Re: SPDY Header Frames

On Jul 13, 2012 11:44 PM, "Poul-Henning Kamp" <phk@phk.freebsd.dk> wrote:
>
> In message <20120714052933.GL16256@1wt.eu>, Willy Tarreau writes:
>
> >If you're talking about Roberto's mail, I read quite the opposite in
> >fact. It was said that Google was open to have anything provided that
> >the concepts raised in SPDY were preserved.
>
> Well, which way is it Roberto ?
>
> Care to clarify ?
>

Willy has it right.

-=R

> >I have talked long hours with the SPDY team at IETF83. [...] They
> >clearly said they were open to changes. What else do you want ?
>
> I want a sensible and ethical engineering process, one that doesn't
> dive headlong into serialization of HTTP headers until we have
> established what problems HTTP/2.0 should solve.
>
> Sneaking things like server-push through the backdoor is simply not
> acceptable, without structured decision to give up on HTTP's strict
> request-response model.
>
> It may be that we want to loose the strict R-R model, it may be
> that we do not.
>
> But making the HTTP/2.0 process a matter of "who already has a ready
> ID" is not the way to decide a question like that.
>
> --
> 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.
>
 On Jul 13, 2012 11:44 PM, "Poul-Henning Kamp" <phk@phk.freebsd.dk> wrote:

> In message <20120714052933.GL16256@1wt.eu>, Willy Tarreau writes:
>
> >If you're talking about Roberto's mail, I read quite the opposite in
> >fact. It was said that Google was open to have anything provided that
> >the concepts raised in SPDY were preserved.
>
> Well, which way is it Roberto ?
>
> Care to clarify ?
>
> >I have talked long hours with the SPDY team at IETF83. [...] They
> >clearly said they were open to changes. What else do you want ?
>
> I want a sensible and ethical engineering process, one that doesn't
> dive headlong into serialization of HTTP headers until we have
> established what problems HTTP/2.0 should solve.
>
> Sneaking things like server-push through the backdoor is simply not
> acceptable, without structured decision to give up on HTTP's strict
> request-response model.
>
> It may be that we want to loose the strict R-R model, it may be
> that we do not.
>
> But making the HTTP/2.0 process a matter of "who already has a ready
> ID" is not the way to decide a question like that.
>
> --
> 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.
>
>