Writing JS is its own pain

Delayed pain vs delayed gratification. All the kids use JS, all the femboys use Rust 😎.

@@anonymousalexander6005 Dude just called whole IT industry a 'kid'

It’s a food chain 😅

​@@anonymousalexander6005bro just called all the youtube, twich, microsoft, google, etc etc website developers kids

i praise him a lot

Yeah, until those tokens are specified in JSON format, e.g. from a node module. Makes it a more cumbersome for that case.

@@dealloc Well, it's a new version of the Tailwind, and this is a breaking change, so if a node module is exporting their tokens as a node module, they should update to the new version and start exporting as CSS. But yeah I agree it is a bit more inconvenient, that's why there's still support for JS config

Curious how you're supposed to do programmatic stuff then - like currently I have a function to generate all the 50-950 color variants for me. How do I do that in css file? Use sass?

@@huge_letters Depends. There are better ways to generate tokens for different formats. One solution is to conform to the Design Token spec and use tools to generate to the desired formats for you. That way you also get two-way integration with design tools like Figma, Sketch, etc. and keep them synced without tying it to specific platforms or formats. For hobby projects that may be overkill, and in such cases I'd either stick with the JS config, or just copy-paste values as-needed, since it's often overkill to spend too much time on these decisions in such cases. I would personally pre-compile colors and tokens (e.g. through a command) and not generate at runtime as it adds overhead for no reason.

I see what you did there

I remember using Emotion, Styled and Radium all in the same project. With Redux and Mobx of course. I'm older and more scar(r)ed now. Which is why I'm not using Tailwind. It's also why I'm more on the backend these days.

Radium was the best.

I actually use both, tailwindcss for most styling and if I hit a wall I just use emotion js. Not sure how "good" this is tho.

@@SpinatMixxer If it gets the job done that's pri 1. Pri 2 is whether you'd hate yourself for making a choice 5-10 years from now (abandonment, build tooling evolution, etc). That's why I'm ok with styled-components, but aprehensive about tailwind. While there is an escape hatch (just use the css), it's not ideal. Styled components or React are way safer bets (though I prefer SolidJS).

its arc browser, i tried it but for development it felt terrible

It is not impossible, it is just difficult to find examples online. Mostly because the TW Typography plugin is a completely different use case. You want to use the addUtilities function in your them to create typography classes. You can put media queries in them as well to fully abstract away the choices made by your design system on font-size, line-height, etc. based on breakpoint.

Leptos + tailwind is nice

They are in American English, too. A sentence that wraps a few lines isn’t necessarily a run on and some authors like Nathaniel Hawthorne would write correct sentences that spanned half a page.

Arc Browser, I think it's macos only for now

The new JavaScript

​@@derciojds but actually a really good language 🤣🤣

Or use Mojo. But I’m just a Chris Lattner fan

Arc's early access program. New update coming

he was using arc browser, it has built in search command palette

To each their own I suppose. Personally, I truly disdain context switching for something as minute as css

Try it, you might like it. I worked with CSS Modules for last 3-4 years and switched to Tailwind in last 6 months or so and would never go back. Syntax is nothing special, it's really not that different compared to actual CSS properties so it's not like you are learning a totally separate language. It's one thing watching videos, reading about something, but it's a totally different thing trying it out. Don't dismiss it until you evaluated it properly.

So, you have css module for every component? Isn't it better than for everyting to be in the same file?

I can relate to the initial resistance but just try it once. I was too just into modules but twcss is just awesome

They are not comparable.

Tailwind is closer to plain CSS than to Bootstrap so it mostly depends on how comfortable you are with CSS. Then again I picked it up when I was mostly using Bootstrap and Material Design and the documentation was so good I quickly picked up CSS via Tailwind. Regarding capabilities: Tailwind allows you to very quickly and easily build your own version of ‘btn’ and ‘card’ etc. It’s basically as capable as vanilla CSS

Bootstrap is so much harder and overcomplicated compared to Tailwind. Tailwind is just CSS with classes. Also yes, tailwind does everything better

@@xrexy CSS already has classes, so no. Not even close.

@@TheOriginalBlueKirby What??? CSS has classes???? What are you gonna tell me next, maybe the sky is blue?

That complexity has to reside SOMEWHERE. What better place than the markup itself.

For content. Divisive opinions generate the views. Go watch his Rust video and just cringe.

Oh boy, grab the 🍿

Easier to write plain CSS? Maybe. Easier to maintain? Fuck no.

Nope. In a FAANG codebase without scoped styling you have to have QA team regress a page every time you change you remove or override a class.

Unocss just manual tailwind

Yeah people now a days rather use over engineerd tools then actually learn the basic.

Just build on raw JS, not even TS. Go big or go home. React is over-engineered

All these hobbyist solo devs who have never worked on a team giving “advise”. 🙄

Raw React 😂

I was on the same page but then I changed my mind. Tailwin team researched and used all scientifically proven knowledge about color schemes, typography etc. to come up with the list of css classes instead od letting you use all the custom sizes and colors. You can read a lot about that on their blog. I was amazed when I read it. And even after learning that, I still didn't like it for a long time, because for me it was better to use css selector nav>li, if I wanted for example to change the style of all navigation lists on my site, instead of changing tailwind classes in all of them separately. However, the re-use problem with tailwind is not an issue if you use a framework (e.g. React) where you can make a NavList component that you can re-use accross your site by passing different props (item list, orientation etc.). So if you don't use new school JS frameworks, it is fine to stick with the old way of thinking, which is to separate content (HTML) from presentation (CSS). But if you use a framework where you define your reusable components then it is good to use Tailwind instead of changing CSS in your defined class without knowing if you broke some other component that used the same css class.

And you're not alone

Tell me you've never worked with other developers without telling me

@@xrexytell me you never worked with real developers thaty know the basics without telling me.

@@tjaytje It's not about knowing the "basics". It's about how CSS is nowhere ideal for bigger projects.

Also the inline mess argument never fails to make me laugh. With CSS you just hide the "mess" in another place behind a selector

People are arguing in the replies but let me take a real crack at showing you why people like tailwind over css classes. If i were to sum up the main benefits, Id say its for the colocation, speed of development, and its predictability Colocation: This is what everyone tends to hate tailwind for and I get it to a degree, it can 3x-10x the size of your html tags. The reason people enjoy it is because when you first get started, it feels like you have to learn a whole new way of writing css, but the tailwind team has put a ton of effort into making the process as close to writing real css as possible. Id say, if you try it, youd be able to read and reason about tailwind code in about an hour. Speed: Like I said earlier Tailwind has a TON of tooling around it such as class sorting and intelisense with really good fuzzy search in all major editors. This makes the css abbreviations really efficient mentally for me. I type flexc and flex-col will be my first result. I have memorized these patterns over time, but starting up i didnt really have to memorize it either. Ive probably checked the documentation maybe 3-5 times in my life? But since I dont have to name anything and just write the classes I want, it becomes so much faster to write and maintain. Need to turn a parent div into a flex? You are one click and four keystrokes away from that, instead of going go the correct css file and finding the classname, which you may have forgotten and changing the display to flex. Predictability: This one is my favorite. Now i would like to preface this with the idea that if you are not using a composable/component architecture, Tailwind is not the solution for you. If that was the case, i would fully agree about your points on maintainability and readability. But when you actually work on tailwind on a component, say maybe 50-100 lines of code, it becomes really easy to maintain. It benefits you a lot to right correct semantic html too but in my opinion its not necessary. When i want to change something or I see something breaking, I can usually narrow it down to a component->tag level immediately, and now since my classes are in a predictable order, I can usually predict exactly where in the codebase I need to look all before I have seen the code. It becomes a massive win in large teams for this reason. Weve all had team members write garbage code and css, and put it in the wrong file, and then put it in a weird spot where they relied on cascading and specificity, then it inevitably breaks on the next change. Those are the worst to debug. That goes away with tailwind. I completely understand your concerns about readability, as it definitely is a hurdle. I think its one to generally overcome easily though, as when your html is readable, your classnames come second to that. And if you know where you need to go in terms of the html, you find yourself at or near the line where your css that broke the html is. No need to change your mind because of this, but really try it for a quick tester app, it really makes the skill floor for good css go way up, while still achievable, making junior devs much more reliable.