W3C home > Mailing lists > Public > whatwg@whatwg.org > November 2013

Re: [whatwg] The src-N proposal

From: Yoav Weiss <yoav@yoav.ws>
Date: Tue, 19 Nov 2013 08:49:13 +0100
Message-ID: <CACj=BEgBqKvTxj_=fRtafhXMC0C=ikiOSJs1A86_Jb10voGcwg@mail.gmail.com>
To: Bruno Racineux <bruno@hexanet.net>
Cc: whatwg <whatwg@lists.whatwg.org>, Ilya Grigorik <igrigorik@gmail.com>, Marcos Caceres <w3c@marcosc.com>
On Tue, Nov 19, 2013 at 5:53 AM, Bruno Racineux <bruno@hexanet.net> wrote:

> On 11/18/13 8:21 PM, "Ilya Grigorik" <igrigorik@gmail.com> wrote:
> >
> > On Mon, Nov 18, 2013 at 8:13 PM, Bruno Racineux <bruno@hexanet.net>
> wrote:
> >> Because these (only 0.2% uzing gzip) stats do not look good at all in
> >> support of your theoretical argument:
> >> http://trends.builtwith.com/Server/GZIP-Module
> >
> > That measures "mod_gzip" adoption.
> >
> > HTTP Archive tracks top 300K (Alexa) sites, and the actual number has
> been
> > hovering ~70-75% for a long time:
> > http://httparchive.org/trends.php#perCompressed
> >
> > </aside>
> I suspected I was missing something.
> Though httparchive stats  are "number of compressed responses over the
> number of HTML, CSS, and JavaScript requests" which is different that a per
> website coverage. Correct?

My personal experience tells me that the number of "gzipped HTML" is most
probably higher, since most of the non-gzipped text resources I see are JS
& CSS. Besides, if authors are not gzipping their HTML, DRYing the URLs
will be like putting a band-aid on a major carotid artery trauma. (i.e. not
very helpful, even if it makes you feel like you "did something")

In any case, since URL DRYing does not solve any of the use-cases and
doesn't bring any new functionality, this is a tangent concern that IMO
would be best tackled later on, once we'd have actual data on the usage
patterns, and we'd be able to estimate how important DRYing the URLs really
is. (No need to wait for author complaints, as we can crawl sites and look
for patterns)

If you'd like to further discuss that concern, can you please open up a new
thread? This one discusses solutions to the responsive images problem
use-cases (e.g. src-N, picture, srcset, etc)

> I assumed the mod_gzip sits looks at the: 'Content-Encoding:gzip ' header.
> Or what else could it look at?
> Wouldn't this  mean that low end site not in the top 300K on Alexa have a
> much higher non-gzipped rate?
Received on Tuesday, 19 November 2013 07:49:39 UTC

This archive was generated by hypermail 2.4.0 : Wednesday, 22 January 2020 17:00:14 UTC