Hey Watson, Apologies if I should respond directly to the mailing list - my old W3C profile has disappeared and I'm trying to get it back...
If the consensus is that the SETTINGS frame is the best place for it, that's fine. Initially I decided on a new TLS extension because it seemed simpler and I didn't want to mess with SETTINGS. I agree that it will require both a new registry and updating it will probably be under the purview of the HTTP WG - as you say, a once-in-a-while RFC or similar. Note that the original QPACK header analysis was done in 2018 - that's 5 (almost 6!) years ago - a lifetime in the world of the internet. If I were to do similar analysis today, it would be clear that there are some very common headers in use which barely existed back then - for example, all the Client Hints headers. Primarily this isn't about specific headers - it's about future proofing. If *something* isn't implemented, then we're at risk either of companies coming up with their own QPACK static tables (which only work with *their* client devices/servers) or a sub-performant scenario where common headers are relegated to the QPACK dynamic table. Thanks, Rory -----Original Message----- From: Watson Ladd <watsonbl...@gmail.com> Sent: Monday, September 25, 2023 10:08 PM To: Hewitt, Rory <rhew...@akamai.com> Cc: HTTP Working Group <ietf-http...@w3.org>; TLS List <tls@ietf.org>; Bishop, Mike <mbis...@akamai.com> Subject: Re: [TLS] New Internet Draft: The qpack_static_table_version TLS extension On Mon, Sep 25, 2023, 2:44 PM Hewitt, Rory <rhewitt=40akamai....@dmarc.ietf.org> wrote: > > Dear HTTP-BIS & TLS folks, <snip> > I would appreciate any comments, positive or negative. Rory, Great to see someone new jump in and the text is pretty clear. I do have one question: why not use the SETTINGS frame? This seems like it would fit there. Also I think IANA isn't equipped to perform that analysis, but it's something the HTTP WG could do and periodically publish a very boring RFC for, or maybe even abuse expert review in the IANA process to get the changes done. I do think something like this is worthwhile in principle, but I wouldn't be surprised if the analysis comes back "HTTP usages haven't changed enough" for a bit. Sincerely, Watson Ladd _______________________________________________ TLS mailing list TLS@ietf.org https://www.ietf.org/mailman/listinfo/tls
