I think the term "trapdoor" in this context merely refers to known characteristics of the Clipper chip and does not imply any additional abilities to decrypt messages other than what has been revealed. A "trapdoor" is generally some secret which allows you to decrypt messages which could not be decrypted without the secret. It is normally applied in the context of one-way functions but I suppose an encryption system could be thought of as one-way. In this case I think the trapdoor may just refer to the Law Enforcement field which, upon decryption, allows information to be recovered so that the message can be read without the user's cooperation. The comment about the trapdoor issue being independent of the escrow issue harked back to Denning's original suggestion, which was that all users of encryption would have to escrow their keys with the government. This could be done in the context of any cryptosystem, such as DES or RSA. The trapdoor makes the system more convenient for users, hence perhaps more acceptable. But the escrow provision is the aspect that the government really cares about. BTW, there were some comments on sci.crypt about how this review document showed that the government was not as opposed to our views as had been thought. On the contrary, my take was that this document reflected something of a maverick view, one which was not politically acceptable with those in authority. Note the critical comments about the document containing unsupported assertions, a common buzzword for "views I don't agree with." I don't think the author of this review has helped his career. Hal Finney hfinney@shell.portal.com