Re: [css3-multicol] [css3-grid] and other layout managers

On Dec 15, 2007, at 12:59 AM, Andrew Fedoniouk wrote:

>>> Otherwise redefining solid background color for some element  
>>> requires explicit disabling of all other background attributes  
>>> like images and gradients.
>> It sounds like you are suggesting to have the 5 methods you  
>> mentioned by all mutually exclusive. But I would prefer to be able  
>> to combine solid with background-image, as we can today. Gradients  
>> could be combined with background-image in the same way.
>
> Well, it was just thoughts aloud. Sample with background is  
> probably not the best one as, indeed, background image can be  
> combined with solid and gradient fills.
>
> But again solid and gradient fills are different types of fill method.
>
> background-fill: solid { color:#ddd; } - mapped on background-color
> background-fill: gradient-linear { start:...; end:...; stops:...; }
> background-fill: gradient-radial { start:...; end:...; stops:...; }
> background-fill: gradient-diamond { a:...; b:...; c:...; }
> background-fill: fractal-plasma { color:...; }
> background-fill: fractal-cloud { :...; }
>
> And these are mutually exclusive, isn't it?

Yes. They also could all be considered gradient types. In theory,  
even a solid fill could be a gradient between two like colors,  
although no one would want to specify them that way. I would want to  
indicate a solid background-color first as a fallback for UAs that  
did not support the fancier fills.

For instance, a fractal cloud, using the gradient syntax I proposed  
in another thread, could be written like this:

	background-color: gradient(fractal-cloud, blue .5in 123, #67f 50% . 
5, transparent 100% 1.0)

In this example, fractal clouds would take 3 values in each comma  
separated grouping. In the first grouping the third value is the  
random seed, and in successive groupings it would be the strength of  
the cloud.

Your idea is interesting for an alternate way of specifying  
something, but I prefer the shorter, simpler forms.


>
>> Disabling the image is only necessary if you no longer want the  
>> image on top of your solid or blended color, so I have no problem  
>> with the way that is already done.
>
> background-image is by itself an example of various types of  
> rendering methods where each method may use their own set of sub  
> attributes
>
> background-image: repeat {  ... }
> background-image: repeat-x {  ... }
> background-image: repeat-y {  ... }
> background-image: no-repeat {  ... }
> background-image: expand {  ... }
> background-image: stretch {  ... }
> etc.

I'm not sure what the advantage would be of specifying these this  
way, rather than as a single parameter for "repeat". I suppose the  
more verbose form it makes it more explicit about what each value is  
for, but I would still prefer to write it the other way, because it  
is shorter and simpler, and doesn't require me to remember the  
parameter names.

>> In regards to disabling a gradient to show a solid color, this  
>> becomes a moot point if the gradient is used as a color value.  
>> There is probably no need to have a gradient with translucency in  
>> front of a solid color on the same object, or a solid but  
>> translucent color in front of a gradient on the same object. Which  
>> is one reason why the idea to use a gradient as a color value for  
>> background-color, color, border-color, etc. was so brilliant.
>
> background-color: gradient() sounds scary for me to be honest.
> And color:gradient() and border-color:gradient() sounds scary even  
> more if you will look on them from computation point of view.

>
It's OK, no need to be scared. It won't bite. ;-)

I don't know about computational difficulties. It doesn't seem to me  
to be any worse than some of the other proposals involving SVG or  
@rules.

There have been various times when I would have liked to use a subtle  
gradient in a border, to complement a different gradient in the  
background fill, but I could not practically do so without a lot of  
extra work, so I did not. And who hasn't occasionally wanted to add a  
gradient to words? [don't answer; it's a rhetorical question.] If  
used with outline-color, and a shape-gradient type that radiates out  
like a radial gradient (but from the inner dimensions of the outline  
to the outer), you could create glows similar to what Apple does with  
selected UI elements in Mac OS X.