We went down a similar path (using Hotwire, non the less!) and got to many of the same conclusions. It is impressive to see what CSS does and how many of the headaches are gone and :has is very powerful.

Alas, after several months we tore it all out and went back to React+Tailwind.

We still use native HTML popovers and :has selectors and other things we've have learned.

But writing UIs across three files (template, stimulus controllers, css) is such a tremendous bore. Concepts that belong together are spread out and I needed to be diligent with placing attributes and classes and remember to remove them all when removing functionality again. Obviously no compile-time checks, just magic strings and runtime errors. The Hotwire docs were also surprisingly hard to work with. All in all a lot of friction.

This just was not worth it.

inb4 rage, it is possible to use React for the UI alone and pass in fully formed view models, use form submissions and links.

>let me be clear: there is nothing wrong with Tailwind

But there is.

Yeah think it’s best to push html css js to the limit & avoid frameworks unless obviously needed. They’re just such a rabbit hole of complexity fragility and lately supply chain risk

I mostly agree. I prefer vanilla CSS to frameworks but I find a lot of value in CSS modules.

In particular I use:

1. The `composes` feature to do something like base classes.

2. The import feature to get something like namespacing.

Given the recent advancements in CSS, I won’t be surprised if they eventually build something like `composes` into the base language.

The namespacing seems more like an artifact of how you package your website and how you stay organized within your package system (I use webpack).

I like the idea of using vanilla CSS for my personal website but, not being a designer, making something that looks good from nothing is difficult. I've looked at some templates to get started with but they are generally a mess of a million classes I don't need

I used to be like this, until I realized Tailwind makes me faster.

In the context of determining responsive layout:

> Using characters as the unit of measure ensures that we get the right behavior no matter which device you’re using and in a number of other scenarios such as multitasking on iPad or even if you simply enlarge the font size past a certain point.

Nice idea! I haven't seen this before.

The section “Utility Classes: Yes, They Still Exist” is unfair in its comparison of Tailwind with pure CSS. It doesn’t take into account Tailwind’s recommendation in https://tailwindcss.com/docs/styling-with-utility-classes#us... that “if you need to reuse some styles across multiple files, the best strategy is to create a component” in your front-end framework or templating language. So its example of a “typical Tailwind component” is incomplete.

A better comparison would use, for example, a React component:

  function Button({ children }) {
    return (
      <button
        className="inline-flex items-center gap-2 px-4 py-2 rounded-full
                   border border-gray-300 bg-white text-gray-900
                   hover:bg-gray-50 focus:ring-2 focus:ring-blue-500"
      >
        {children}
      </button>
    );
  }
  
  // Usage:
  <Button>Save</Button>

This would counter all of the article’s arguments in favor of pure CSS. If the website used a `Button` component like this, it would also be true that the “HTML stays readable”, that “changes cascade”, that “variants compose”, and that “media queries live with components”.

A better argument against Tailwind would be the added complexity of having a build system and a front-end framework or templating language, if your project doesn’t already have those for other reasons.

(adapted from my better-formatted comment at https://lobste.rs/c/oznzzj)