Re: [css-color] color-correction into color-interpolation

On Sun, Apr 7, 2013 at 9:28 AM, Dirk Schulze <dschulze@adobe.com> wrote:

>
> On Apr 4, 2013, at 6:40 PM, Rik Cabanier <cabanier@gmail.com> wrote:
>
> > Since color-interpolation is of such questionable value, why don't we
> just deprecate it?
> > It's use can be replaced by using the Lab color spaces since it is
> linear.
> >
>
> Even if I would agree on deprecating for the general usage on arbitrary
> shapes, 'color-interpolation' is used on masking and filters as well. All
> mayor browsers and most viewers support color-interpolation on filters. I
> think the same is true for masking, but I did not verify that yet.
>

I would hope that it is not active for masking since it would force masks
to behave differently from alpha compositing.
Is it active on filters? I thought color-interpolation-filters was used
there.


>
> > If a UA implements color-correction properly, it is also unclear what
> the 'linear' version of a non-sRGB color is.
>
> This is part of the request. If both properties are not combined,
> color-correction would need to clarify that.


I'd rather see something more specific.
Maybe there could be a css property that sets the colorspace of a document.
Color-correction could then become a property that says if the object is in
that colorspace or another one.

>
> > Rik
> >
> > On Thu, Apr 4, 2013 at 4:32 PM, L. David Baron <dbaron@dbaron.org>
> wrote:
> > On Thursday 2013-04-04 16:07 -0700, Dirk Schulze wrote:
> > > In the behalf of the SVG WG, I would like to ask it if is possible to
> merge the 'color-interpolation' property with the 'color-correction'
> property. We think that they do share similarities.
> > >
> > > - 'color-interpolation' is a CSS property and should be part of a CSS
> specification anyway. CSS Color makes most sense here.
> > > - SVG has a similar problem as CSS: while SVG requires an sRGB color
> space, most (maybe all?) implementations use DeviceRGB in reality.
> > > - We think 'auto' should be added to 'color-interpolation' to
> represent this difference
> > > - The keyword 'linearRGB' would need to be specified as well, as it is
> in SVG to switch the color space to linearRGB. 'color-correction' is coming
> from the WebKit CG implementation which, in theory, is able to use
> linearRGB.
> >
> > I think this is a bad idea.
> >
> > 'color-correction' is designed to solve a legacy problem and allow
> > documents to opt in to what the spec has required all along (but
> > hasn't been implemented, at least partly due to lack of color
> > matching with plugins).  It's designed so that if we solve that
> > legacy problem (e.g., with changes to the plugin API) we can change
> > the initial value or potentially even entirely remove the property's
> > effect.  It's also designed to be the smallest possible solution to
> > the problem it's trying to solve, allowing authors to specify colors
> > in a globally-meaningful way, and thus hopefully the most likely to
> > be implemented.
> >
> > -David
> >
> > --
> > 𝄞   L. David Baron                         http://dbaron.org/   𝄂
> > 𝄢   Mozilla                           http://www.mozilla.org/   𝄂
> >
> >
>
>