W3C home > Mailing lists > Public > www-style@w3.org > January 2004

Re: Negative Border Widths

From: Gabriele Fava <gabriele_fava@virgilio.it>
Date: Wed, 24 Dec 2003 12:06:28 -0500 (EST)
Message-ID: <3FE9C789.6070902@virgilio.it>
Cc: www-style@w3.org

Dris wrote:

> This may or may not have been discussed before.  I just had a hunch 
> about border widths.
> Specifying a positive width means applying the border in a measure of 
> units outside the padding.  My thoughts say that specifying a negative 
> width would apply the border along the inside of the padding.  This 
> might provide an alternate way of working with the box model for 
> designers who are confused by it, or it could add more confusion.
> I don't know how this would fit into CSS3.  I don't even know how it 
> would fit into CSS2, because other logic might say the total 
> dimensions of the box would have to be decreased.  One could even go 
> crazy and say the border color would have to be inverted, citing some 
> weird physics formulae (I should hope nobody would look *that* far 
> into it).
> Most likely, this will be tossed into the "...right, no," pile.
> ________
> "Irony is a voluntary survey with required fields."
>     ~ Dris ~
I think this is a good idea, and I don't think there would be complexes 
like "invert the border color" or such: the desired effect (a border 
which expands into the padding) will have to be anyhow in case explicity 
specified, because in curret drafts negative values in border-widths are 
not permitted ( 
http://www.w3.org/TR/CSS21/box.html#border-width-properties ).
Still there's a problem, in that as those values are illegal, current 
browsers won't consider the declaration ( 
http://www.w3.org/TR/CSS21/syndata.html#illegalvalues ), and will 
probably consider the default value which is unfortunately for 
border-width medium; if we say that a negative border should develop 
into the padding, starting from the padding edge, we'll have different 
box dimensions in current browsers and in browsers supporting negative 
I think that the possibile spec-side solutions to this problem (like 
forcing padding, or margin, to expand properly with border-width 
negative values, ) are too tricky, but I don't either like very much 
asking authors additional attention (setting a border-width: 0em before 
a negative border-width) for backward compatibility; there are anyway 
similar precedents like text-shadow, so I commit to other's decision.
Received on Tuesday, 6 January 2004 11:37:24 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:10 UTC