5 ms·

Technology choices, and life's choices in general are like a spectrogram of different wavelengths with positive and negative values... Like "this will help you

Technology choices, and life's choices in general are like a spectrogram of different wavelengths with positive and negative values... Like "this will help you today, but bite you in the rear tomorrow", or "this will help you tomorrow, but be annoying today". And it also varies by different conditional developments, if this happens, if that happens, can you implement a change, or the solution is "write-only" and so on. One of the tests of intelligence is can you defer immediate gratification, so you can reap higher rewards in time that compound themselves, or you lack impulse control and therefore you essentially have the model behavior of a house fly. This incessant focus on "here, now, me" that I see glorified, which doesn't consider at all anything outside a point in time and space that demands the impulse to be immediately satisfied... is not a good direction for tech, or for society at all. It's like choosing porn over a family. Obviously, I'm describing a broader problem here, not just some CSS tool, but Tailwind and many of the "frameworks" we see pushed in the industry do also embody this philosophy, quite clearly. Many things do these days. Ship faster, do less, faster, faster, faster. To what end? Are your projects so worthless, that you need to ship them today and then leave hell to yourself and others few months down the line? Not to mention that all the claims Tailwind is "faster", somehow, are written in weasel language to hide the fact Tailwind adds tangible, often significant lag and load to people's devices when they use your site. Does this matter? Of course, a fancy experiment is all good. You won't see me writing long posts about how Brainfuck is the end of humanity. But Tailwind has a good size community that actually believes their approach is superior, they're on a mission from God, and it's become viral, dragging in more and more people in, and influencing other projects into the same stupidity. What else does it mean to write a post called "Why Tailwind won". Won what? Was there a war going on? Won over what? Bootstrap? Bootstrap is a framework. Won over CSS? It runs on CSS, it's not an "alternative to" CSS, although it sure makes it impossible to use most of the benefits of CSS in a sane way. Yay, what a win. If Brainfuck was totally serious, and you saw your company is starting to write critical projects in Brainfuck, wouldn't that give you a serious pause?

jonwinstanley 3y ago
What makes you think tailwind would be hell in the future though? It's just a load of classes on html elements. Maintenance is pretty easy. I do agree with you on the headline though, of course tailwind hasn't won anything!
- 3cats-in-a-coat 3y ago
  First, the site uses some weasel language to speak about how much they're "focused on performance" in such a way as you think Tailwind is literally faster than doing normal CSS. It's not. It's a heavy approach and while you can always find edge cases to the contrary, in general it leads to pages that add lag and load to people's devices. If you have a very simple page this doesn't matter, but as your site/business grows, it does. Second, Tailwind obviously promotes a ton of repetition in the same way that using presentational attributes from the 90s and style="" does. This not only adds, again, burden to the CSS engine and DOM, but this "ease of copy pasting" that the author boasts about in the linked blog means you can easily end up with tons of subtle or not so subtle inconsistencies in your site's code and components, UI. Copy pasting is easy. But there's a reason why we define classes, functions, modules, and reuse them, instead of copy-pasting our code everywhere. Copy pasting is a strategy for a quick start and a very short journey as you paste more and more variations of what you're trying to do around your site. Now change them all consistently. Oh shit, I guess... find/replace + regex? This works sometimes, but it's a deeply unserious approach to change management in an application/site at scale, and means your velocity and quality both will suffer. I'm aware of the way to define custom classes as sort of "components" in Tailwind, but at this point... you're literally using Tailwind as you would CSS, but with a ton of indirection, gibberish, and very very unclear precedence/specificity rules. You will also get many people complaining about compilation speed for larger sites. A problem that is quite bizarre to have for styling with a language that doesn't require compilation when you use it in a sane way (and SASS etc. are extremely fast and don't have this problem, there's even a Rust SASS parser I believe, that's even faster). Tailwind also severely restricts your precision and expression. I probably don't have to explain the limitations of inline styles when those styles have to change from user interaction, or for different devices, and container size and so on. Tailwind makes all this extremely bizarre, if it addresses it at all. In "plain" CSS (and other more reasonable tools like SASS), it's trivial.
  - TOGoS 3y ago
    I would donate more upvotes to this multi-part tirade if I could. I haven't used Tailwind, or any other CSS framework, because I have never understood why I would want to, having attended a talk about a similar system back in the day where styles were turned into weirdly specific CSS classes, as if they were trying to reinvent font tags. The way you've described it is pretty much what I would expect from such a thing. And yet it's oddly familiar, because all of the 'MVC' frameworks that I've been forced to use and hated[1] seemed to come from the same design philosophy: Throw a bunch of shit code together, then build a cult around it so that people use it whether it makes any sense or not. Laravel's a great example of this because if you search the web for it you'll mostly get results saying that it's "elegant" and "for web artisans", but none of them can explain why it's better than...not using it. They even have some class in there called "Illuminate", which is ironically not very illuminating w.r.t. what the heck it is or does. As if writing code that actually makes any damn sense is everyone's biggest fear and the solution is to acquire a magic wand from one of the Framework Gods that does away with pesky old logic and predictability. "What do you want to do, write everything from scratch?", your coworkers will say, as if we were making computer programs by banging rocks together before Laravel/Rails/Spring/Django/etc came along, and give you a blank stare if you try to explain that writing regular old well-factored OO/functional code is a thing you can do and that slathering culty frameworky paste all over everything doesn't make life any better or your programs any faster, even if said goo has "good documentation", by which they mean lots of activity on Stack Overflow because nobody can figure out how they're supposed to do anything without copy-pasting from some example that mysteriously worked for somebody one time. [1] I don't remember much about the ones I didn't hate, because they did very little and stayed out of the way.
overflow897 3y ago
"Ship faster, do less, faster, faster, faster. To what end?" You're on hacker news, which is pretty heavily focused on startups. Startups usually have a limited runway, wherein if they don't deliver they cease to exist. That seems like a pretty reason to ship faster? Comparing "choose a framework that saves developer time and energy" to "choose porn over your family" seems a bit much.

jonwinstanley 3y ago

What makes you think tailwind would be hell in the future though? It's just a load of classes on html elements. Maintenance is pretty easy. I do agree with you on the headline though, of course tailwind hasn't won anything!

3cats-in-a-coat 3y ago
First, the site uses some weasel language to speak about how much they're "focused on performance" in such a way as you think Tailwind is literally faster than doing normal CSS. It's not. It's a heavy approach and while you can always find edge cases to the contrary, in general it leads to pages that add lag and load to people's devices. If you have a very simple page this doesn't matter, but as your site/business grows, it does. Second, Tailwind obviously promotes a ton of repetition in the same way that using presentational attributes from the 90s and style="" does. This not only adds, again, burden to the CSS engine and DOM, but this "ease of copy pasting" that the author boasts about in the linked blog means you can easily end up with tons of subtle or not so subtle inconsistencies in your site's code and components, UI. Copy pasting is easy. But there's a reason why we define classes, functions, modules, and reuse them, instead of copy-pasting our code everywhere. Copy pasting is a strategy for a quick start and a very short journey as you paste more and more variations of what you're trying to do around your site. Now change them all consistently. Oh shit, I guess... find/replace + regex? This works sometimes, but it's a deeply unserious approach to change management in an application/site at scale, and means your velocity and quality both will suffer. I'm aware of the way to define custom classes as sort of "components" in Tailwind, but at this point... you're literally using Tailwind as you would CSS, but with a ton of indirection, gibberish, and very very unclear precedence/specificity rules. You will also get many people complaining about compilation speed for larger sites. A problem that is quite bizarre to have for styling with a language that doesn't require compilation when you use it in a sane way (and SASS etc. are extremely fast and don't have this problem, there's even a Rust SASS parser I believe, that's even faster). Tailwind also severely restricts your precision and expression. I probably don't have to explain the limitations of inline styles when those styles have to change from user interaction, or for different devices, and container size and so on. Tailwind makes all this extremely bizarre, if it addresses it at all. In "plain" CSS (and other more reasonable tools like SASS), it's trivial.
- TOGoS 3y ago
  I would donate more upvotes to this multi-part tirade if I could. I haven't used Tailwind, or any other CSS framework, because I have never understood why I would want to, having attended a talk about a similar system back in the day where styles were turned into weirdly specific CSS classes, as if they were trying to reinvent font tags. The way you've described it is pretty much what I would expect from such a thing. And yet it's oddly familiar, because all of the 'MVC' frameworks that I've been forced to use and hated[1] seemed to come from the same design philosophy: Throw a bunch of shit code together, then build a cult around it so that people use it whether it makes any sense or not. Laravel's a great example of this because if you search the web for it you'll mostly get results saying that it's "elegant" and "for web artisans", but none of them can explain why it's better than...not using it. They even have some class in there called "Illuminate", which is ironically not very illuminating w.r.t. what the heck it is or does. As if writing code that actually makes any damn sense is everyone's biggest fear and the solution is to acquire a magic wand from one of the Framework Gods that does away with pesky old logic and predictability. "What do you want to do, write everything from scratch?", your coworkers will say, as if we were making computer programs by banging rocks together before Laravel/Rails/Spring/Django/etc came along, and give you a blank stare if you try to explain that writing regular old well-factored OO/functional code is a thing you can do and that slathering culty frameworky paste all over everything doesn't make life any better or your programs any faster, even if said goo has "good documentation", by which they mean lots of activity on Stack Overflow because nobody can figure out how they're supposed to do anything without copy-pasting from some example that mysteriously worked for somebody one time. [1] I don't remember much about the ones I didn't hate, because they did very little and stayed out of the way.

overflow897 3y ago

"Ship faster, do less, faster, faster, faster. To what end?" You're on hacker news, which is pretty heavily focused on startups. Startups usually have a limited runway, wherein if they don't deliver they cease to exist. That seems like a pretty reason to ship faster? Comparing "choose a framework that saves developer time and energy" to "choose porn over your family" seems a bit much.