Hacker News Reader

Comment by fishtoaster | Hacker News Reader

Not gonna lie, I've been waiting for that to start to seem like a good idea. :) React seems to have more staying power than previous trends, though. Off the top of my head, the frontend library supremacy* went something like:

- 2008 to 2012: Jquery

- 2012 to 2014: Backbone

- 2014 to 2016: Angular

- 2016 to present: React

And although there are serious contenders, React still seems like it has a solid lead over all others. If it lasts another couple years, it's been the dominant tool for longer than most.

* Based entirely on "what I was hearing the most about at that time," not on any rigorous methodology. There was still plenty of Jquery code being written last year and there are large, mature codebases out there in Backbone and Angular.

View on HN

> there are large, mature codebases out there in Backbone and Angular.

When it comes to large long-historied Angular codebases, I prefer the term "metastasized" over "mature," though now that we're in the 2020s we can almost use "legacy."

You know what's funny? When I was finishing a Backbone.js job (circa 2014? I can't recall) I looked at Angular and React and my conclusion, based on thought and research, was that while React looked better in the moment Angular was conceptually superior and would be the future (or the future would be something like Angular but better).

That was really a thing I thought, I posted it too and people were like "hmm yes good point."

I know I've been wrong before and since but for whatever reason that one stuck with me as a humbling moment, I had it all backwards.

Hey I was certain that the iPad would be a flop too and that one year from the release nobody would talk about it ...

I've been wrong about every US Presidential election since 2000! That takes special talent.

So next time tell us who you see win, then we'll know the opponent will win! :)

As an atheist I realize this makes absolutely zero sense, but I can't help but feel that "abusing" the power in such a way would somehow "jinx" it and cause it to backfire!

Backbone never got really popular here (Sweden). I still see some legacy apps running Angular but React took over from jQuery quite fast in 2015 or so. What I have noticed thouhg is that people who like MVC-frameworks tend to like using libraries similar to Vue and those who have more of an application approach to web apps tend to like React. Personally I think I got a bad first impression from Vue and have had a hard time trying to get to like it (I know it is popular). I think libraries such as Svelte and dev environments similar to Snowpack will gain traction the coming years. Focus on Developer Experience and doing the heavy lifting in the dev environment.

I keep wondering what everyone is going to complain about when it becomes clear React and Vue & a couple other decade old offerings are what everyone in JS land is using. The jokes about new JS frameworks being invented daily have started to fall off already. Stagnancy & ossification are setting in. Although functional components in React have, fairly recently, re-simplified/re-writeen the DX again.

Backbone was the first time there was framing at all. So long ago & so not that long ago!

The frontend world is maturing!

I remember having similar complains in the smaller rails world about a decade ago: every 6-12 months there'd be a new community-favorite gem to accomplish X, where X is any number of common needs: file upload, auth, json serialization, background jobs, etc etc. It was damn frustrating. I had the early-rails equivalent of js framework fatigue! But after a while it quieted down. I wouldn't call it stagnancy and ossification - I'd call it maturity and stabilization. :)

I was in those same trenches and I have the same scars. However... IIRC the difference between JS fatigue and magpie development with Gems (ooh! Shiny!) is that switching to the latest and greatest gems was usually painless. If you were willing to learn yet another DSL. With JS you have to throw everything away and start over with each new thing. But maybe Im just wearing ruby-tinted glasses. I actually liked the energy of the ruby community back then. But with JS I feel like it’s the blind leading the blind with the confidence of a used car salesman.

I like the term “magpie development”.

React has staying power this time, thank God. It's almost 8 years old and popularity is still increasing. Angular is dead to me though, which means big rewrites coming for me in the next few years...

The core React design pattern is nice, and the community is great, but the noise in the code.

What do you mean by noise?

Sh*t that sounds nice, but when stretched outside the simple documentation, has no purpose or meaning and ruins your life.

I’m looking at you -> { blah && }

What's the matter with that notation? I work in a code base many millions of lines large, see those every day, and think nothing of it.

This is just personal opinion: The one-way flow has made functional React components very noisy in terms of state and effects management.

For examples, here's a comparison implementing the same functionality using React hooks and Svelte: https://github.com/joshnuss/react-hooks-in-svelte

Now, there are ways to handle all this with less code in React as well (for example, MobX), but the current de facto defaults with hooks and Redux create so. much. boilerplate. for. every. little. thing.

I like and use React on almost a daily basis but I'll admit it can definitely be overly verbose sometimes (at least WidgetFactoryFactory's aren't common like in some other languages). I know next to nothing about Svelte though, I'd be interested to see how it compares to React in a complex app and what the big tradoffs are.

I think it's safe to say React wins on ecosystem but loses on verbosity w.r.t Svelte.

The noise is the code.

Play the spaces between the notes.

Old school Perl coder here - rolling my eyes... ;-)

The Pathologically Eclectic Rubbish Lister.

Talk to a sendmail v8 admin who had to hand edit sendmail.cf rewrite rules.

With or without m4?

Without m4. You'll likely speak to a gibbering mess though.

I’m triggered.

Phoenix Live View pattern is where it's at now. So much better.

you forgot prototype before jquery

And script.acoulo.us! What a duo they were. Actually I just checked and scrip.acoulou.us site is still alive, with last version from Dec 2010 :)

I use the browser extension Wappalyzer, which tries to detect the tech used in every site, and you would be surprised how many use the duo :)

Oh my god, the flashbacks!

Circa 2005 doing client work during summers off of uni in this newfangled “Ruby on Rails” (v0.8.x). I was new to all of it. Coming from Flash/Flex/AS and jumping both feet into Ruby, Rails, JS via Script.aculo.us! Great times!

Oh yeah, I'd forgotten about prototype! I also used Dojo heavily at one of my college internships. Ancient history in javascript-land. :)

Dojo[0] actually reinvented itself to be a full VDOM react-like framework, focused on enterprise

[0]https://dojo.io/

Seems they also had the good luck to have missed out on jQueryMobile too...

:shudder:

(I did make a bunch of money with that one project that went live for about 8 days before the company we did it for sold out and the new owners shut it down immediately. _So_ glad I was never on the hook for production support on that...)

mootools, anybody??

prototype was horribly named. I was just becoming aware of prototypal inherence and it was so confusing because search results mostly pointing to the prototype library.

And yahoo!

That timeline is missing vue and vanilla.js. Also, while personally I'm not a fan, Angular seems to continue to do well in "enterprise" projects where the web app isn't the main deliverable, and having a separate Node.js stack is undesired next to the Java or .net main backend stack.

You might be surprised how many sites are still using jquery..

Reply on news.ycombinator.com