[ietf-dkim] BCC Recipients

[Hallam-Baker, Phillip]

> -----Original Message-----
> From: william(at)elan.net [mailto:william at elan.net] 
> Sent: Tuesday, August 23, 2005 3:33 PM
> To: Hallam-Baker, Phillip
> Cc: ietf-dkim at mipassoc.org
> Subject: Re: [ietf-dkim] BCC Recipients
> 
> 
> 
> On Tue, 23 Aug 2005, Hallam-Baker, Phillip wrote:
> 
> >> This doesn't help for BCC recipients at the same domain.
> >
> > The only way to sign BCC in my view is to provide a per 
> user signature 
> > constructed by means of an HMAC.
> >
> > For example message is "Hello World", Sending it to dot at dotat.at
> >
> > So I construct a BCC identifier HMAC ("dot at dotat.at", SHA1("Hello
> > World"))
> >
> > Or something of that nature. That means that the BCC recipient can 
> > verify it was sent to them while preventing any To: or CC: 
> recipient 
> > knowing anything more than that there is a BCC.
> 
> While its cool idea, I fear it may not be 100% doable because 
> when message 
> is sent to bcc recipient, the address originally in bcc (and 
> which becomes address in 2821 RCPTTO) may not be final 
> address seen in RCPTO when message is delivered (i.e. if 
> message is further forwarded for example).
> 
> BTW - why HMAC? You could do just SHA1("dog at dotat.at","Hello World")

All you need is a salt that is message dependent. HMAC is better from a
purist point of view.