A brief note on the choice of separation character: The + was first
used, I believe, back in the mid-1980's as part of the Andrew Mail
system. We chose "+" after an exhaustive analysis of every ASCII
character for potential conflicts. It was one of a very small set of
characters that were suitable, i.e. invariant across a huge range of
email systems, Internet and non-Internet.
Since then, a few *new* systems have been built that have trouble with
the "+", so I sympathize with those who now prefer "-". I'm not sure I
remember properly, but I think the main objection to using "-" was that
quite a few mailing list names already used it, although there may have
been others. I remember that our second choice was "#".
The problem with ":" is that it can cause parsing ambiguities with
valid RFC 822 addresses. (In particular it was one of the "specials" in
the 822 grammar.)
I'm still partial to "+" but "-" is certainly fine for this too. --
Nathaniel
On Feb 13, 2004, at 4:24 PM, Ken Hirsch wrote:
From: "Seth Honeycutt" <gwest39(_at_)yahoo(_dot_)com>
This is a suggested possible idea for Internet Mail with RFC 822 /
RFC2822,etc. It is similar to "disposable e-mail addresses",
This is already widely used. See http://tmda.net/config-pre.html. It
was
discussed back in 1997 on this list whether it should be standardized.
See
http://www.imc.org/ietf-822/mail-archive/mail8.html for discussion of
"subaddressing".