Re: [css3-background] box-shadow spread radius and rounded corners

On Apr 28, 2010, at 1:06 PM, Sylvain Galineau <sylvaing@microsoft.com>  
wrote:

> Brad, I don't understand what is 'extreme' about my example or why  
> it's more extreme than your own.

Mine was a little extreme too, but mainly because it is harder to see  
where spread stops and blur starts when using smaller values. Usually  
when I spread shadows in PhotoShop, it is by a few points, usually  
less than the blur amount. When you have a lot of spread and not much  
(or any) blur, it starts looking less like a shadow, and more like  
something else. I can't recall the last time I used more than 10pt of  
spread on anything that was supposed to really look like a shadow.  
Sometimes, when using it as a faux glow (blur, but no offset), I add a  
lot of spread in order to get the solid color fully outside the box,  
beefing up an otherwise large blur (and the falling off of opacity  
associayed with that) with more visible color.  But even then, I don't  
think I've ever had 50 points (or pixels) or more of spread except  
when working at very large sizes.

> ...- then I do reach a point where the shadow's shape is very  
> different from that of the box
> being shadowed, as well as that of previous less spread-out shadows.


>  ...But has the *shape* of the shadow been preserved all along ? I  
> don't
>
> get that part at all.

>  ...And then inferred it
> from a requirement to preserve shape when a spread radius is  
> applied, requirement I do not see
> being fulfilled with an even surface spread.


> ...Most importantly, what shape is being preserved, or
>
> how it is preserved ought to be clarified. Maybe it's just a matter  
> of making it clear that *only* sharp
> corners retain their border-radius ?

Yes. All this talk about shape is, I believe, due to reading too much  
into that sentence, which was added after a telecom where we decided  
that we should preserve sharp corners. I don't think there was ever a  
question about not preserving the shape of curves in the same as as  
with borders, ie with the outer radius being bigger than the inner  
radius so that a single curved shape has an even amount of spread  
following it.

I agree that we can probably improve the text to be more clear.

> I think an example like the one above annotated with arrows may be  
> helpful in clarifying the meaning
> of spread radius and I can try to produce one.

We can use the one I created and linked to in an earlier post today,  
if you like, or I can improve it if that is not enough.