- From: Lucas Pardue <lucas@lucaspardue.com>
- Date: Wed, 01 Jul 2026 22:52:54 +0100
- To: "HTTP Working Group" <ietf-http-wg@w3.org>
- Message-Id: <6bfacbf1-da73-47e7-ae7d-ad516bdf8511@app.fastmail.com>
I'm not tracking the value proposition(s) being discussed. Is it saving bandwidth, processing latency etc? What are concrete rought values of things and are they actual practical challenges on Internet deployments? On Wed, Jul 1, 2026, at 22:30, Julian Reschke wrote: > Am 01.07.2026 um 22:35 schrieb Mike Bishop: > > I think the difference in this proposal is that it's client-driven. If > > the server exposes an equivalent resource, great! We don't need this for > > repeat requests from the same client. > > > > Two cases where this has potentially more value: > > > > * > > First, on an intermediary. A server that doesn't produce an > > equivalent resource /also/ prevents intermediaries from caching > > responses. This mechanism could make it possible for an intermediary > > to cache even where the server isn't. > > Does it? An intermediate *can* cache the result of a QUERY (that was one > of the harder things to specify), just as it can cache GET responses. > It's just that the cache key is more complex. > > > * > > Second, cross-client. Multiple clients sending the same request > > won't benefit from equivalent resources, but could generate cache > > hits from this. > > Why wouldn't they benefit? > > I'm sure I'm missing something here, but what??? > > > Handwavy concerns about normalization apply to both cases. > > > > There's value there. I'm not certain it's /enough/ value to warrant the > > work, but it's definitely non-zero. > > Best regards, Julian > >
Received on Wednesday, 1 July 2026 21:53:22 UTC