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

[Matthew Wilcox]

On 7 February 2012 20:05, Nils Dagsson Moskopp
<nils at dieweltistgarnichtso.net> wrote:
> Matthew Wilcox <mail at matthewwilcox.com> schrieb am Tue, 7 Feb 2012
> 19:38:31 +0000:
>
>> Can we not turn this into an option in the same way browsers handle
>> requests to get the users location? With configuration too?
>>
>> Allow browsers to see my:
>>
>> screen size
>
> That would be yet another way to push “This web site was designed for
> 800x600, 16 colors, IE 5.5 or later.” – only with the added “feature”
> of “sucks to be you, get a bigger screen, this content will not
> display”. UA sniffing, now even worse! Oh, I would certainly be mad.

I tire of this viewpoint.
Of course it can be used that way. There isn't a technology in
existance that can't be mis-used. Does that mean it's a bad idea to
try to solve problems with new technology?

>
>> connection speed
>
> This is an interesting case – but today well handled by minimal sites.
> OkCupid, Twitter, Wikipedia all have well-known minimal alternatives
> that can be used when on a bandwith-constrained connection. Often, they
> are arguably better (not full of ads etc.).

Handled how? "By minimal sites" is not an answer, that's a design
approach consisting of "bring everything down to the lowest common
denominator" if you use it for one site, or an extension of "just
throw up a plain-text version" if you supply it as a separate URL.

The point of responsive designs is to allow one URL to work well on
multiple un-known devices whose characteristics vary enormously. The
screen can have a ten fold difference in width and a greater
difference in processing power and network connectivity. One of the
main issue in this arena is how to adapt images, because you don't
want to send big images to small devices. So, how do you see this
particular problem as already solved? Please - myself and a few dozen
other people who've been at this for the last year would like to know
- how do we serve a small image to small devices and a big one to
large devices? Reliably? Without the flaws of all the existing
attempted solutions, which are well documented elsewhere?

http://adaptive-images.com
http://filamentgroup.com/lab/responsive_images_experimenting_with_context_aware_image_sizing/
http://www.alistapart.com/articles/responsive-images-how-they-almost-worked-and-what-we-need/
etc

PS: UA sniffing and dedicated 'mobile' site's are expressly not the answer.

> Since that type of request is not narrowly constrained to specific
> circumstances (e.g. location-based services), there is a possibility
> for this to become very annoying.
>
> Externalizing the inability to provide content that works across
> multiple devices is a very easy cop-out for developers. I for one
> advocate control at the UA level (like with CSS media queries) so there
> won't be yet another UA string nightmare.

I would hate another UA string nightmare too. Hence why I want
headers. Explicit, direct, non-assumed reporting of actual, current
properties. Without guessing.

> Also, I am writing this on a laptop via a throttled mobile connection.

It'd be nice if sites had the capability to adapt to that throttle
then wouldn't it...

> --
> Nils Dagsson Moskopp // erlehmann
> <http://dieweltistgarnichtso.net>