On Mon, 1 Mar 2004 14:29, Doug Royer wrote:
- A mini protocol between the MUA and the opt-out sites that allow
a user to opt-out in a standard way.
Even better, IMO, would be the following.
* Users do not want to "un-subscribe" from something that they never
subscribed to in the first place. The default policy should be for
subscription and un-subscription to be entirely under the control of the
recipient, such that no third party can "subscribe" or "un-subscribe" them to
This is, perhaps, a re-formulation of the "uniform interface to mailing lists"
goal in conjunction with various "recipient control" goals. Note that the
above does not recognise any form of spamming as a subscription: it only
recognises recipient-initiated requests for mail as subscriptions. This
doesn't exactly cover "opt out", which tacitly assumes a right for a sender
to initiate a subscription. For those, we have the following requirement.
* Users want to be able to trivially blacklist any given sender.
This is more effective than an "opt-out", because opt out requires sender
cooperation. It's still not entirely effective, though, because any given
person can create an unlimited number of sender identities from which to
spam. No sooner do you block amazingoffers.biz, than you start to get
remarkably similar spam from bestdeals.info, or something like that.
Neither of these options preclude a formalised opt-out mechanism, but opt-out
isn't necessarily something that we want to encourage, since it can
facilitate abuse. Even if a spammer honours an opt-out request (on all
current and future spamming runs), this does not prevent him selling the
address to other spammers as a "confirmed active" address. This is
The best summary of the current suggestions with regards to recipient control
of messages is currently at the following URI.
James Seng's Wiki is more up-to-date, but not as conveniently organised so far
as this particular subject is concerned.