RE: [css3-background] Curved borders intersecting backgrounds of inner boxes

On Apr 12, 2010, at 6:47 PM, Brian Manthos wrote:
>> Given that neither padding nor border widths can be negative, it remains conveniently simple.

On Apr 12, 2010, at 10:37 PM, Brad Kemper wrote:
>I'm not sure I follow what you mean about the negative widths. Can you elaborate?

Sure.

http://www.w3.org/TR/CSS21/box.html#padding-properties
"Unlike margin properties, values for padding values cannot be negative."

http://dev.w3.org/csswg/css3-background/#border-width
Links to <border-width>...
http://dev.w3.org/csswg/css3-background/#ltborder-widthgt
"<border-width> = <length> | thin | medium | thick
The <length> may not be negative."

If padding were allowed to be negative, the content-box could grow larger than the padding-box.
If border widths were allowed to be negative, the padding-box could grow larger than the border-box.

With both of these possibilities ruled out, we can be assured that border-box is never smaller than padding-box which is never smaller than content-box.

This assurance allows us to ignore the border widths and padding to identify which of the 3 boxes is biggest.

Further, it means we don't have to worry about cases where the padding-box has grown wider but shorter than the border-box; similarly for the content-box and padding-box.  Were that not the case, then term "biggest" would mean area calculation which doesn't buy us much from an aesthetic rendering perspective at all; it would open the door for performance pain as well as visually undesirable renderings.

Received on Tuesday, 13 April 2010 06:06:27 UTC