W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 2022

(wrong string) Éric Vyncke's No Objection on draft-ietf-httpbis-targeted-cache-control-03: (with COMMENT)

From: Mark Nottingham <mnot@mnot.net>
Date: Sat, 15 Jan 2022 12:11:04 +1100
Cc: "httpbis-chairs@ietf.org" <httpbis-chairs@ietf.org>, "draft-ietf-httpbis-targeted-cache-control@ietf.org" <draft-ietf-httpbis-targeted-cache-control@ietf.org>, The IESG <iesg@ietf.org>, "tpauly@apple.com" <tpauly@apple.com>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
Message-Id: <B7D220FA-2C4E-44ED-A4F7-53DC286D5AA4@mnot.net>
To: "Eric Vyncke (evyncke)" <evyncke@cisco.com>


> On 15 Jan 2022, at 1:35 am, Eric Vyncke (evyncke) <evyncke@cisco.com> wrote:
> 
>    The behaviour needs to be deterministic (there's a long history of caches taking liberty with directives, at the expense of interoperability). That's not to say it's inflexible; the target list is under control of the cache, and can be adapted on a per-request basis. It's just that when a request 'hits' the target list, it needs to operate in a deterministic fashion.
> 
> EV> how can it be deterministic (as seen by upstream/origin) when the target list is local. Still wondering but not blocking.

In the sense that once the semantics are invoked, they're consistent. The target list only affects invocation, not the semantics themselves.

>    Also, if we specified as SHOULD, we'd need to describe what conditions satisfy the SHOULD...
> 
> EV> this is indeed the usual issue when using "SHOULD"...

:)

--
Mark Nottingham   https://www.mnot.net/
Received on Saturday, 15 January 2022 01:11:24 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 15 January 2022 01:11:25 UTC