g'day,
stanislav shalunov wrote:
...
> The second proposal ... assigns a
> problem for the sender to solve before the message is delivered. The
> problem must have predictable complexity that the sender must be able
> to estimate before solving the problem. The sender can then spend the
> resources to solve the problem to have the message delivered or simply
> abandon the attempt.
>
> A reasonable problem could be, perhaps, to find a string that has MD5
> sum that has first n bits equal to a given string.
Cool, so we could hook the SPAMers up to the SETI project.
Alternatively, you could make one humongous chess position
evaluation system, or perhaps work on a few more digits of
PI. Heck, you could even let senders negotiate which "good
cause" they'd like to sponsor with their CPU cycles. The
marketing folks at the ISPs could then use it as a
differentiator - "10mS of every mail transfer from our
systems goes to support the mapping of the Human Genome..."
;-)
Seriously, the concept here seems to be "impose a cost onto
the SPAMer to make him/her pick up the tab for his/her
actions". Put in these terms, there would actually be
something for the IETF to work on here at the protocol
level. You could certainly imagine extensions to allow a
recipient to specify that, like the Knights at the Bridge,
they must first "answer these questions three" and only
accept the mail if the sender agrees to negotiate the
challenge/response algorithm and exchanges the challange
data and the response.
Technically this seems feasible, although you immediately
run up against network effects. Until enough folks agree to
install the software that handles the extensions (i.e.
agrees to pay an additional cost to send mail) all you'd be
doing is cutting yourself off from the net. And of course,
hanging out a red flag for the SPAM community, who will
immediately seek new ways to circumvent/lower the cost. I've
seen it posted in other contexts, but remember the old saw
"Build a better mouse trap and all you get are smarter
mice..."
Still, I for one think it's an interesting idea.
Deliberately seek to raise immediate costs, to avoid longer
term pain. We could promote it the way they sold pollution
controls for the past 30 or 40 years...
- peterd
---------------------------------------------------------------------
Peter Deutsch pdeutsch@gydig.com
Gydig Software
The one who says it cannot be done should never
interrupt
the one who is doing it....
- The Roman Rule
---------------------------------------------------------------------