Re: [css-variables] RE: CSS Variables in Last Call

n you On Tue, Sep 3, 2013 at 1:37 PM, François REMY
<francois.remy.dev@outlook.com> wrote:
>> Given that this email doesn't appear to have any concrete comments at
>> all, just a statement of general unease with something (unstated)
>> about the spec, I'm not sure what I or the group is supposed to do
>> about this.
>>
>> ~TJ
>
> To be honest, I don't know either. But the fact I still feel highly uncomfortable with the spec is not a good sign. There are too many things I think are wrong, and I'm pretty sure we should think back about the use cases we want to support and what's the best way to support them. My intuition is that this spec is the wrong solution to them.
>
> This may be totally personal, or not. This is what the group should maybe try to find out.

I can't act on your intuition that something is simply wrong, so shrug.

> On that matter, did someone take some *realworld* stylesheets, css polyfills and webcomponents-scoped styles to see how the css-variables spec would make them easier?
>
> I'm under the impression the spec took some road and continued ahead as far as it could, but the end result looks far from satisfying to me. I'm pretty sure that we could shift the lines toward what preprocessors do could make sense and at least cover properly that use case, something even preprocessors couldn't do completely since they're not dynamic on the page they live in.

Yes, "someone" did - me, and coworkers on their own personal projects.
 When I was first putting it together, a coworker who sits next to me
was just getting into a personal web project, and was very frustrated
with dealing with CSS.  To make himself feel better, he rewrote his
stylesheets with variables and nesting, to see what they'd be like in
the future, and was *much* happier with them.

I've done similar things with my own projects.

> I feel like the current draft isn't going to evolve anytime soon into something I could use for polyfilling and web components styling purposes, and a lot of complexity arise for trying to support these use cases as a backdoor. If this is just not going to work, maybe we should refocus our attention on what may end up working (pure variables and computated expressions), and eventually start something else along the lines of the web components & polyfilling scenarios at a later time. I recently
>
> It is just an idea though, what do you think?

Variables appear to be perfectly suited for web components styling
(modulo the fact that, at level 1 of the spec, dealing with default
values is a slight hassle if a variable is used multiple time).  Can
you elaborate on what you think is wrong?

Similarly, can you elaborate on what you think is wrong with using
custom properties for polyfilling?  Again, there's a level-1
limitation that it's annoying to *find* the elements with custom
properties, so you currently have to iterate through the document
looking for them, but I have plans for level 2 to make that much
easier.

This email and your OP have been *extremely* vague about what you
actually have problems with, which is not useful feedback.  I'd
appreciate something concrete. :/

~TJ

Received on Tuesday, 3 September 2013 21:38:18 UTC