[cryptography] [Cryptography] the introduction problem, was prism proof email, namespaces, and anonymity

John Levine johnl at iecc.com
Sat Sep 14 14:05:48 EDT 2013

Hash: SHA1

>   Jabber, Facebook and other services where all or essentially all
>   communications require a bi-directional decision to enable messages
>   for years now, and there is virtually no spam in such systems
>   because of it. So, require such bi-directional "friending" within
>   our postulated new messaging network -- authentication is handled
>   by the public keys of course. 

This is an old approach, trying to reduce the spam problem to the
introduction problem.  It works, sort of, but it's not as simple as it

I know people who do security at Facebook, and the lack of spam is due
more to the fact that it's a closed system with people whose job it is
to keep spammers from annoying the customers than to the introduction
aspect.  For Jabber, I expect it's that other than gmail (which has
its own security department) there aren't any Jabber networks large
enough to be worth spamming.  There's plenty of spam in AOL's instant
messaging system, where you can send anyone one message asking to be

Introductions have terrible scaling properties.  If you want a
messaging system that can do what email does, it needs to be able to
handle mail sent by robots.  For example, when you buy a plane ticket
online, you probably want to let the airline send you a confirmation,
and also updates if the flights change.  How do you authorize that,
short of allowing anyone to send you a request that you look at?  And
more importantly, how do you tell the flight updates from valuable
offers sent by the same company?  Or, remarkably often, an
organization's contact list will leak (it's hard to tell whether by
malicious employees, incompetence, or malware) and now you have to
abandon the existing token and set up a replacement.

This isn't a useless technique, and it's very useful for some
situations like small children whose parents manage their list of
correspondents, but I don't think you'll find it a very useful way to
keep out unwanted messages in general, unless you're also willing to
lose a lot of wanted ones.

Version: GnuPG v2.0.21 (FreeBSD)


More information about the cryptography mailing list