I think it's fair that some people have stronger feelings about and more appetite for discussing a topic than I do.
To pick one of her examples, I've done a couple of toy projects with Tailwind. I think it's pretty neat, and I don't mind saying so! But I accept that people working full time on huge Tailwind projects at work probably care way more deeply about the topic than I do, are much more impacted by aspects of Tailwind that might be trivial to me in my toy use of it, and are going to invest an amount of time and passion discussing it that would frankly bore me to death. If I were running into that repeatedly, I would stop bringing Tailwind into discussions that I didn't want to be about Tailwind, and I wouldn't feel that bad about it.
Sure, but on the other hand people can learn to read the room. I am interested in lots of things, but that doesn't stop me from trying to fit the context and the interests of the other participants. E.g., I have a lot to say about cryptocurrency, but there was a time when it seemed like almost any HN discussion ended up with a tenuous cryptocurrency tangent. It was annoying and exhausting, and I'm glad we've mostly gotten past it.
I wonder if we've reached that point for LLMs/AI/ML yet. Sometimes it definitely seems like the discussion on HN is very influenced by the VC fad of the day.
You know, that's interesting. I tend to avoid Tailwind discussion because I feel like I don't see enough opinions from people working full time on huge projects using it. Most of the time, it seems like the strongest opinions in favor come from folks who have only used it in toy apps, side projects, or completely greenfield small-to-medium applications.
There's nothing wrong with using it in that way, of course, or of thinking it's the best tool for the job in those scenarios. But for some reason I rarely see that caveat mentioned, and the idea of refactoring a legacy application to use it makes me queasy haha.
> But for some reason I rarely see that caveat mentioned, and the idea of refactoring a legacy application to use it makes me queasy haha.
Easiest way to check is to just add tailwind with PostCSS and the `@tailwind` directives and see what happens to your app. Unfortunately, if there's substantial CSS it will most likely be a disaster because the Tailwind reset will clobber the assumptions made by the previous CSS. It can be really hard to refactor custom styles when you can't see how they were supposed to look without running a separate older commit.
You can try disabling the CSS reset tailwind adds but then you'll have to chase bugs in the assumptions Tailwind makes so it's a bit of a catch 22. If it's a really large project with dozens of developers in its history, chances are there is already a custom CSS utility framework in place that uses the obvious formats like "p-2"/"px-2" for padding (with a different scale, yay!) and other naming collisions. You can however use the `@apply` directive to inline the Tailwind styles in the legacy classes to make it easier.
To pick one of her examples, I've done a couple of toy projects with Tailwind. I think it's pretty neat, and I don't mind saying so! But I accept that people working full time on huge Tailwind projects at work probably care way more deeply about the topic than I do, are much more impacted by aspects of Tailwind that might be trivial to me in my toy use of it, and are going to invest an amount of time and passion discussing it that would frankly bore me to death. If I were running into that repeatedly, I would stop bringing Tailwind into discussions that I didn't want to be about Tailwind, and I wouldn't feel that bad about it.