Re: [Ohttp] Scenario/use-case description ?

On Mon, Aug 2, 2021 at 7:31 AM Toerless Eckert <tte@cs.fau.de> wrote:
>
> Will the authors of the proposed ohttp documents and/or the proposed WG
> produce some comprehensive example scenario description in which ohttp
> is embedded ?

My email to the list shows one case where it is useful. Other cases
discussed at the BOF include telemetry, etc.
https://mailarchive.ietf.org/arch/msg/ohttp/XN6mMVTDxV8u6ZkVc4MTJe0zCEY/

>
> In the absence of such scenario description, i have a hard time vetting
> if or how well the proposed solutions will work to achieve anything.
>
> E.g.: Starting with some user writing herself a browser and ultimately
> wanting to use ohttp for privacy to access random (or selected) resources
> across the Internet. What type of discovery/lookup would the browser need to
> incooperate to learn all the required parameters to use ohttp and no
> divulge the target resource information in the process already.

I believe from the conversations at the BoF that this sort of scenario
wasn't considered in scope by most of the participants, or at least
wasn't a central application. Note HTTP proxies today aren't
"discoverable" or looked up, but there is a semistandardized
configuration format for them.

>
> The BoF sounded to me as if the knowledge of how this was supposed to work
> was in many participants heads, yet it is uclear from the discussion if they
> all thought about the same scenario and if they all assumed he same
> functionality. Without any written down agreement against such a comprehensive
> reference scenario, i can not see how we could honestly vet the proposed
> technology against the overall goals. Whether those goals are just
> privacy and/or maybe more (user-choice, non-collusion, etc. pp) is
> a secondary step. Aka: not much use discussing them without scenarios
> being documented.
>
> Cheers
>     Toerless
>
> --
> Ohttp mailing list
> Ohttp@ietf.org
> https://www.ietf.org/mailman/listinfo/ohttp

--
Astra mortemque praestare gradatim