W3C home > Mailing lists > Public > www-style@w3.org > April 2009

Re: [css3-background] Issues open for feedback

From: Brad Kemper <brad.kemper@gmail.com>
Date: Thu, 16 Apr 2009 23:08:03 -0700
Cc: www-style@w3.org
Message-Id: <2632883E-6978-4991-AC82-F5463444D817@gmail.com>
To: fantasai <fantasai.lists@inkedblade.net>

On Apr 13, 2009, at 4:10 PM, fantasai wrote:

> Hello www-style,
> I've posted a summary of the major open issues we've been discussing
> here on css3.info: http://www.css3.info/border-image-issues/
>
> Feel free to comment here, there, or in the appropriate threads. :)
>
> ~fantasai

I had a few more thoughts about the numbered red list on that page:


	• If there is a border image, the ‘border-style’ is ignored (i.e.,  
drawn as if it were ‘none’)

That's not quite right. Ignored, yes, but not as if 'border-style'  
were 'none'. If it was 'none', it would be the same as if 'border- 
width' was zero, and thee element would take up less space if it had  
dimension, or the padding would be in a different place if it was auto  
sized. A transparent border-color would be a better way to described  
the intended effect (to avoid jumping around geometry as the images  
load).

	• It is unknown if it is a requirement that edges can be set to be as  
thick as ‘border-width’ (e.g., to set the thickness to ‘thick’).

So far, without a way to scale the images evenly by a percentage,  
'border-width' is the only way to specify a single number that would  
scale all sides evenly (and that's only if all sides have the same  
'border-width'). So I don't know if scaling by 'border-width' is a  
requirement, but there should be a simple way to scale all sides by  
the same factor OR individually by side (and percentage is the obvious  
choice). Scaling by border-width is, meh, O.K. and maybe useful  
sometimes, but considerably less desirable and useful.

	• It is unknown if it is a requirement that the outer edge of the  
border image can be placed elsewhere than at the border edge – one can  
probably just move the border edge itself, by giving the box more/less  
padding and margin.

It is considerably less desirable to have to resize the box or give it  
negative margin in order to adjust the image position, and then have  
to readjust padding, etc. to put the contents back in the right place  
(and this is very ugly when the images haven't loaded yet or will  
never load).

	• The border image has no influence on the size of the box: the  
margin edge, border edge, padding edge and content edge do not depend  
on it in any way.

I (obviously) agree with this point, above, but it seems to be  
contradicting the one before it.

	• It is unknown what should happen when the border image is scaled to  
be too big for the box: do the nine parts overlap, are they clipped,  
scaled down?

I initially thought it should clip, but I've changed my mind and now  
favor scaling. It may not be possible to get reasonable clipping. For  
images optimized for border-image its almost guaranteed that any  
clipping would cut off something important. Also, the fact that border- 
radius scales when there is not enough room suggests similar behavior  
would be appropriate. This could even be used intentionally with an  
over-large image to intentionally get a fluidly resizable border  
pattern.
Received on Friday, 17 April 2009 06:08:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 22 May 2012 03:46:58 GMT