A classic problem with end-to-end encryption backdoors is that most of the time there is no benefit to users or software vendors to offset the cost and risks th

A classic problem with end-to-end encryption backdoors is that most of the time there is no benefit to users or software vendors to offset the cost and risks they must take on to provide law enforcement access. I can think of just one exception, which is identity based encryption, which allows e.g. email addresses to be used as public keys. IBE requires a master secret key that is used to generate the private keys associated with each user's ID, and therefore has an inherent backdoor, but at least the users can benefit from the reduced complexity of key management. For the most part backdoors require some kind of central coordination, so there are situations where it is basically impossible to support one. HTTPS is a good example: anyone can run a web server and there is no permission or coordination needed to do so, and so HTTPS backdoors would be difficult to create (otherwise I cannot run my web server without coordinating with whatever authority controls the backdoor, or I cannot run a secure web server without such coordination). That is what the conversation is really about: whether or not we are willing to sacrifice the most useful and valuable aspects of the Internet's design (distributed, loosely coordinated, no-permission-needed), or do without security and hope for the best, for the sake of a law enforcement backdoor.