On Nov 2, 2024, at 21:54, John Levine <johnl@xxxxxxxxx> wrote: > > I can imagine situations on private networks where it might be useful. Despite > what some other people (not you) keep asserting, we do not know all of the nooks > and crannies where devices might be using features like TURN in ways that make > sense. So my inclination would be to change nothing, or at most a short note Not having an opinion on TURN myself, I find this reasoning odd. If we would apply this reasoning to TLS, we’d still insist people implement SSLv2, SSLv3, TLS 1.0 and TLS 1.1. Because who knows what devices people still run anywhere. It basically means you can never obsolete anything. I heard TURN is not implemented in any opensource SMTP server. I heard it is insecure. We haven’t heard anyone step forward to say they require TURN. If we can’t kill off things that aren’t visibly implemented nor visibly deployed, when can we kill it from a spec ? I don’t think that line is “I could imagine a use case”. Paul -- last-call mailing list -- last-call@xxxxxxxx To unsubscribe send an email to last-call-leave@xxxxxxxx
