[css3-lists] remove "Complex Counter Styles" and "Optional Extended Counter Styles" sections

The current draft of the CSS3 Lists spec includes two sections "Complex
Counter Styles" and "Optional Extended Counter Styles".  I don't think
either of these belong in a spec for use with simple lists, they are
either tackling problems that are overly general or are clearly beyond
what is needed for simple lists.  This spec should be about how to style
simple lists and not about how to capture the infinite diversity of number
formatting.

This was discussed as an issue on the list back in April [1] and 
Håkon is again raising many of the same issues.  The only issue marked
in the spec in in the section entitled "Predefined Counter Styles":

  Issue 6:
  Should this chapter and the next be made an informative appendix
  rather than a required UA stylesheet? 

I think it would be best to remove sections 11 and 12.  At the very
least both sections should be marked for further discussion with an
appropriate issue:

  Issue:
  Is there really a strong use case for supporting these list styles?

I think part of the problem here is that some folks view the existence
of list styles in some document form as an indication of a use case for
inclusion of these styles in CSS [2].  But I think many of these examples
are actually not presentational at all.  For example, the section
headings in legal documents are part of the content, not part of the
presentation.

Section 12 is marked as "optional" but I don't think we should be
defining optional features if there isn't a strong use case to begin
with.  Optional features always lead to divergent implementations and
headaches for authors.  Either there's a solid use case and it's a
required feature or we should omit the feature from the spec.

I think it would be much simpler for this level to define
'@counter-style' as simply as possible, see how it's used in actual
practice, then refine and extend if necessary at a later point.  Spec'ing
out a whole smorgasbord of complex features in one level is a recipe for lots of
underutilized features that later need to be deprecated.

Cheers,

John Daggett

[1] Of lists and castles
http://lists.w3.org/Archives/Public/www-style/2011Apr/0705.html

[2] Tab's defense of the need for long lists
http://lists.w3.org/Archives/Public/www-style/2011Apr/0744.html

Received on Thursday, 24 November 2011 05:32:07 UTC