Richard Reynolds wrote:
On Sat, 24 May 2008, Andrew Lentvorski wrote:

Every single challenge email I have ever seen now gets marked as SPAM. This caused me a *huge* nightmare with three different companies that I was caught in between.
then your system is broken. but your not saying "i cant fix my system" your saying "this type of system cant work" those are different

No, I'm saying that a system which I *do not control* is capable of causing huge amounts of breakage *for me*. I have had to personally forward emails between three different parties because their respective challenge response systems wouldn't let the challenges through.

And this is invariably what happens when challenge/response systems come in contact with one another. The main reason why this doesn't happen more often is that once you implement challenge/response for a large enough block of people, adminning the stupid thing eventually becomes too much work and it gets ripped out. Fortunately.

Yes, in theory, challenge response systems can be set up perfectly, maintain lists so they don't DDoS systems, create mutable challenges that won't be cloned by spammers, etc.

And, BTW, this *still* doesn't work if there are enough challenge/response systems because then you use 10,000 of these systems to bombard the poor victim rather than just one. Any form of unverified backscatter is simply broken. Period. Fortunately, challenge/response systems are few and far between.

In reality, the spammers have more incentive to do nasty things than the good folks have to keep up with it. Once any system hits critical mass, the spammers figure out how to take advantage of it. See the current battle between Craigslist and spammers, for example.

Even worse, in real reality, the challenge/response system is being run by some random sysadmin who doesn't know jack about it and is afraid to touch it. Thus, you wind up with dueling challenge response systems and two intransigent sysadmins who are determined not to be the one who has to make the change.

Lots of things are useful, in theory. In theory, there is no difference between practice and theory. In practice, there is.

-a


--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to