@sean yeah that seems wrong =)

@sean if come around to this behavior being superior than responding to the n+1 request with a failure, even though I originally had your point of view as well. It requires potentially more logic/state server-side but it's a better client experience I think

@sean Looks like this is not yet in the core HTTP spec, but here's a paragraph in the not-yet-standard idempotency-key RFC:

https://www.ietf.org/archive/id/draft-ietf-httpapi-idempotency-key-header-07.html#section-2.6

The central idea is that if you receive 2 identical idempotent request, it's […]

@sean there's circumstances where the second case is now codified correct too, in case of idempotent multiple requests.

@mayintoronto Yathkyed Lake

it didn't 😢

@me seems neat but it makes me nervous without a comprehensive accessibility angle.

@tomjennings trains beat traffic