Comment by rglover | Hacker News Reader

Comment by rglover | Hacker News Reader

This is a symptom of a lot of developers thinking they have to write code exactly like everyone else (or at least, strictly adhere to "best practices"). It's a very subtle disease, but I've noticed it again and again over the years.

Reading between the lines, this is a criticism of hooks if they're viewed as a wholesale replacement for classes; from experience I'd argue they're not—they're just a convenient tool for simplifying common patterns. I'd imagine the author knows that to be the case and instead of just using classes where appropriate (or where they wanted), they had to rationalize using hooks because of the aforementioned "but everybody else is using hooks" problem.

I suffered from this behavior for years before I realized it was impeding my work. The term that came to mind for the phenomenon was "The Invisible Developer:" a non-existent developer sitting over your shoulder always judging you for your programming choices. That developer doesn't exist. If instead how "in fashion" your code is is the standard on your team: you're on the wrong team.

That's not a developers fault, that is the insane JS ecosystem. Every year, everything old breaks and everything new is different. Of course no one can keep up, develop intuition and good practices, so all that's left to do is copy. That is how learning works, only the curve never flattens (and nothing is really learned, nothing is gained).

Nothing old broke with the introduction of hooks.

This is true, but even as a fan of hooks, I must admit that you'd be fighting a losing battle if you tried to convince your coworkers to merge your PRs with classes in them, in most places.

That is a problem with your coworkers, certainly.

That's just the illusion that's created by the inexperience of most JavaScript developers. You don't have to continually learn the bleeding edge stuff. What they don't tell you (arguably, because it doesn't pay their bills) is that what you should learn is the language. Once you understand that, you can comfortably adapt to any new library/framework/whatchamacallit relatively quickly.

Most of what you've described is just peer pressure. The only reason I know that is that I shared this exact opinion until I said "this is stupid" and started doing things my own way (to much success and peace).

Learning the language is one part, but what enables you to make an application are patterns and paradigms: inflating, mvc, mvvm, immutability, functional. These work for all languages, you learn one once, you can apply it everywhere. It just so happens that maybe every 10-20 years a paradigm replaces another. Since i started writing code in the 90s this happened maybe three times. As far as i am concerned all this ever-changing js fatigue talk is complete rubbish. Even the tooling, we're using the same tools that we used 5 years ago (Webpack, Babel, TS, React, Jest, ...). Microsoft/Apple change their platform toolkits quicker than that!

Class lifecycles being replaced by hooks is a radical but much needed change. Learning it isn't complicated and it will enable you to venture on as everything around you sheds the class based component approach. Swiftui, Vue, Svelte and ofc React are just the beginning.

> Once you understand that, you can comfortably adapt to any new library/framework/whatchamacallit relatively quickly.

I disagree. If you know typescript, it's still quite a leap to be able to code with angular, and debug weird stuff where you stare at a blank page, and the back trace in the dev tools include not even a single line of code you wrote.

Also, the language isn't really enough. You need to be familiar with the tooling (npm/yarn, babel, whatever) to use most libraries; so unless you do everything from scratch, you have to deal at least with the churn of the toolchain.

>Every year, everything old breaks and everything new is different.

It's great for online educators and others who monetize devs though..

The JS ecosystem almost seems like it's self promotional, where the monetization opportunities result in a higher self marketing activity of individuals on platforms like Twitter, which of course builds up hype and draws more people into it.

I get so tired of the churn claim. Let's look over some stuff

ES6 -- 2015

jQuery -- 2006

React -- 2013

React Native -- 2015

Redux -- 2015

VueJS -- 2013

6to5 (now Babel) -- 2015

Webpack -- 2012

ESLint -- 2013

jslint -- 2002

ExpressJS -- 2010

Lodash -- 2012

UnderscoreJS (basically the same API as Lodash) -- 2009

D3 -- 2011

Blockly -- 2012

Node Red -- 2013?? (I'm not sure when IBM first released the project)

Bootstrap -- 2011

PDF.js -- 2011

WinJS -- 2012

SocketIO -- 2012

ThreeJS -- 2010

dotenv -- 2015

momentJS -- 2011

node-sass -- 2012

Winston -- 2014

yargs -- 2011

Modernizr -- 2009

Jasmine -- 2010

Mocha -- 2011

Jest -- 2015

All of these are at least 5 year old and most are 8-9 years old. Look over the most downloaded npm packages and you'll find this to be generally true for all the top packages.

angular.js -- 2010 - 2016

angular -- 2016

That's an example where, when you jumped on the bandwagon at the height of the hype, you were thrown into a backwards incompatible rewrite pretty quickly.

And while npm has been around for quite some time, my outsider impression is that was, for some time, basically discouraged from use, because yarn was so much better.

Compare that to Ruby on Rails -- 2004, Catalyst (Perl) -- 2005, Django (Python) 2003.

Time moves slower in enterprises, with upgrade cycles more on the scale of 3 to 10 years.

Angular 2 shared a name for marketing reasons, but was/is 100% unrelated as every major idea in Angular turned out to be bad.

Yarn solved some npm issues they refused to tackle, but npm got motivated and has now solved most of them. Meanwhile, yarn 2's release has been a disaster and has seemingly resulted in most people (myself included) moving back to npm.

I've done Rails and Django work. Rails in particular took years to migrate to the next version. I'd add that Rails has seen 6 versions in 16 years which gives around 1 major breaking upgrade every 3 years (in truth, there's been about 1 major change per year as point releases are pretty big). Likewise, wikipedia gives almost 20 major django updates (I haven't used it in several years, but point updates used to have a very big risk of breaking the project I was working on).

> Angular 2 shared a name for marketing reasons, but was/is 100% unrelated as every major idea in Angular turned out to be bad.

Be that as it may, Angular 2 meant nobody was willing to commit to any fixed time span to support Angular.js, so it basically forced everybody to migrate off of Angular.js, introducing the toil and churn we were discussing.

Quite a few of these packages have undergone significant backward-incompatible changes over the years and are intertwined with specific version ranges of other packages. The churn complaints are caused by the cadence of this happening and how difficult it is to stay on a combination of stuff that actually works together. Babel may have been released in 2015 but nothing written back then would work in babel today. Same with webpack. I can't write something and then come back in a year and expect even an ounce of stability in my project. It's good that things keep improving but it's also exhausting having so much time spent performing upgrades and re-learning best practices in tools you were previously comfortable with.

There is a real pressure from the ecosystem though. When looking for a library there is good chance that anything recent is written with hooks in mind, and the examples are primarily hook based and the issue tracker is full of questions about hooks... so while you can try to alter your own behavior it feels like you're swimming against the tide.

See my reply to _stefan above (responding to the need to use something out of necessity for lack of alternative).

In respect to swimming against the tide, this is where experience comes in and knowing both how and _why_ something works the way it does. Over time, you can develop your own opinions and not have to worry that much about what everyone else is up to.

Admittedly because I mostly work alone that's a privileged POV of sorts, but I think teams can "act as an individual" and adopt a similar attitude (scale of team permitting).

I'm an Android developer who despises Kotlin, reactive programming, and "the Google way" in general. My main gripe with all this is that it needlessly complicates everything that can already be done in Java with vanilla SDK. All these additions are touted as a way to write fewer code and do it faster, but the problem is, it's all rather complex under the hood and requires you to hold a lot of context in your head at all times. And if you want to read someone else's code, you gotta have the encyclopedic knowledge of Kotlin, because characters are apparently a dollar each and it's not like we can spare some to write code with actual words people can search for. And to make it even worse, "the Google way" is a moving target. It's constantly evolving just for the sake of change.

In the end, I'm sticking with Java and vanilla SDK because it just works and only changes ever so slightly with each major OS release. Also my apps run pretty darn fast on 7-year-old devices because they only use the CPU to do useful work.

Good example, React is to web what Kotlin Android is to Android, over complicated and bloatware.

I find Kotlin more readable than Java in most cases -- less noise to drown out the actual logic. Could you provide an example of a language feature which requires you to "hold context in your head"? It's possible that you have just worked in bad codebases abusing Kotlin sugar. Inferred types for example can improve readability, but only if the variables are named appropriately.

> Could you provide an example of a language feature which requires you to "hold context in your head"?

Extensions. Basically every project has its own dialect of Kotlin that you have to know everything about if you're to understand anything. And good luck doing that without an IDE.

> It's possible that you have just worked in bad codebases abusing Kotlin sugar.

I've never actually worked with Kotlin, I've just seen many examples of it. The problem is, the language itself pushes the developer to use all the sugar, and they have to control that themselves. As opposed to Java, where writing unreadable code would take extra effort because of how dumb and simple the language itself is.

Also, to me it feels like an abstraction on top of Java that gets in my way as opposed to being helpful.

I love dumb programming languages.

Exactly. I wrote this in another comment but I've been following a Github issue about hook support in Flutter, and I understood React hooks much better after reading that than I ever did before. It is clear now what exactly the problems hooks solve are, composable life-cycle state reuse. I think that because most people don't really understand why hooks were even needed, they can't wrap their minds around them.

https://github.com/flutter/flutter/issues/51752

I, for one, do not mind developers adhering to some common patterns/best practices. Whenever I open a codebase foreign to me, I bank on those patterns to understand the app to be able to contribute, learn, etc. Seeing new patterns would be helpful to myself, of course. But, there's a good chance I would bail out if I don't have a huge need learning that codebase.

Hooks do draw you in the direction of a wholesale replacement just because of their incompatibility with classes. I wouldn't consider that to be a bad thing but it does take some time to adjust to the change at first.

Well, code reviewers can choose to enforce best practices, so you kind of do have to write code in a certain way according to them.

Reply on news.ycombinator.com