> complicated-to-implement and largely ineffective solution such as subverting draft-rhrd-tls-tls13-visibility for improper purposes? The phrase “subverting for improper purposes” is inaccurate, and perhaps misleading. We would be providing another cleartext signal and we have to expect that someone will use it; anything else would be naïve. All the discussions about “ossification” in the past two years should make that painfully obvious.
As for “improper purposes,” it’s something enabled by the protocol, even if it’s not what the proposers intended it for. But isn’t the whole story of the Internet? If the purposes are really improper, define things in a way that it can only be used properly, for whatever that definition is. So far, we’ve seen that it can be used to segregate clients, open them up to stream modification, and the only justification has been that this is perceived easier to keep visibility as currently built by sharing static RSA keys. Do I have that right? _______________________________________________ TLS mailing list TLS@ietf.org https://www.ietf.org/mailman/listinfo/tls