Re: [css3-page] Idea for simplier page-margin boxes

Le 25/03/2013 04:15, Robert O'Callahan a écrit :
>
>     Using abspos also creates some dependencies. In the code above, I'm
>     guessing that the page number will fit in a box with a with of 4%
>     (100-88-8). For page numbers, that's a reasonable guess. For other
>     types of content, using abspos may cause overflow/overlap.
>
>
> CSS3 Paged Media solves this by introducing what is really a new layout
> model just for margin boxes. That is not simple. We propose to render
> actual elements in the margin boxes, so one could use flexbox to
> position the page number and chapter title instead of the pseudo-flexbox
> features in CSS3 Paged Media.
>
>     I'm not against adding abspos margin boxes for marginalia. However,
>     I've used the current css3-page model to publish real-world books [3]
>     and I find it simple & powerful -- not over-complicated & inadequate.
>
>
> It's simple and powerful from the point of view of authors who want to
> do exactly what the feature was designed to do. Since you designed the
> feature, I'm not surprised it's a good fit for your needs :-). But 16
> new pseudo-elements and a lot of spec text defining positioning and
> sizing for the margin-box layout model is not simple for implementors,
> or people trying to write tests for the feature, or authors trying to
> understand the behavior they're getting, or authors trying to understand
> how to get the behavior they want. And it will get more complex if/when
> we add features for the use-cases Julian introduced at the start of this
> thread.

I think there are two separate issues:

1. What you can put inside page-margin boxes. css3-page only gives the 
'content' property, where each component creates an anonymous box that 
can not be styled individually. I expect css4-page to provide a way to 
flow elements into page-margin boxes. With that, page-margin boxes could 
contain eg. flexboxes.

2. The layout of paged-margin boxes themselves. The "fixed dimension" 
part is relatively simple (similar to inline dimension of block boxes), 
but I agree that the variable dimension is more complicated than it 
should be. The previous algorithm was even more horrible to implement 
(look at the 2006 WD if you’re curious) but I tried to make it better. 
I’m open to changing it again, possibly using the flexbox algorithm. 
This a big sub-issue, so I started a new thread:
http://lists.w3.org/Archives/Public/www-style/2013Mar/0557.html

Based on what Daniel said, I don’t expect his proposals for css4-page to 
have anything about 2.

-- 
Simon Sapin

Received on Monday, 25 March 2013 13:10:50 UTC