People often suggest that adding strong identification to the Internet will solve many security problems. Strong, useful identification isn't possible and wouldn't solve the security issue; trying to have it will create privacy problems.
The problem with security in cyberspace is that exploits target bugs in software to make them do things the designers didn't intend. Authoritarians have this fantasy that if they can just design a system that requires everybody to be tracked and tagged they'll be able to arrest those dirt bags who commit crimes on the Internet. One problem with this idea is the assumption that the identity system will be any less prone to software bugs then any other part of the infrastructure. It won't be, so it won't work. Another problem is the idea that you can actually manage an identity system for everybody in the entire world. You can't.
This, much like the ill considered efforts at "whois accuracy," will only serve to make it easier to target, arrest, or sue people who aren't intentionally out to commit crimes but for some reason run afowl of well heeled interests.
The secret is that some of the supporters of these systems know this, and thats exactly what they want. Other supporters don't care - they stand to benefit financially from these requirements regardless of how effective they are.
Bellovin RE: The National Strategy for Trusted Identities in Cyberspace