jcaron - 11:05 am on Sep 28, 2010 (gmt 0)
I think the idea is not so much to be able to decrypt the communications, but rather to make sure there is a point where the communications are not encrypted, and that wiretaps could happen at that point.
I.e. there would be no end-to-end encryption, but rather two encrypted client-server communications, and the wiretaps could happen at the server, where the communication is decrypted before being sent on to the other endpoint.
This is actually already what happens in many (most?) communication systems that use encryption.
Now this would not necessarily prevent the use of encryption by the users directly and from end to end, though this is of course not necessarily applicable to all services (the service must be able to carry random digital payload unaltered), and of course it would depend a lot on the exact wording of the law (e.g. if it includes providers of encryption software).