[hashcash] Re: PR Problem?

On Sun, Nov 12, 2006 at 06:09:11PM -0500, Eric S. Johansson wrote:

> BCC is a royal pain in the butt.  I took the brute force bloody

I can see the problem. But isn't the problem caused by hashing the
destination address, requiring a hash for each recipient?

What if--and feel free to say something if I'm missing something
obvious--hashcash was hashing something less variable such as the
messageID or the message body? Granted, it would mean that a stamp is
now per-message rather than per-recipient and therefore lowering it's
value in preventing long recipient lists, but the value wouldn't fall to
zero.

I'm not actually advocating this as a change, BTW. I'm simply suggesting
that the proof of work is only as valuable as the inbound filtering
makes it, and that's where the smarts need to be.

So, as a thought experiment, consider this:

    1. You hash the message body instead of the recipient list.
    2. You mint a stamp.
    3. The recipient MTA gets the email.
    4. The recipient's filtering pipeline considers the length of the
       recipient list.
    5. Starting at 20, the filter will require 1 additional bit for each
       5 recipients.
    6. If the recipient is the envelope recipient but not in the To: or
       CC: headers, it's a BCC, so require 22 bits plus 1 for each group
       of 5 listed recipients

I'm pulling the numbers out of my ear, and they should probably be
configurable, but the idea is to make the stamps per-message while still
requiring more work for multiple recipients.

I think this is workable from a practical standpoint because of the
nature of the problem to be solved. Yes, theoretically a spammer could
mint a single stamp, and then send out individualized messages, but that
would increase the overhead at the sending server, essentially
accomplishing the same thing as minting stamps (e.g. increasing the CPU
cost).

I'm sure someone can poke some holes in my logic. That's okay; I'm just
trying to consider the problem from all angles.

> This solves the BCC problem nicely but it does cause a mail list
> explosion.

Exactly. So, wouldn't just requiring a more-expensive single stamp solve
the problem in a more manageable way?

-- 
Unabashedly littering the information superhighway with detritus like
this for over 15 years now.

Other related posts: