[whatwg] RWD Heaven: if browsers reported device capabilities in a request header

[Henri Sivonen]

On Sat, Feb 4, 2012 at 9:28 PM, irakli <irakli at gmail.com> wrote:
> Responsive Web Design [http://bit.ly/f6TPB7] is an extremely important
> approach/technique/movement for making web mobile-friendly.

It may look like it's the most important thing today, but most of the
Web isn't and won't be doing it. What goes in every HTTP request
affects the number of bytes transferred for all sites, though.

Ten years ago, advertising XHTML support in every HTTP request seemed
like the most important thing. And we got a bloated Accept header. (I
was guilty of advocating that. I have learned since.) And how often
did sites care? Very, very rarely. And once IE9 finally added XHTML
support, it also added SVG-in-text/html support, which pretty much
removes the need for application/xhtml+xml. It would have made more
sense to get the major servers fixed so that Accept: */* would have
become unnecessary and then to get rid of the Accept header.

It's been less than a week sense we got rid of the Accept-Charset dead
weight (http://hsivonen.iki.fi/accept-charset/) and you are already
proposing adding more. Everyone wants to have a piece of the UA string
or the HTTP request for their idea. None of those ideas are so great
in retrospect.

We should Just Say No to any and all proposals to add more stuff to
HTTP requests.

-- 
Henri Sivonen
hsivonen at iki.fi
http://hsivonen.iki.fi/