On Mon, 25 Sep 2006, Anne van Kesteren wrote:

> On Mon, 25 Sep 2006 14:54:30 +0200, David Latapie <david@empyree.org> wrote:
>> -- <em role="0">    default
>> -- <em role="+1">   equivalent to em
>> -- <em role="+2">   equivalent to strong
>> -- <em role="-1">   less important, may be rendered as font-size:smaller
>
> This proposal doesn't cover nesting.

Why not? Isn't the possibility of nesting an immediate consequence of the 
syntax? Or do you mean that the _meaning_ of nesting is left unspecified? 
If <em> means emphasis with respect to the environment (or, syntactically, 
the parent element), then obviously
   <em role="+1"><em role="+1">foo</em></em> bar
would make foo emphasized twice with respect to bar. Whether this would be 
semantically equivalent to
   <em role="+2">foo</em> bar
might need to be defined. That is, do the role="..." attributes express 
something that is additive?

I must admit that the role="..." stuff looks like a pointless game. 
Instead of defining an attribute with suggestive or descriptive name,
an almost dummy name is used.

I think what would really be needed is <b>, <i>, and <small> with new 
names and more abstract semantics as well as an attribute that lets the 
author declare a block element as more or less important. But I'm afraid 
that would be far too practical - useable and implementable.

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/