- From: Martin Thomson <mt@lowentropy.net>
- Date: Wed, 15 Nov 2023 09:59:46 +1100
- To: "David Benjamin" <davidben@chromium.org>, "Ben Schwartz" <bemasc@meta.com>
- Cc: "Ilari Liusvaara" <ilariliusvaara@welho.com>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
On Wed, Nov 15, 2023, at 08:30, David Benjamin wrote: > I also think this allergy towards minting ALPNs is bad for the > evolvability of our protocols. Using the pejorative "allergy" isn't helpful here. Ben is simply pointing out the non-trivial cost of minting a new token, which we have to carefully weigh against the benefits. Right now, it's quite possible that a breaking change in HTTP/2 -- which is what an ALPN change implies -- could be interesting, but the more we do, the harder it is to justify doing the work. If we can't agree that deploying a very minor update like the one described in the draft under discussion is feasible, then I don't hold out high hopes for a breaking change with a suite of changes. Obviously, if you think that you have to have a new ALPN token in order to break the 0.5 RTT impasse, that's another thing, but it does push the cost-benefit further into the red (at least for me).
Received on Tuesday, 14 November 2023 23:00:13 UTC