spec/css3-box: edited

Author: aprowse

spec/css3-box.txt

@@ -77,7 +77,7 @@ From css3-images:
 
 ===Current thinking===
 
-I think the whole concept of anonymous boxes is unnecessary and should be removed from CSS.  Properties are set on elements, and elements generate boxes based on its properties and on the formatting context in which its contents participate.  How an element imparts its properties on the boxes that it generates is the business of that element, and I don't see that we need a special term to describe a box whose property values all arise in some particular way.  For example, I see no difference between the way that a table element distributes its property values between the wrapper box and the table box and the way that a block element distributes its properties to a so-called anonymous box; inheritable properties are inherited down the box tree (although CSS doesn't state that explicitly anywhere) and non-inheritable properties get their values from somewhere as defined by the element "type".
+I think the whole concept of anonymous boxes is unnecessary and should be removed from CSS.  Properties are set on elements, and elements generate boxes based on its properties and on the formatting context in which its contents participate.  How an element imparts its properties to the boxes that it generates is the business of that element, and I don't see that we need a special term to describe a box whose property values all arise in some particular way.  For example, I see no difference between the way that a table element distributes its property values between the wrapper box and the table box and the way that a block element distributes its properties to a so-called anonymous box; inheritable properties are inherited down the box tree (although CSS doesn't state that explicitly anywhere) and non-inheritable properties get their values from somewhere as defined by the element "type".
 
 If the definition of principal box is modified so that an inline element generates a principal box (possibly fragmented across multiple line boxes), then perhaps we can define an anonymous box to be any box which is not a principal box.  In that case, "derived box" might be a better term since such a box might not be anonymous in the sense that there may well exist a selector which can directly address some properties of that box (eg table element selector which addresses the table box, list item selector which addresses the marker box or even a ::marker pseudo-element).