I’m qualify as a developer and I’m confused by what you mean when you say delegation. It makes it harder for me to understand CSS. Variables, however, wouldn’t. As they are pretty much the simplest concept in the universe. This delegation thing seems really abstract and hard to hold in your head.

In a roundabout way, Usme Cah, I think you’ve proven my point. :) Delegation is functionally equivalent to variabilization but implemented from the perspective of a designer rather than a developer. And perspective coupled with function is the whole point, isn’t it? If we want designers to use CSS, then implementing capability from the perspective of developers and not designers is clearly self-defeating.

Variables, however, wouldn’t. As they are pretty much the simplest concept in the universe. This delegation thing seems really abstract and hard to hold in your head.

x, however, everyone learned in algebra somewhere between the ages of 12 and 17. And I was always saying I’d never use it in real life–ok, I don’t, but the concept of a token standing in for an unknown value, at least…

Plus, CSS keeps adding modules and properties. Has anyone done a proper count of all the properties likely to exist in CSS3? I know many of them are just proposals at this point, but still, several have advanced beyond that. By CSS4 aren’t we talking about a vocabulary of hundreds of properties? Have you seen the syntax of gradients? Talk about something that is hard to hold in your head.

Basing your arguments on “what other people will be able to understand” is a pretty shaky foundation. You have no clue what other people will understand. Walk up to a designer and say “variables” and, yeah, maybe they’ll be like, “what? that’s programming stuff! scary!” But say, “hey, there are two things I’ve wanted over and over in CSS. one is not having to copy and paste the hex code of a color, the other is the ability to set padding to something like line-height minus font-size. how about you?”

Casual content authors are perhaps unlikely to have written CSS on such a scale that they’d feel as rousing a need? So perhaps they’d ignore it until they needed and keep on doing things the way they always have….

Then, your proposal is awfully limited. From the syntax and examples you give, it looks like you would need to give unique ids to basically every container in the page. I can see a lot of use cases where I would need grid-like behaviour (“I want this set of elements to act as a column, this other set to act as a row with same-height cells, etc.”) but where your proposal wouldn’t help if I don’t have a unique ID for each element. It would be fine for short pages and websites that don’t change much, but for big pages and very changing websites built with many reusable components… that would be a nightmare. I also think the ASCII-art syntax AND use of the display property for grid layout is nonsense, but the scope and power of the feature itself is quite good, and much more interesting for layout than hackish use of CSS delegation. Don’t throw the feature baby with the syntax bathwater. :)

Finally, you’re suggesting a very broad feature but give only one use case as an example. Why implement such a broad feature (access to any computed value of any element, through CSS, and use for any property of any declaration block…) when the main concern is layout, and maybe only a small subset of layout needs? Sadly, that’s not very convincing.
You need more compelling use cases before you can convince a majority of CSS coders that a new language feature is needed.
(That’s true for CSS constants too, obviously.)

PS: I liked your recap of the CSS variables debate. I think Bert Bos in his insistence on simplicity. He doesn’t seem to realize that CSS is a professional tool too, not exclusively something for hobbyists and graphic designers that occasionally dabble in code. There are webdesigners who are great expert coders, and front-end developers who specialize in HTML, CSS, and JavaScript (+accessibility, front-end performance and related concerns). What’s more, fewer and fewer content creators (in proportion) are using CSS; instead, they use CMS platforms and software. So I’m more in line with what you write: IF the new feature is worth the additional complexity and a good solution to real needs, go for it. It’s not as if the complexity will be forced down the throats of hobbyists and beginners (are they forced to used complex CSS3 gradients? no, they can stick to background images or aim for simpler designs).

Do you know any project out there that would take this syntax and compile it into a mix of “current” CSS and JS code, as LESS does for variables, etc.?