On 16 September 2015 at 08:02, Peter Gutmann <pgut...@cs.auckland.ac.nz> wrote: >>HTTP-2 did this kind of thing, and IIRC are the first to do so. > > Some PKI standards have done it too, but mostly because the base standard was > such a mess that you needed a profile just to sort out what needed to be > implemented for anything to work (for some level of "work"). They're such a > design counterexample that I didn't want to mention them in my original > message :-).
Yes. I wouldn't recommend following this path to others; it's not easy and the return on that investment isn't all good. The mess we were attempting to clean up with HTTP/2 was the state of TLS deployment on the web, not so much the spec itself. _______________________________________________ TLS mailing list TLS@ietf.org https://www.ietf.org/mailman/listinfo/tls
