W3C home > Mailing lists > Public > www-style@w3.org > October 2013

Re: [css-regions] Named Flows, Elements and Box Generation

From: Håkon Wium Lie <howcome@opera.com>
Date: Sun, 27 Oct 2013 01:47:09 +0200
Message-ID: <21100.21629.441979.596514@gargle.gargle.HOWL>
To: Johannes Wilm <johannes@fiduswriter.org>
Cc: www-style@w3.org
Hello Johannes, 

 > you might remember I wrote a little while ago. We have created a little
 > javascript app that renders pages in the browser the way they are printed
 > using CSS Regions. ( http://sourcefabric.github.io/BookJS/ )

Yes, here's the conversation:

   http://lists.w3.org/Archives/Public/www-style/2013Apr/0210.html

I recoded your examples into pure HTML/CSS here:

  http://lists.w3.org/Archives/Public/www-style/2013Apr/0631.html

My conclusion at the time was that it was simpler and easier to write
this in HTML+CSS, without using the complex book.js. 

  http://sourcefabric.github.io/BookJS/book.js

I'm not dissing your code; I think it's amazing what you've been able
to do in legacy browsers by adding JavaScript. But I do think that
(say) footnotes and top floats should be "first-class citizens" in a
page-based rendering systems, and not just the outcome of a javascript
reordering algorithm.

They are described as first-class citizens here:

   http://books.spec.whatwg.org/
   http://figures.spec.whatwg.org/

 > This includes not only splitting the text into individual pages, but also
 > adding footnotes, margin notes, cross references, word index, top floats,
 > etc. . The idea is that something can be a footnote on one device, and on
 > another device it may make more sense to show the same content as a margin
 > note, or a floating box or some third thing.

Indeed, a noble goal.

 > What worries me a bit more is reading here and also in a few other fora,
 > that the Mozilla camp is opposed to CSS Regions with named flows entirely.
 > 
 > I can understand that overflow:fragments make a lot of sense in many cases
 > when nothing more is required. But if overflow boxes all have to be
 > siblings, that would certainly be problematic if you do stuff like we do
 > it. Or what about doing subflows from flows?

Could you point to one of your example where you use this?

 > I do not get the problem with the HTML-element either. Whenever one makes
 > just the most simple websites, one includes a lot of extra elements to be
 > able to position other elements within it.

We're trying to change that.

That's why, for example, you can now have multiple backgrounds on a single element.

 > One of the most basic things to
 > position elements is to have an outer element with the position:absolute
 > inside an position:relative element. I don't see how adding an extra
 > element for CSS Region-flowing is much different.

The fact that 'position: relative' establishes a containing block has
led to much abuse. We're trying to not make the same mistakes.

Cheers,

-h&kon
              Håkon Wium Lie                          CTO °þe®ª
howcome@opera.com                  http://people.opera.com/howcome
Received on Saturday, 26 October 2013 23:47:46 UTC

This archive was generated by hypermail 2.3.1 : Saturday, 26 October 2013 23:47:47 UTC