On Wed, Apr 09, 2003 at 11:04:58PM -0700, Chuq Von Rospach wrote:
I think this is a key issue -- if you send an e-mail to someone, I
think there has to be an assumption of consent to accept a response.
But OTOH, I'm not sure it implies a permanent willingness to accept a
communication relationship.
so perhaps there ought to be two forms of consent here: a one time use
key and a revocable consent key, something that can be embedded in a
header or in some other way in the return mail to unlock the door.
You're making it too hard. There are too many mailers. In fact, you want
to try every possibly means to accept replies to mails you actually wrote.
That starts with of course, putting a unique string in your outgoing
message-ids, time-encoding, and accepting anything with that in the in-reply-to.
Also good is whitelisting all mail targets of course, and remembering all
outgoing subject lines, and outgoing bodies (to see replies that include
text from your messages.)
Now you keep a different standard for mail you send to a list than mail you
send to ordinary people. You want to try _everything_ to accept replies to
mail you wrote to ordinary people. There is nothing more aggravating than
having somebody mail you, and you reply, and you get back a challenge or
frankly any autoresponder.
There is of course a very simple solution here. Use a different address
for your person to person mail, a totally unfiltered address. You just have
to be sure people don't put that address out in public and not use it to
mailing lists. If they do, you need another one.
I believe Zoemail is using a technique of a different address for each mail
out. That's overdoing it but can be useful with list mail.
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg