Yet another badly-broken challenge-response mail system

[Dave Farber <dave () farber net>]

> Date: Fri, 01 Aug 2003 08:02:00 -0400
> From: Rich Kulawiec <rsk () gsp org>
> Subject: Yet another badly-broken challenge-response mail system
> To: Dave Farber <dave () farber net>
>
>
> ----- Forwarded message from Ted Dolotta <Ted () Dolotta ORG> -----
>
> > Thank you for sending me your email with the subject "Re: [IP] An 
> interesting
> > perspective on the latest DARPA brouhaha". I really want to receive 
> your email.
> > In an effort to eliminate junk email, I am using MailFrontier Matador.
> > Matador has placed your message on hold.
> >
> > Please click the link below so you will be added to my Allowed people list,
> > I will receive your email, and we will be able to communicate freely going
> > forward.
> >
> > <http://c.mailfrontier.net/XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX>
>
> This presumes that I am
>         a) on a TCP/IP network
>         b) not behind a firewall which blocks HTTP
>         c) on a computer with a web browser
>         d) willing to start up a web browser just to have my mail delivered
>         e) using a web browser that works with their site -- any bets on
>                 a text-only cookie-rejecting Javascript-disabled browser?
>         f) willing to verify for mailfrontier.net, a third party,
>                 that this is a known-working mail address
>         g) not blocking mail from the third party (as opposed to
>                 the person I was sending mail to)
>         h) willing to do this jump-through-the-hoop exercise (with
>                 individual variations, of course) every time I send
>                 mail to someone
>
> I *probably* could subvert this by forging Dave's address or the
> address from which IP is sent into the headers, but I'm not going
> to do that.  Instead, I'm going to drop mailfrontier.net into the
> local blocklist, permanently, so that I am no longer asked to
> jump through hoops merely to send a mail message.
>
> The irony of all this is that a simple, low-cost anti-spam system built
> from open-source software and databases -- one which should already
> be in place on all competently-operated mail servers -- is capable of
> blocking upwards of 90% of spam with a false positive rate ranging
> tenths to hundredths of a percent AND simple mechanisms [which use
> mail themselves] for those accidently caught to report the problem
> and get it fixed.  It works; it works efficiently; and it doesn't
> require that senders/recipients use anything other than the mail
> client that they are obviously already using.
>
> But *this* system has a 100% false-positive rate, until corrections
> are incrementally applied by correspondents.  It thus shifts the work
> of achieving accurate results (low false positive and low false negative
> rates) to everyone BUT the vendor who's presumably the one being paid to
> perform the task of screening mail!  Even so, application of thousands
> of corrections would still leave it with a 99.999% false positive rate --
> which is ludicrous.
>
> ---Rsk

-------------------------------------
You are subscribed as interesting-people () lists elistx com
To manage your subscription, go to
 http://v2.listbox.com/member/?listname=ip

Archives at: http://www.interesting-people.org/archives/interesting-people/