On Thu, Jul 28, 2022 at 10:07:09PM +0200, Carsten Bormann wrote: > On 2022-07-28, at 22:02, Ilari Liusvaara <[email protected]> wrote: > > > > I don't think the references are good to go before 2024. And I would not > > be surprised if that turns out to be 2025… > > That could give us a five-digit RFC number! (RFC 100xx) :-) > > I do think there is a lot of value in having a long-lived I-D first, so that > people can run experiments without squatting on code points and guessing on > the best structure to represent something.
Sure, not squatting on codepoints is worthwhile. But it also pays to put some thought in in advance about what to do when the references go and change the protocol parameters. If you can't enumerage all implementations and ensure they are either updated or gone by some deadline, you probably have to get a new codepoint. Which is not so terrible (see TLS ExtensionType Value 53), but good to make sure everyone is on the same page before we get there. -Ben _______________________________________________ COSE mailing list [email protected] https://www.ietf.org/mailman/listinfo/cose
