I was just thinking the other day how everything Meta has put out on the frontend, I felt compelled to use it because of the hype. But I've ended up being frustrated by almost all of it.
It really feels like Meta has led us on a shitty frontend path for the past 5+ years that we need to leave...
It really feels like Meta has led us on a shitty frontend path for the past 5+ years that we need to leave behind.
- Flow
- Jest
- Metro
- React
- React Native
- GraphQL
- Flux/Redux
Everything they do is just so overly complicated, and so many layers of abstraction rolled up in a messy ball of mud.
Letting such a large company lead the frontend space is a bad idea, because if they make a bad architectural decision they have to just keep rolling with it because they roll it out so widely internally, and improvements are very slow because of the business impact.
Simplicity in my book comes down to how comprehensible is your stack trace. A few lines of code calling into a giant magical function is not simple.
Comments
Wow, a legitimately unpopular opinion. Strongly disagree, I worked on complex frontends before React, and React was SUCH a breath of fresh air
By threatofrain 2023-08-0117:00 It's a legitimately unpopular opinion only because React is on that list. Take React off this list and the opinion will be stale.
I also don't know what it means for FB to be "leading" in this space because they don't push any of their other solutions on you. In fact, React being a stand-alone library may be a long-term strategic mistake as it seems the ecosystem is tending towards mega-frameworks.
By gjsman-1000 2023-08-0116:571 reply I'm at the point where I question whether almost all tools that we use for the web need to be revised. HTML, CSS, and JavaScript have served us well... but I think all of them are showing their age, and everything built above them is a hack at this point. But they are all so engrained that nobody can imagine how things might be done better...
JavaScript was obviously never intended, or designed, for 90% of the things it's used for. Also, as a language, it's so flawed that we have stuff like `left-pad` and `is-odd` on npm which is just atrocious...
CSS, and styling pages in general, has always been a mess. Either have a disaster trying to name all your classes, or abuse HTML by using a billion utility classes a la Tailwind.
HTML was obviously never made for utility classes, encourages bad practices like inline styles, was never designed for DOMs of such massive size, and the number of strange new HTML tags that don't quite do the job is quite large.
By Gualdrapo 2023-08-0117:15 > CSS, and styling pages in general, has always been a mess. Either have a disaster trying to name all your classes, or abuse HTML by using a billion utility classes a la Tailwind.
In my book this is not a problem inherent to CSS but, well, how to name classes and stuff in your HTML. At this very moment writing CSS is just amazing and many of its "new" features are what many of us have dreamed about for years.
Still I have another unpopular (and for many of you, stupid) opinion: I'd like it if HTML, CSS and JavaScript had all the same syntax. Like QML's.
By jcollins1991 2023-08-0116:57 I mostly agree, and usually feel like a grouchy old man for it. Been doing FE for a long while, jQuery into stuff like Meteor, then early Angular and Ember... There were a lot of frustrations but I found that they were more easily overcome, and past a certain point I've felt like we aren't getting any more technologically superior with reimagined frameworks, but rather just finding a new flavour of the month.
Conversely, React Native (and other similar solutions, even Electron which I love to hate) are AMAZING and I feel like it's such a breath of fresh air compared to pure native app development. They have their ups and their downs, and they're not always the best solution for a given problem, but they've brought so much opportunity for dev efficiencies within companies and allowed solutions to be built that would've otherwise been cost-ineffective in the past.
