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

RE: Possible inadequacies in CSS3 Border-Image proposed specification

From: Ernest Cline <ernestcline@mindspring.com>
Date: Sat, 17 Jan 2004 19:37:29 -0500
Message-ID: <410-22004101803729562@mindspring.com>
To: "Vincent Starre" <thebitman@comcast.net>, www-style@w3.org




> [Original Message]
> From: Vincent Starre <thebitman@comcast.net>
> To: <www-style@w3.org>
> Date: 1/17/2004 3:51:20 PM
> Subject: Possible inadequacies in CSS3 Border-Image proposed specification
>
>
> I've been browsing through the CSS3 drafts to see what things may come in 
> the future (deciding on whether I should wait for widespread CSS3 support 
> before changing the way my site is currently structured) and I noticed 
> some things which seem lacking in the draft regarding the module: Borders.
> Now this message may be taken as unwelcome, but as long as the 
> specification is still a work-in-progress it seems the right time (if not 
> the right place) to make note of such things. I am hoping that this 
> message will be seen and perhaps considered by whoever it is that 
> ultimately makes decisions.

The working group makes those decisions and while those of us not in
the working group may occasionally disagree with their decisions, I have
never known them to show offense at constructive criticism such as what
you have presented here.

> The current draft ( http://www.w3.org/TR/2002/WD-css3-border-20021107/ ) 
> would allow for up to 16 images to be specified for a border. I realize 
> that it is impossible, and doesn't make sense to try, to make everyone 
> happy. However, by my count, the logical number of images to allow for is 
> not 16, but 24. That is, 3 per edge and 3 per corner.
>
> Now, it may be possible to emulate this effect by setting
negative-padding 
> on the contained content, placing clones of the background-image or 
> background-color within the images used by the border itself. This would 
> seem like a dirty hack, however, and may in some cases be impossible to 
> implement (if, for example, the bottom-edge's position is variable, it 
> would be impossible to line up the background image with the border image)
>
> Yes, its usefulness would be limited, similar layouts using tables are
not 
> widespread, but they do exist. Because of their current difficulty, most 
> such layouts which I have seen resort to using a fixed-size. For these
few 
> people, allowing 3 images to be specified for each corner would be a
boon. 
> Come to think of it, I personally can recall more instances of seeing 3 
> images used for corners than instances of 3 images used for edges.
>
> I guess I've made my argument for allowing the use of three images per 
> corner. Does anyone have an argument against?
>
> Of course, if this is the wrong place to even mention such things, please 
> send me a private e-mail telling me who I should talk to.

This is absolutely the right place to be talking about such things.

I think the current sixteen border image model can be adapted to do
what you want without having to add eight additional images.
What is needed is a way to specify that the images should tile
in from the corners instead of out from the center of the border.
I can't think of any reason one would want to do both at the same time,
which would be the only reason one would need twenty four images
instead of sixteen.

The way to do that given the current model, (which now that I've
taken a closer look at doesn't make me happy, but I don't have
an alternative to present just yet) would be to add some additional
values to the 'border-fit' properties so that there exist versions
of the "repeat" and "overflow" values that would fill in from the
corners instead of the center of the border, as well as values
that would fill in from a particular corner, so that the:

ABBBBBBB
C
C
C
C
C

effect you described could be done.
Received on Saturday, 17 January 2004 19:37:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:25 GMT