Re: Fwd: I-D Action: draft-reschke-http-jfv-00.txt

Now that would be an interesting idea. My only challenge with it is that it
would be easy and tempting to abuse by stuffing way too much into the
"header", conflating "header" and "payload" and just plain over
complicating things. My WS-Therapy sessions are just now starting to do
some good, the twitching is nearly completely gone.

Sanity around header structure is a good thing, but I'm -1 on forcing
dependency on a json decoder. We are already heading the binary route.
Let's invest time in an efficient binary encoding that gives us easy,
reliable parsing. I know Julian has said before that he wants something
that would work in h1 bit, for me, that's simply not interesting.

- James
On Jul 14, 2014 6:22 AM, "Poul-Henning Kamp" <phk@phk.freebsd.dk> wrote:

> In message <53C3D406.1060101@gmx.de>, Julian Reschke writes:
>
> >FYI: this as a rough proposal how new header fields could use JSON as
> >syntax.
>
> I would *love* to see HTTP/3 impose a uniform well defined format on
> headers.
>
> It might even make sense to drop the concept of headers, and send
> them all in a single JSON metadata object.
>
> --
> 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.
>
>