Re: Minor error in CSS2, section 14.2; 'background'

Hello,

"L. David Baron" wrote:
> 

[snip]

> > display properties. User agents should observe the following
> > precedence rules to apply background-color and
> > background-image to the canvas: If any property is declared
> > for the HTML element, extend the HTML background onto the
> > canvas, otherwise extend the BODY background onto the canvas.
> > If the resulting background-color is 'transparent' or
> > 'inherited' then the background-color of the canvas is
> > undefined.
> 
> I think that's actually fundamentally against CSS cascading rules,
> because it should not matter whether a property has its value (be it
> 'inherit' or not) because that value was declared or because the
> property was given its default value.  So far, nothing in CSS has
> broken this (to my knowledge) [1], although the temptation is sometimes
> great to propose such rules.  The rule should depend only on what the
> values are, not how they came to be.  Requiring knowledge of the source
> of values would, I think, impose considerable difficulty on CSS
> implementations.
> 
> David
> 
> [1] For some properties (those that inherit by default), the default
> value is different for the root element, though.  A more logical way of
> handling this might be to describe a super-root element where all the
> properties' values are UA-dependant.  This might also explain:
> 

Perhaps I'm misunderstanding the point here, but wouldn't using such
a super-root element somewhat defeat the purpose of CSS? I mean I 
thought part of the raison d'etre for CSS was to get away from 
properties being UA-dependent, and to define a design and layout
mechanism that rendered similarly with ANY UA? Wouldn't a better 
solution be to better define the root defaults and how they are 
inherited by properties that inherit by default?

Rick J.
firespring@nfx.net

Received on Wednesday, 15 September 1999 12:40:15 UTC