Re: [9fans] ISP filtering - update

[Lucio De Re <lucio@proxima.alt.za>]

On Fri, Sep 26, 2003 at 09:18:42PM +0200, boyd, rounin wrote:
>
> > The opposite may also work: try the MX for the destination first,
> > with a bit of luck you'll have a certificate to grant you access.
> > If not, you're stuck with your most accommodating ISP.
>
> say there are no MX's but the host you are mailing has multiple
> IP addresses and they might be forwarders or whatever.
>
Well, you could tighten up the rules, we are after all trying to
solve a problem.  But multiple MXs remain a problem, as you'd want
to talk to the exchanger that knows you and there is no guarantee
that all MXs have the same certificate list.  Part of the weakness
in RFC-821/2 lies in its fallback recovery features, which is really
where SPAM has hurt the most, forcing Internet services to be
trimmed.

It may be a philosophical point, but I'd still like to make it:
abuse has led to the loss of the socially desirable features of
the Internet.  Addressing the technological issues without restoring
the trust will not bring back the good ole days and that, at the
very minimum, is a shame.

> the whole thing is a nightmare.  it's like a jigsaw puzzle (or IKEA
> furniture) and the pieces don't fit.

I agree, but there are enough people interested that one can seek
and will probably find a solution.  It is a matter of critical mass
(what an ugly pun!) and a touch of unselfish development.

In fact, against my principles, I'd advocate prototyping something
that even half works at identifying the participants and spreading
it as far as it will reach.  Hard to do, but there are places like
slashdot and even google where one can broadcast the details and
presumably find a receptive audience (no comments required about
the neurological complexity of these collective brains!).

In my opinion, the biggest failing in RFC-821/2 lies with the
redundancy provided by the envelope and the message headers and
the contradictions these can lead to.  I believe this feature ought
to be the first thing to be eliminated, which is no small change.

I also can't help agreeing with Heinlein that a society that demands
identification documents is best abandoned, but I fail to see a
viable alternative.  Whichever way one turns, one needs to be able
to create a boundary between the perceived good and evil. Wasn't
that the first step in the evolution of life on Earth?

Just for the sake of stirring the pot, I've been considering an
anonymising mail server ever since anon.penet.fi (it's been such
a long time I can't even tell if I got the right name) was established
and recent thinking was along the lines of a public pool of messages
that could be indexed by the user's certificate hash value (I guess
Venti has had quite an impact on my thinking).  If the message is
encrypted, only the recipient would be able to decrypt it, more or
less minimising the likelihood that anyone else would want to
collect it.  I don't quite see that this would assist in our current
objective, but it may be worth something elsewhere in the development
of a different electronic mail exchange scheme.

++L