[commanderkeen08]

Every time Tailwind does something great and gets posted here, the conversation devolves into the same arguments: “I don’t get it. It’s just inline styles.”

What is it about CSS that gets people so offended and opinionated? If it were a new JS framework, few people would be saying “I just don’t get XYZ. Use React”. Is it because Tailwind is so drastically different and breaks people’s core ideas about separation of concerns? Maybe it’s because we all learned CSS very early and were told to do it XYZ and now challenging that is painful.

If you honestly want to “get” Tailwind, go use it in a project. If you don’t like it, don’t use it. Nobody is going to change your mind in a comment and you’ll never convince anyone to stop using it.

Idk. CSS is what I least care about. It’s a thing. I use it to do a thing. And I move on. I used SASS. Now I don’t.

[jastr]

When you first hear Tailwind's concepts, they seem to contradict everything you know about good software design. When you try it out and begin to remember some of the class names, you get into a great state of flow - it's the high developers are always chasing.

I can't help but feel that Tailwind detractors have never actually tried Tailwind or are too square to give it a chance.

[alphabet9000]

regarding 'contraditing everything you know about good design' i feel this way about designing webpages using <table> elements. the constraint really forces you to be creative with using image elements to make everything fit on the page just right. styling with tables used to be popular, but then it went out of fashion, unfortunately. i'd like to see it make a comeback. i'm not being sarcastic either.

[iovrthoughtthis]

i think we too easily fall for the heuristic of separate files == separate concerns

but your concerns arent really separate if, while in their separate files, they are concerned about the same things

and, so long as you want relative styling, they always will be. you can either (1) attach properties to your tree or (2) flatten your tree into a list of paths and attach properties to nodes that match the paths

in (1), changing the structure can only impact an individual node or it's children, only by the properties attached to the node and it's parents; changing the style can also only impact an individual node or its children

in (2), changing structure can impact the individual node, it's children, siblings or parents by any rules that now match any of the mentioned nodes; changing style will impact all matching nodes (not all of whom can be identified statically) and their children and may interact with other rules which apply properties to matching nodes.

either way, you cant escape the structure of the tree and one strategy lets you scope your changes easily, the other does not.

[winstonewert]

For me, the situation is: I gave tailwind a try. I hated it. I spent a lot of effort of ripping every last tailwind class out of my project.

And yet, a lot of people swear by it and think its great. This really confuses me, and I'd kind of like to understand: how can other people like this thing that I think is terrible?

[aniforprez]

> how can other people like this thing that I think is terrible

Cause it works for them? You cannot be the arbiter of what other people like or don't

I hate CSS and I find the TW classes being right there with the HTML more helpful than class-hunting through a bunch of CSS files. React solves that somewhat with styled components now. I like having design guidelines set loosely about things rather than writing reams of CSS myself. I like having media queries defined right there in the HTML. I LOVE the flexibility it offers me and how quickly I can iterate through concepts and styles

[johnfn]

Yeah, it flies in the face of conventional wisdom and I think that bothers a lot of people. It’d be like if a new JS library came out that said you should make every variable a global.

And just for the record I like Tailwind :)