On 7 Oct 2008 at 20:38, James Knott wrote:
....
 Isn't it necessary to impersonate the address that the mail is
> originally sent to, in order to request unsubscribe?  Otherwise, we're
> back to the situation where anybody could unsubscribe anyone.

Don't forget unsubbing is a two-step process.

The original request needs to identify /who/ is to be unsubbed -- 
this comes either from the 'from' mail header or by cutting up the 
'=' form of the 'to' unsub address.

The request needs authenticating because it's trivial to forge such a 
request.

OOO does this by sending to the I-want-to-be-unsubbed address a magic 
cookie. It assumes that whoever finally receives this magic cookie is 
a legitimate user of that address, and so anyone who can quote that 
cookie is allowed to confirm the unsubscription request.

If a malicious unsub request is sent in, the attacker won't 
ordinarily receive the cookie, so cannot complete the request. 
Instead the legit email subscriber will get it, wonder what's going 
on (and hopefully take action!!)

Quite reasonable, as it covers a situation where an address is 
forwarded. It does mean a malicious mail admin (or anyone with 
similar access  /could/ intercept mail and play daft tricks. I don't 
believe that's a significant issue though.





-- 
Permission for this mail to be processed by any third party in 
connection
with marketing or advertising purposes is hereby explicitly denied.
http://www.scottsonline.org.uk lists incoming sites blocked because 
of spam
[EMAIL PROTECTED]    Mike Scott, Harlow, Essex, England



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to