[CSS21] response to issue 142b (was Re: [CSS2.1] Assigning property values, Cascading, and Inheritance)

On Monday 2003-10-20 12:09 -0400, fantasai wrote:
> S6.2: Inheritance
> 
> # When inheritance occurs, elements inherit computed values. The
> # computed value from the parent element becomes both the specified
> # value and the computed value on the child.
> ...
> # Each property may also have a specified value of 'inherit', which
> # means that, for a given element, the property takes the same
> # computed value as the property for the element's parent.
> 
> There's a dichotomy in how the "specified value" for inherited values
> works, depending on whether it's explicit or implied. It would be best
> if the initial behavior of 'color' and "color: inherit" meant exactly
> the same thing. And this way the meaning of "color: inherit" on the
> root element would follow from the inheritance logic without having to
> be handled as another exception.

The working group doesn't want to change this at this time.  The rules
on how inheritance applies to specified and computed values can have
complicated effects and have been checked to ensure that they lead to
the outcome we expect.  In particular, these rules have not been
modified since CSS2.


My original proposal for changing the way computed values work actually
removed this dichotomy -- it made the specified value always be the
computed value of the parent, which seemed rather ugly, but I wanted to
avoid having to deal with 'inherit' in every "Computed Value:" line.
However, someone else came up with a better solution in what's now the
third paragraph of section 6.1.2 [1].

If we were to fix this now, I'd much prefer making the specified value
consistently be 'inherit'.  This would require, at a minimum, changes to
6.1.1 and 6.2.1.  However, the changes to 6.2.1 would introduce another
inconsistency that couldn't be fixed until CSS3 -- an inconsistency
between the second and third bullets in 6.1.1.  (CSS3's 'initial' value
could be used in the third bullet point.)

Since this change doesn't fix any real problems and is the type of
change that might introduce inconsistencies in other parts of the spec
that we won't find for a long time, I think it's better not to make the
change.

-David

[1] http://www.w3.org/TR/2003/WD-CSS21-20030915/cascade.html#computed-value

-- 
L. David Baron                                <URL: http://dbaron.org/ >

Received on Tuesday, 17 February 2004 17:37:41 UTC