W3C home > Mailing lists > Public > www-style@w3.org > December 2011

Re: [css3-multicol] splitting floats across column boundaries

From: L. David Baron <dbaron@dbaron.org>
Date: Thu, 1 Dec 2011 21:17:10 -0800
To: Alex Danilo <alex@abbra.com>
Cc: robert@ocallahan.org, "www-style@w3.org" <www-style@w3.org>
Message-ID: <20111202051710.GA8389@pickering.dbaron.org>
On Friday 2011-12-02 15:50 +1100, Alex Danilo wrote:
> >On Fri, Dec 2, 2011 at 1:48 PM, Alan Stearns <stearns@adobe.com> wrote:
> >
> >Take a look at this file. It has a div floated left with six lines of text
> >in a 2-column layout containing eight lines of text
> >
> >http://hg.csswg.org/test/raw-file/50587184a512/contributors/adobe/incoming/f
> >loatLeftMulticolBreak.html
> >
> >In WebKit the columns are balanced by splitting the float in two. In Gecko
> >and Opera the float remains intact and the columns are not balanced. Should
> >I expect more interop here, or is this a vague area that css3-pagination is
> >meant to resolve?
> >
> >I think it should split. I don't know how to spec that, but it should be specified.
> Why? I'd be curious to know the reasoning.

My reasoning (though I'm not Robert) would basically be this:

First, I'd be very skeptical of specifying rules that are different
for page breaking or column breaking.  I'd expect the rules to be
the same.  And if they're different, it's likely to add extra
complexity to implementations, tests, and specifications, which
(given that breaking generally isn't the highest of priorities) is
likely to just mean it's buggier.

Second, while floats were designed for a particular layout concept
(for which they are widely used on wikipedia, but not a whole lot of
other places), they are widely used on the Web as a tool for layout.
Thus, when printing most Web pages that use floats, it makes sense
to break them as close as possible to the idea that you're
displaying the Web page on a canvas and then just splitting that
canvas at horizontal lines to separate it into pages (except that
you don't split a line between sheets).  (I've heard this called the
toilet paper model of pagination.)  So I think making floats
implicitly page-break-inside:avoid would be bad for most uses of
floats on the Web, even though those uses aren't what floats were
designed for.  (Then again, if we have real layout systems designed
for user interface, like flexbox or grid, this might change.)


𝄞   L. David Baron                         http://dbaron.org/   𝄂
𝄢   Mozilla                           http://www.mozilla.org/   𝄂
Received on Friday, 2 December 2011 05:17:46 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:08 UTC