Re: [css-regions] responsive and semantic use of named flows

On Jan 26, 2014, at 4:11 AM, Håkon Wium Lie <howcome@opera.com> wrote:
> 
> I have later learned that the web is different from desktop
> publishing. Web devices come in different shapes and sizes, and by
> clinging to the desktop publishing metaphor

I wouldn't say I was clinging to it. I would say it was useful enough then that I miss it even now, even after many years of specializing in Web-only design and creation. I would say we don't have to re-invent the wheel, when it is long overdue to adapt it to similar needs. 

> we risk having designs
> that look great if you happen to own a device like the one used by the
> designer. If you don't, the design crumbles easily.

Any design can crumble if the designer is short-sighted or rushed. The Web is full of fixed width layouts that do not adapt well to the variety of devices they might be viewed on. 

> Media queries can
> be used to provide alternative designs for alternative devices, but it
> gets verbose, tiresome and expensive. It is better to make sure CSS
> features are responsive from the outset, than patching them up later
> with media queries.

There is nothing I see in the regions spec that prevents flexible or adaptive layouts, or make media queries the only way to achieve them.  There are some simple examples designed to illustrate the basic features of regions in a simple way. I don't think it is fair to attack the examples on the grounds that they are not also great examples of page layout and adaptive designs. Few of the examples in any of the specs meet that requirement. Regions exist to define ways of flowing content between boxes and applying different styling based on which box the content ends up in. How the boxes are created or positioned in the first place is outside the scope, and is already being dealt with by other specs (such as Grid). It is not fair to attack Regions because it is not also Grid, or because something outside its scope might be used to create non-responsive designs. Regions doesn't attempt to solve every problem, but it does deal very well with the problems it does exist to solve. 

> As such, the first example in the Regions specification is telling. It
> creates a three-column design where the number of columns is fixed.

It compactly illustrates what it is attempting to illustrate, and no more. 

> David Baron has expressed this in a draft:
> 
>  http://dev.w3.org/csswg/css-overflow-3/

I think that shows a lot of promise as a way of creating region boxes out of fragmented content, but it doesn't negate the need for a regions spec.

Received on Sunday, 26 January 2014 19:55:23 UTC