RE: Media Queries and optimizing what data gets transferred

Sorry, I believe users should come before authors in web infrastructure
designs.  Users will just disable the new header if it does not work for them
and nothing is advanced.

The Javascript/DOM design does have security problems that allow the
server to download codes that leak UA state back to the web, but there
are efforts underway to address this.  Building more infrastructure
on these security holes does not help this effort.

The damage that the proposal does to caching would likely cause much
more latency and bandwidth usage than it tries to solve.  Every time
that the hints change the cache becomes stale and the UA needs to
download new resources.  As more and more media query variables
are added the space of hints grows and caching become ineffective.

For example, someone will propose a media query based on the ambient
light level and then every time the lighting changes the cache will become
stale.

If you could give some consideration to users who want to avoid an
increased fingerprint surface and who consider some of the state that
media queries access as private it would be appreciated.

cheers
Fred

> Date: Fri, 25 Jan 2013 17:31:03 +0200
> From: hsivonen@iki.fi
> To: www-style@w3.org
> CC: ilya@igvita.com
> Subject: Re: Media Queries and optimizing what data gets transferred
> 
> On Fri, Jan 25, 2013 at 5:09 PM, Fred Andrews <fredandw@live.com> wrote:
> > Why is a 'promise' required?
> 
> Web authors want things and JavaScript lets them have what they want.
> If CSS doesn't promise that they get what they want, they can and
> route around CSS to make sure. Thus, deciding that they want the wrong
> thing and not giving them what they want and CSS does not mean that
> they stop doing things that they want to do.
> 
> However, if CSS let them get what they want, there'd be less latency,
> which would be better for users.
> 
> -- 
> Henri Sivonen
> hsivonen@iki.fi
> http://hsivonen.iki.fi/
>