Re: [css-color] vendor named color enhancement

> On Jul 24, 2016, at 01:31, Alex Cohen <AlexCohen@XRITE.com> wrote:
> 
> It’s great to see so much interest. We love where the color module level 4 is going and think it will bring color to a whole new level on the web.

Hi Alex,

I am extremely happy to see Pantone engage us on this topic, and this is a very opportune time since we have been working on colors recently.

> As a quick note, what I’m about to say is not a promise in any way, shape or form that Pantone will change it’s ways or that it will change it’s product structure, this is just a discussion, for now.
> 
> I’d like to address Pantone IP. Pantone has been, to say it lightly, a bully regarding it’s IP. Some consider it non-sensical, other consider it to be the basis of their business model. I consider it to be a bit of both. There is a growing feeling that Pantone needs to loosen their grip.
> 
> To me, Pantone is a nomenclature for color ( just like everyone knows what green looks like, a lot of people know what PANTONE 101 C looks like - it’s a yellow ). It is also a business that sells the information on how to reproduce that color. The separation of the two is very important.

The way we have designed it now as you can see in the color module level 4 involves two parts:
- the color function, which lets you pick a color either by numeric parameters or by name out of a color profile
- the @color-profile rule, which lets you import color profiles from ICC files.

We designed the color function with named colors in mind, and we certainly considered Pantone while doing that. With the syntax as it is proposed, authors would be able to write:

.logo {
  color: white;
  background: color(pantone '101 CP');
}

Note that the spec does not specifically call out pantone, this is just a normal application of the syntax we defined. "color(my-favorite-colors 'Mapple leaf 7')" is just as valid.

That's for syntax. The remaining question is how the browser gets to know how to match that name with a color definition.

> We need more than nomenclature to render a color on the web, which is why Pantone might be willing to loosen it’s Terms of use regarding the sRGB data for it’s colors. This is a discussion that we would need to have, but it’s looking good. I think sRGB will do the trick as a first step. We don’t want or expect perfect reproduction on the web, we just want a good approximation.
> 
> The real goal is to specify the color using the names, which in a sense, means so much more than the data.

Unless someone comes up with something else, how the mapping between the color name and the color definition works boils down to 2 options:

1 - Pantone names *and their corresponding colors definition* are built into browsers. If so, the rule above just works, there's no need to do anything else. There's also no delay caused by having to fetch color definitions over the network, since they're built-in. One things that is important to note here is that this approach pretty much locks down the color definitions: once the web starts depending on certain names looking a certain way, it is very hard to change things.

2 - Authors are required to profile an icc file which defines the colors they want to use. Think of how web fonts work.

@color-profile pantone {
  src: url('http://www.example.com/pantone.icc');
}

@color-profile my-brand-pantone {
  src: url('http://www.example.com/pantone-colors-used-for-my-brand.icc');
}

This has downsides, as there's now more syntax to deal with for authors, and there's a network request in the middle which could slow down things, but it may also be more flexible in terms of providing different profile files to cater for different subsets, different device values...

> Tab, If Pantone is something the W3C is interested in, I can definitely help out, I want to help out. Anything I can do, I will do.


Count me in. We (Vivliostyle) develop a CSS formatter for print, and support for pantone colors would be great.

 - Florian

Received on Tuesday, 26 July 2016 04:40:16 UTC