Fwd: [CSSWG] Minutes and Resolutions 2009-08-18

Sending to public list, instead of just WG list...

Begin forwarded message:

> Begin forwarded message:
>
>> Hyatt: ... SVG shadows
>> ChrisL: That's how I implemented these
>> Brad: Once we have SVG shadows, then anything we do for border-image
>>       just becomes redundant
>> Hyatt: For box-shadow, my concern is what the author expects.
>> Hyatt: I think there's really only two options: either shadow the
>>        border-image or suppress it
>> Hyatt: I don't see what the problem is with doing shadows on border- 
>> image
>> Hyatt: Is it just that we can't come to agreement on how it works?
>> fantasai: no, that's not the problem
>> Brad: We could add switches to controll what the shadow gets  
>> applied to
>> fantasai: we can add switches later
>> Hyatt: If you really want a separate border-shadow then we can add  
>> that
>>        and copy the box-shadow syntax
>> ...
>> Hyatt: If we have a box-shadow, then it tries to do this filled  
>> shadow
>>        effect where it's drawn outside the box
>> Hyatt: A border-shadow effect would shadow whatever's drawn for the  
>> border
>> Peter: Does border-shadow really change the shape of the box?
>> Hyatt: People expect it to
>> Hyatt: It looks like they change the shape of the box, but it's  
>> kinda fake
>> Bert: I don't want to have a non-continuous border change the shape  
>> of the
>>       box. I still want the box to be rectangular.
>> Hyatt: Most of our use of border-image is to round things
>> Brad: I think using border-image for interesting dotted patterns  
>> will be
>>       as interesting as using it for changing the shape of the box
>> Hyatt: It's going to be used for a lot of things. Any case where the
>>        built-in borders aren't good enough
>> ChrisL: So you're saying people want this immediately ... ?
>> Hyatt: Feature requests we've gotten:
>> Hyatt:  1. do border-shadow, which is just a property like box-shadow
>>            and text-shadow that would exactly shadow the border  
>> drawing
>> Hyatt: 2. A full-blown shadow property, that would shadow everything
>>           drawn inside the element
>> Hyatt: 3. Wanting to shadow pieces of SVG.
>> ChrisL: You can do that already with the filter property
>> ChrisL: The latest draft is adding some syntactic sugar for common  
>> effects
>> ChrisL: I believe that's already implemented in WebKit, actually
>> ChrisL: If we go ahead and do a border-shadow property, then I'd like
>>         the box-shadow property to be only affected by the box, not
>>         by the border-image
>> ChrisL: You'll only get a rectangular shadow
>> ChrisL: It's not what people want. It's just clear and simple.
>> Hyatt: The argument for suppressing the box-shadow instead of just
>>        drawing a rectangle
>> Hyatt: Is that if the image don't load you show the border with a  
>> shadow
>> Hyatt: I'm fine with just rendering the rectangular shadow
>> Hyatt: I think that's what we currently do
>> Hyatt: I think it's an important point that border-image doesn't  
>> change
>>        the shape of the box
>
> In thinking this over after the discussion, I am willing to relax my  
> stance a little. Even though I am in favor of having a box-shadow  
> fallback for when images don't load (or may still be loading, in the  
> case of progressive rendering), I am appreciating the argument that  
> we could just say that box shadows only apply to box shapes (that  
> is, either rectangular or rounded-corner rects), and not to other  
> apparent shapes layered on top of that, such as border-styles and  
> border-images.
>
> So, I propose [we just] add a simple note about that to border- 
> shadow, to finish that topic off in the backgrounds and borders  
> module and help advance its status. Then, separately, we should  
> explore the following...
>
> 1. A separate CSS drop-shadow property (alpha-based), with syntax  
> similar to box-shadow, and a keyword to say how much it applies to:  
> e.g.  that can be applied to just borders and border-images, or the  
> entire background-clip area of the element (including borders and  
> border-images, if background-clip:border-box), or to everything  
> including contents (that is, border-box and border-images, plus  
> children inherit it).
>
> 2. A way to specify fallbacks for no-image-loaded (or no-image- 
> loaded-yet) situations. I rather like Andrew F.'s solution using  
> pseudo-classes, but perhaps a media query would work as well.

Received on Wednesday, 26 August 2009 15:46:36 UTC