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

[css-book] Re: [css-page] Meaning of :first page selector (was: [css3-gcpm] Page groups, named pages, and :first)

From: Håkon Wium Lie <howcome@opera.com>
Date: Tue, 22 Oct 2013 14:40:39 +0200
Message-ID: <21094.29255.317270.93135@gargle.gargle.HOWL>
To: Simon Sapin <simon.sapin@exyr.org>
Cc: "Cramer\, Dave" <Dave.Cramer@hbgusa.com>, "www-style\@w3.org" <www-style@w3.org>, MURAKAMI Shinyu <murakami@antenna.co.jp>, Michael Day <mikeday@yeslogic.com>
Simon Sapin wrote:

 > > I would like to note that solutions that involve the content of margin
 > > boxes (like first-except) don't meet all our use cases. A chapter-opening
 > > page may have a background-image, different margins, etc.
 > 
 > Again, I definitely agree that that page selectors need to be more 
 > powerful, including to select the first page of a group. (For some 
 > definition of "group" that remains to be specified.)

I've tried to specify what a page group is here:

  Named pages can appear in sequence, stemming from different
  elements. A sequence of pages with the same name is called a page
  group. It is sometimes necessary to split one page group into
  several page groups in order to, e.g., address the first page in the
  group. The page-group property expresses whether an element starts a
  new page group or not.

  http://books.spec.whatwg.org/#page-groups

 > For 15 years since CSS 2.0 (which already had named pages), ':first' has 
 > been defined as "first page of the document":
 > 
 > http://www.w3.org/TR/1998/REC-CSS2-19980512/page.html#q8
 > 
 > Instead of changing it now, I’d rather have a new page selector such as 
 > :first-of-group and possibly :nth-of-group() (together with :nth() which 
 > counts in the document.)

To me, it seem quite intuitive that 

   @page :nth(1) { ... }  

selects the first page and that

   @page funky:nth(1) { ... } 

selects the first page of the "funky" group

I don't see a need for further syntactic differentiations. Also, this
has been implemented and used for years.

 > As to how page groups are delimited, a boolean 'page-group' property 
 > does not seem very elegant. Perhaps an optional keyword on the 'page' 
 > property?

I agree that boolean properties should be avoided. This is probably
why it took so long to bring it into the spec. But optional keywords
will turn the page property into a shorthand and someone will quicly
ask to expose it in a separate property.

-h&kon
              Håkon Wium Lie                          CTO °þe®ª
howcome@opera.com                  http://people.opera.com/howcome
Received on Tuesday, 22 October 2013 12:41:27 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 22 October 2013 12:41:28 UTC