Re: [css3-multicol] Intrinsic Widths of Multi-column elements (Take I)

Also sprach fantasai:

 > I've been thinking about multi-col shrinkwrap wwidths a lot due to
 > the interesting situations we get into with vertical-horizontal
 > mixed text. I think it's not very well defined in the multicol
 > spec right now, and I'd like us to do better and solve some more
 > use cases

The current spec doesn't try to define shrink-to-fit, it just points out
that it exits an uses the result in the pseudo-algorithm:

   shrink-to-fit: this variable represents the result of a
   shrink-to-fit computation. CSS does not define the exact algorithm.

   http://www.w3.org/TR/css3-multicol/#pseudo-algorithm

Shrink-to-fit is only used when the available width is unknown and
either column-count or column-width is auto. So, this is an edge case
and not a central part of multicol.

Still, if we can find an accurate and sensible definition, it will
be a good thing.

There are some good arguments for keeping the current definition, though.

First, the spec is in CR. 

Second, the current defintion will be one of least surprises. For
example, if a float is changed into a multicol element, it will most
likely retain its width.

Third, the current defintion seems to be the natural fallout: all four
browser implementations (webkit, mozilla, opera, IE10 (welcome!))
retain the width as per the current spec. (Prince seems to turn off
floating for multicol elements.)

I'd therefore be hesitant to make changes at this stage.

Cheers,

-h&kon
              Håkon Wium Lie                          CTO °þe®ª
howcome@opera.com                  http://people.opera.com/howcome

Received on Friday, 15 April 2011 12:41:45 UTC