I absolutely loath frontend developments in companies with more than 5 engineers. In the past you may have found some spaghetti code and various abuses of jQuery but you could probably grasp the codebase in an afternoon. Nowadays every mid company codebase is a spaceship. For sure, you'll find a pseudo technical VP of engineering which enable whatever cool technology to be used and resume driven developers happy to pile on the latest Big Tech framework - which is likely to be an over-engineered, over complicated exercise in engineer retention and engineer roles marketing.
Then you have Pieter Levels pulling a mil per year with PHP and jQuery.
My first job out of college involved a Rails codebase littered with conditionally rendered jQuery snippets. It was a nightmare to figure out what code was even loaded on any given page. Give me a modern JS codebase over that any day.
Might it be possible that those expectations, and those requirements for doing "more", are not coming from the actual users/visitors to the websites - but rather from the companies owning the websites?
Because I, as a visitor of HN, come here for the IT news and the nerd stuff. I happily put up with the layout - because the sites actually meets my expectations for content. At the same time, the best, most modern, snazziest website ever... why would I go there (more than once), if there's no content of interest to me?
And that - to my experience - holds true for almost ALL visitors. They come for the content, and they put up with the layout. And this table-based layout with spacer gifs right here, imho is a lot easier to put up with, than some desing-heavy website that's perfectly optimized for conversion, engagement and retention rates and whatnot.
I think the actual problem is, that when we talk about "expectations" and "requirements" - we talk about what the HOST/SELLER wants - not what is best for the visitor.
I don't know. I just tried to use Netlify to upload a simple static Html/Javascript page I made. Now I'm trying to figure out CORS errors. Pretty sure that didn't use to be a thing. For better and worse.
CORS will eventually click for you and then it will always make sense. Until then, sorry you’re going through the muck.
(1) Tailwind uses utility classes that feel like syntactic sugar over full CSS, so I don't feel like I'm "using Tailwind for layouts" as much as "using useful CSS grid presets"
(2) Much like with margins and paddings, text sizes, and colors, Tailwind helps me "pare down" the number of different values available to me. Much like having preset "m-1", "m-2", etc. values helps me be more consistent, having preset grid columns and gap spacing helps me stay consistent and not go too crazy.
(3) Because I'm using something closer to CSS-wrapped-in-classes rather than a set of components built by someone else, and because tailwind gives both a naming convention and an easy mechanism for adding my own classes that fit right in to Tailwind's, if Tailwind doesn't give me the exact values I want for grid spacing, then I can modify theirs. If I need to add a couple new presets in addition to Tailwind's, then I can do so easily.
(4) Early on when CSS Grid was new, I got bit by some bad layout bugs that were very hard to find, and impossible to fix (they were browser bugs). So, while I'm sure the bugs I encountered have since been fixed, I'm a bit gun shy on grids. I occasionally use Grid when it makes sense (via Tailwind), but more often than not I find it easier, more flexible, and safer to just use nested flexbox cols and rows. For which I also use Tailwind, if the project is of suitable size.
To sum it all up better, I view Tailwind more as syntactic sugar over CSS that helps me build my own design system. So, I would totally use vanilla CSS for small projects that don't need a design system. But for anything more than that, I like Tailwind for the same reason I like using web frameworks on the back end: it helps me keep things structured and consistent and more maintainable, even if I could do it all on my own without the framework.
For (2) however I usually use CSS custom properties to reach for preset values, e.g. setting "column gap" to "var(--margin-inline-wide)", and "row-gap" to "var(--margin-block-short)" etc.
I would also point out that Tailwind only makes sense to me when I'm also breaking up each individual UI element like cards or buttons into reusable components, either using a reactive framework like React or Vue, or a server-side template system. I'm not sure the maintainability of Tailwind survives with more monolithic HTML pages.
Easier to remember and quicker to type. Centering content on a screen is class="place-content-center" which makes things easier.
.container {
grid-template:
"nav . header header header" auto
"nav . . . . " 1ex
"nav . main . aside " 1fr
". . . . . " 1em
"footer footer footer footer footer" auto
/ 40ch 1em 1fr 1ex 15ch
}
The thing is you can still do all of this. But we have since built tools and frameworks to let you still do this, while giving you superpowers. I think Svelte is the best example of this. You can create a dead simple site with it and have the result you are talking about, but the real power is now you can employ many of those advanced techniques that were reserved for large web apps in the same dead simple site.
> Is tailwind really easier to work with something like - let's say - pico.css? Of course if you want to do super-elaborate layouts, then probably yes - but if you embraced a more minimalist approach?
I've used tailwind quite a bit, it's killer feature is the fact that it's just intuitive. With most CSS frameworks, even minimal ones, I always refer back to the docs to see how they do a specific thing. With Tailwind you refer to the guide a few times but very quickly pick up how it does things. I think that's why it's so popular, I've typed out classes thinking "is this a thing", and almost every single time, it was.