W3C home > Mailing lists > Public > ietf-http-wg@w3.org > July to September 2014

Re: Reference set in HPACK

From: Eric J. Bowman <eric@bisonsystems.net>
Date: Wed, 2 Jul 2014 19:38:12 -0600
To: Roberto Peon <grmocg@gmail.com>
Cc: Poul-Henning Kamp <phk@phk.freebsd.dk>, Kazu Yamamoto <kazu@iij.ad.jp>, HTTP Working Group <ietf-http-wg@w3.org>
Message-Id: <20140702193812.3cf329d788bb2bfb7f47e42f@bisonsystems.net>
Roberto Peon wrote:
>
> You can cache an object that has inlined resources within it, but you
> cannot cache those resources separately.
> You can't cache a resource that doesn't have a name.
> 

If I need to cache a resource separately, I don't inline it. I'm just
not seeing the latency problem with inlining, if it only occurs when
the technique isn't used properly?

-Eric

>
> -=R
> 
> 
> On Wed, Jul 2, 2014 at 5:54 PM, Eric J. Bowman <eric@bisonsystems.net>
> wrote:
> 
> > Roberto Peon wrote:
> > >
> > > Inlining is the technique most often used to reduce the number of
> > > request.
> > >
> >
> > I've also always considered it a worthwhile technique for
> > byte-shaving, as combining lots of small images into one mosaic
> > image eliminates whatever redundant bytes are inherent to the
> > chosen image format.
> >
> > >
> > > It is a strategy that works well most of the time for cold
> > > pageloads, but it also harms latency for subsequent navigations
> > > deeper into the site as those resources cannot be cached.
> > >
> >
> > But you lost me there, as I've never had a problem caching those
> > resources? Or does "inlining" not mean what I think it means?
> >
> > -Eric
> >
Received on Thursday, 3 July 2014 01:38:30 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 30 March 2016 09:57:08 UTC