Re: [css3-background] box-shadow spread Multiple Choice Question

On Fri, Jun 4, 2010 at 5:49 PM, Brad Kemper <brad.kemper@gmail.com> wrote:
> On Jun 4, 2010, at 10:41 AM, Brendan Kenny <bckenny@gmail.com> wrote:
>
>> Is there a reason why (1) needs to be specified as a length? It seems
>> more natural to specify a scale as a number multiplier or percentage.
>> This would also preserve shape, regardless of width/height ratio.
>
> That would make it a very different feature. We are not looking to add
> features at this stage, just clarify what the rendering can/should be for
> the features the WG agreed to already.
>
> And as I've said in the past, spread is not something created in order to
> simulate some phenomenon observed in nature. It is intended to give greater
> control of the placement and extent of the shadow at a level consistent with
> what is commonly available when creating shadows in familiar ways using
> tools such Photoshop.
>
> Preserving aspect ratio when specifying a percentage shadow-scaling factor
> would make the feature far less useful to those accustomed to being able to
> set actual spread for their shadows.

Yes, unsurprisingly I had not forgotten your opinion on this topic =)

The question was just procedural, and if changing the unit of the
parameter would be adding a feature, obviously that can't happen at
this point. If the "sabotage" comment was directed toward me, I don't
really understand the point, as both "Change spread to scale" and
"Drop spread radius" are options listed in the original choices.

I'll maintain that, if creating a "box-shadow" property ab initio, it
makes more sense -- logically and from an implementation standpoint --
to use a true (uniform) scale option as that is how a shadow actually
grows and shrinks. I also think that there will be some confused
authors wondering why their curved corners turn square, or barely
rounded rectangles turn more elliptical. Looked at from a
historical/print standpoint, this is really spread with an optional
blur, not a shadow. I've belabored this point already, so I'll stop
there.

All that said, if a number or percentage can't be used, (1) is
actually the least desirable choice, as neither authors expecting the
aspect ratio of the shadow to be maintained or the proper behavior of
a spread value will be satisfied.

I agree with Brad that "spread-radius" instead of "spread" is
confusing and doesn't seem to describe the the property anyway.
"blur-radius" does make sense, but plain "blur" might be better there,
I'm not sure.

Received on Sunday, 6 June 2010 19:53:25 UTC