Re: h2#404 requiring gzip and/or deflate

We know from both experimentation and real; deployment experience that
having compressed resources is usually a large latency win.
Compressing resources certainly can be expensive.
Decompression (which is really the only thing that is required to implement
here) is usually fairly cheap.

It would be silly to require everything to be compressed. That is not what
is happening here.

-=R


On Fri, Feb 21, 2014 at 12:46 PM, Zhong Yu <zhong.j.yu@gmail.com> wrote:

> On Fri, Feb 21, 2014 at 11:50 AM, Roberto Peon <grmocg@gmail.com> wrote:
> > The thinking when we put it into SPDY was that if you do gzip, you've
> > effectively required a deflate implementation, so one may as well allow
> it.
>
> If everybody can use either of the algorithms with ease, there's no
> point to mandate both. I also agree with Bjoern's point that
> gzip/deflate is pretty slow; enabling it may actually decrease
> throughput in a lot of deployments.
>
> Zhong Yu
>
> >
> > In the end, I'm ambivalent so long as there is a reasonable payload
> > compression requirement.
> > -=R
> >
> > On Feb 21, 2014 9:44 AM, "Martin Thomson" <martin.thomson@gmail.com>
> wrote:
> >>
> >> On 21 February 2014 09:38, Roberto Peon <grmocg@gmail.com> wrote:
> >> > No. Reverting to solely identity would be a huge step backwards.
> >>
> >> That's what I thought.  Which would you prefer gzip + deflate or just
> >> gzip?
>