character, but not for
subaddressing, and several of them may appear in a
local-part) for what may be a short-term gain?
More than one deployed product already manages subaddresses as the spec
mentions and I've heard of no infrastructure problems other than the
occasional product which rejects all addresses containing "+". Of
course such products are clearly in violation of RFC822 already.
I also don't view this as a short term gain.