Re: [css3-background] vastly different takes on "blur"

On Jun 22, 2010, at 3:20 PM, Tab Atkins Jr. wrote:

> On Tue, Jun 22, 2010 at 1:54 PM, Simon Fraser <smfr@me.com> wrote:
>> On Jun 22, 2010, at 1:42 PM, Brad Kemper wrote:
>> 
>>> On Jun 22, 2010, at 12:07 PM, "Tab Atkins Jr." <jackalmage@gmail.com> wrote:
>>> 
>>>> What we're all saying is that, when we're thinking about how much to
>>>> blur, what we mentally care about is how much the blur extends out
>>>> from the normal shadow.  The amount that the blur extends into the
>>>> normal shadow isn't relevant to our decision on how much to blur.
>>> 
>>> That's the part that still seems indefensible. Whatever direction it extends, it is part of the blur for which the author provided a distance measure. How can that possibly be irrelevant to the blur value provided? If all you cared about was how far it how far it extended (and I know ghats nit the case), you wouldn't need blur. So when you say you are visually picking an attractive blur in conjunction with that decision, then you are just visually and concurrently picking a blurriness that can be represented by a number, regardless of which way it extends.
>> 
>> There are two another reasons to keep the definition of the blur amount that you disagree with.
>> 
>> 1. We already have text-shadow, whose blur behavior we do not want to change.
> 
> Webkit's and Firefox's text-shadow behavior appears to be consistent
> with Brad's preferred behavior.  I just put together a testcase with
> "font-size:1000px; text-shadow:0 0 100px black;" and it's pretty
> clearly extending roughly 50px out and 50px in.
> 
> If we don't want to change text-shadow (we don't), then we should go
> ahead and make the blur length specify the full size of the blur
> region, like Brad suggests.  While I do prefer the other way, I value
> consistency between nearly-identical properties more.

That's not what I see. Don't forget that shadows larger than 8px are buggy in WebKit. Here's a test with an 8px shadow:
<http://smfr.org/misc/shadow.html>

and here's what it looks like in WebKit:
<http://smfr.org/misc/shadow-webkit.png>

The box and text shadows are identical, and Pixie shows that the shadow extends out by 8px (for a total shadow transition of 16px).

>> 2. In the public-fx group, we will be discussion the addition of filters to CSS, with convenience properties for common filters. Blur will be one of these, and it will have a radius, as input, which should give behavior comparable to box/text-shadow's blur radius. This filter will be based on SVG's gaussian blur, and so the behavior of the radius parameter should match SVG.
> 
> If we want the blur filter to work similarly to text-shadow, then it
> needs to work as Brad wants.

I think we agree that the following should all match:

* text-shadow
* box-shadow
* SVG gaussian blur
* hypothetical future blur filter in CSS
* PhotoShop, probably

and bonus points if we choose something that matches what browsers do today (modulo bugs).

Simon

Received on Tuesday, 22 June 2010 22:39:49 UTC