Re: Encouraging a healthy HTTP/2 ecosystem

On Tue, Jul 1, 2014 at 1:32 PM, Jason Greene <jason.greene@redhat.com>
wrote:

>
> On Jul 1, 2014, at 3:00 PM, William Chan (陈智昌) <willchan@chromium.org>
> wrote:
>
> > Let's be clear here - we don't want to have a spec that doesn't mirror
> reality. If the spec says that something is a core feature, then it's
> terrible for it to fail to interop depending on which implementation you
> are speaking to.
> >
> > I think it's a completely fair argument to discuss whether or not
> CONTINUATION frames should be a core feature in the spec, or punted off to
> be negotiated instead or something. But *if* we keep it as a core part of
> the spec, it really must work, and I think it'd be healthy for the
> ecosystem if major implementations moved first to exercise as much of the
> spec as possible in order to prevent these kinds of interop issues from
> appearing.
> >
> > Please don't hijack this thread to make your debates about CONTINUATION
> frames. There are plenty of other threads for that already.
>
> How is this hijacking your thread? You used CONTINUATION as a specific
> example, and as far as I can tell it was the impetus.
>

It's hijacking my thread because it wasn't the impetus :) Indeed, as I
already said, I specifically caveated the CONTINUATION case because I knew
it was contentious. The impetus is because I've received word of MITM
proxies that plan to do deep inspection of HTTP/2. So I'm becoming worried
about ossification on these TLS connections that were previously free from
intermediary interference. If they move first, then we can't move later in
incompatible ways with these intermediaries. Therefore, I think it's in our
interest to make sure that whatever part of the protocol we think is
important to preserve for use should be actively exercised. And
CONTINUATION would only be one example of such a thing if we decide to keep
it. I purposely tried to avoid that discussion by caveating it, but I now
see my error in even mentioning it because everyone else has seemed to
latch onto it and hijacked this thread :(


> Although my point isn’t specific to CONTINUATION. My point is simply that
> if you are going to create some compliance black/white list, you need to
> ensure that you have fair criteria, and that also means ensuring various
> open ended aspects of the specs (such as allowing infinite data) are
> changed to specify the constraints you intend to enforce. Otherwise the its
> a purely subjective matter.
>
> --
> Jason T. Greene
> WildFly Lead / JBoss EAP Platform Architect
> JBoss, a division of Red Hat
>
>