Re: [css3-gcpm][css3-multicol][css3-regions] page floats, column spans, implemented & revised

On 3/3/13 3:57 AM, "Håkon Wium Lie" <howcome@opera.com> wrote:

>James Holderness wrote:
>
>(snipped)
>
>I have hopes that we can agree on a model with
>the automation that css3-multicol provides, and with the
>configurability that css3-regions offers. Here's my attempt at
>resolving these:
>
>  http://dev.w3.org/csswg/css3-gcpm/#regions

I support the column selector defined in this section - this would be a
very good improvement to multicol. But it falls short of the capabilities
defined in css3-regions. Because CSS Regions can be any non-replaced block
container, they can participate in any positioning scheme. Columns are
constrained to be immediate children of a multicol element. CSS Regions
can be children of separate parents, such as different grid containers in
a paginated view.


CSS Regions are a lower-level mechanism than multicol, so in a lot of
cases it does not make sense to compare the two. Multicol can be built out
of regions (as is being done in WebKit at the moment), and upcoming layout
schemes can use regions (such as slots in grid layout or page templates,
or shadow insertion points). The css3-regions specification is designed to
be that layer that others build upon. Adding column selectors does not
transform columns into the kind of building block that's needed.

>> And in case you're interested, I recently wrote a blog post on some of
>>the
> > tricks we've been using in our implementation.
> > http://www.xn--8ws00zhy3a.com/blog/2013/02/columned-layouts
>
>Wow. Clever. In a weird-code kind of way :)

It's quite clever in how it gets around current limitations of multicol
and floats. I want to ensure that no one has to go to these lengths to
extend what we currently support in CSS. One of the main benefits of using
elements as regions is the scripting access that makes extending web
layout capabilities easier than what James had to do.


Thanks,

Alan

Received on Monday, 4 March 2013 21:30:52 UTC