Re: Introducing a new HTTP response header for Carbon Emissions calculation

On Wed, 12 Apr 2023, 09:33 Poul-Henning Kamp, <phk@phk.freebsd.dk> wrote:

> --------
> Lucas Pardue writes:
>
> > Failure to provide client-side reporting risks creating a perverse
> > incentive to push more work onto the client, which is probably worse
> from a
> > net perspective.
>
> Ohh, absolutely.
>
> But it is tricker to implement because the significant client side
> emissions are from JS which runs after the transaction is nominally
> over, so the charge needs to be cached client-side and sent on the
> next request, if any.
>
> There is also the issue of how to prevent people from just
> charging 1 billion tons of CO2 to Exxon's website because they
> feel like it...
>

Yep. As discussed in the W3C WebPerf TPAC 2022 session on sustainability,
that I linked earlier, browsers already accommodate for such scenarios of
learning post-fact or deferred report delivery, by use of the reporting API
[1].

Preventing people from reporting nonsense is somewhat protected against if
the measurements are done internal to the user-agent. But there are a range
of potential privacy considerations that require thought.

Cheers
Lucas

[1]  https://www.w3.org/TR/reporting-1/#report-delivery



> --
> Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
> phk@FreeBSD.ORG         | TCP/IP since RFC 956
> FreeBSD committer       | BSD since 4.3-tahoe
> Never attribute to malice what can adequately be explained by incompetence.
>

Received on Wednesday, 12 April 2023 08:46:13 UTC