It's hilarious. We live in a time where you can choose whatever library or framework you prefer. I don't like Flutter but I won't go online and rant about it, cause some people find it useful for their needs.
I think you've answered your own question. The "new thing" 10 years from now is a new version of React, because of how it constantly evolve itself towards something better. People will only move away if React stops innovating and adapting, then the replacement will be a new thing that's willing to adapt. I've been using React in production since 2015 and went through all the migration, so I actually appreciate all the updates :)
Chris my boi! In SolidJS there's no need to deal with stale closures or useEffect infinite loop headaches, plus handles global state out of the box, just put createStore hook inside it's own file or outside of component (in SolidJS, hooks don't need to be declared inside components). One last thing, JSX elements are just native elements so if you console.log(hi) it will actually log native div element. In answer to your question at the end of the video, I think React has another 5-10 years.
Solid makes React look like when there's a good, well written book, and then it get's adapted and the history being told takes several weird moves hehe
I don't know why people keep invoking the useEffect loop as an issue. I mean there's a thousand issues with React but this isn't one. If you're getting loops caused by useEffect the problem is not the hook but you who didn't understand how it works or how reference equality works. I haven't had that issue since the first time it happened to me because I actually took my time to read how it works and understand it. It's really simple
React code is unreadable. Complete mess. I can't believe that Svelte or Vue are not in top when it comes to jobs. Framework should be simple to learn and use. "It's not real Javascript" argument is stupid, no one cares. Just make product in simple way.. React is preventing us from that. Great video.
I’m a newbie developer and I have not learned React. Instead I looked at my options and decided to use Svelte(kit). The only downside is that it will take longer to find a job. But meantime I’m building basic projects with Sveltekit for freelance clients whilst working my finance job, no regrets so far. My friend just landed a job with React and the over complexity for the simplest things is throwing him off. I showed him some svelte over the weekend and his reply was ‘fuck you’ :)
I did the same. It feels like react is exactly a mirror of the human behavior towards "an issue/error" in real life. Instead of seeking simplicity and seeking the root of the issue, it implement more layers to the issue to "fix" something. It's what you can see exactly in psychology. But like you i'm a junior dev (in my 30's) so i'm new in mostly anything towards that. and i'm like you (haha) it does take longer to find a job with Svelte.
Why not learn React and get some work experience while learning your preferred framework? In my opinion, it could be easier for you to get that new job with that framework you like with some work experience under your belt.
@@jmse37 I was pondering how to write this without sounding bigheaded or just plain ignorant, but i'll just go for it. Overall, thanks for the comment as I think for the most part, for most people, you are 100% correct. I just don't think I can apply that for myself (and you're not the first person to suggest just learning React). Ultimately React is fine to use and I could land a job faster, you're right. It would probably take me 2-3 weeks to get used to it to an OK level. But I have always benefited from doing things my own way, it helped me stand out. I got a finance analyst job because of sales & graphic design experience (plus finance knowledge) as an example. I usually find a way to mix my skills and knowledge when i apply for very specific jobs. So again, you're right generally speaking. I just never personally benefited from copying/following what other people did, it does not help me stand out and usually slows me down. And my ultimate goal is to build a small agency that does things its own way but lots to learn till then. I'll definitely land a job, at worst I'll pick up Vue as NuxtJS 3 is looking very interesting too!
@@semikolondev We think alike! Whether we're right or not I ultimately think that whatever works for you as an individual is best to do. Following the masses never worked for me personally. I just find myself getting annoyed at everything over time. I also think Svelte will pick up. Don't think Vercel would have hired Rich Harris to work on it full time if they didn't see potential. Building everything faster and having more readable code is amazing for me personally. OF COURSE I wish Svelte had a bigger ecosystem (lets say TailwindUI for Svelte) but that will all come with time. I don't think Svelte is going anywhere but up. And React is amazing in its own right anyway, it paved the way for things like Vue and Svelte. All of these frameworks learned from each other after all :). The friend i mentioned who is learning React is just weirded out by how complex some simple things are in NextJS. I was too when he was showing me and we were comparing by building identical components in React and Svelte. Fun little exercise that was :)
Exactly. If you like Vue the natural way in the future is Svelte, if you like React you go to SolidJs. They both are excellent (I personally prefer solid because I have always liked react). Solid builds upon the foundation of React and smooth its pain points. IMHO, we components is a dead fish
@@victorlongon interestingly enough I don't enjoy writing Vue at all but have had a good time writing Svelte so far. Haven't tried SolidJS yet, but if it's more geared towards the React side of things I'm really curious to check it out.
The problem with web components is that it only addresses one issue, namely encapsulation. SSR is out of the scope, API is needlessly complex, doesn't address state management and data binding, etc. I ditched React for Vue3 mostly because of its reactive system and composable model (vs. Vue2), but having said that I'm quite confident that Svelte and SvelteKit will rule this and next decade. Why? This: let count = 1; const inc = () => count += 1; {count} inc .count { font-weight: bold; } It just CAN NOT get any simpler. Do that in React, Vue, Angular or WebComponents. When compared to anything Svelte is just superior in every aspect.
To be honest, you just boiled down everything that is to be said. Most of my projects are just written in Svelte, Svelte-Kit. Because it just make a lot of sense with it simplicity. Am not batching any frameworks here, but there is a saying that: Svelte is fun, Vue is all rounded, and React just works.😄
I kinda preffer this: It's Solid, and the state can be anywhere, it uses JSX which makes it really easy to learn if you come from React and you don't get "why the f do I need to write so much extra code just to read a value?" moments, add to that the fact that is probably unbeatable in performance and size, yes, Svelte is small and performant too but Solid is small and performant even when there'sm a big ammount of components and Svelte can't handle that too well
@@elultimopujilense It does, but Svelte made it even simpler (no reference vs. reactive objects), TS integration is far better when defining component properties and slots. Also templating is a bit more powerful with "awaited" keyword, so there's no need to wrap things inside AsyncComponents. Exposing properties is also simpler (you just "export" them)...and probably a lot more (I'm not Svelte expert). No VDOM. All in all, it's Vue3 on steroids and better TS integration which is very crucial for me. This is where Vue3 is probably the worst option, TS support. While it's way better then Vue2, it's still behind Svelte and especially React which has great support for TS (since it has no templating and it's all function anyways).
I've only been writing React for about 3 years. And the answer is yes. Also because you touched on it, it blows my mind how terrible their docs are. Like this is React. Made by Facebook. And their docs are UTTER garbage for nearly everything.
@@IvanRandomDude I had a job interview the other day and he gave feedback that he found it “weird” I don’t use redux libs. It’s because I don’t have to!
@@CodingAbroad From What I remember one of the redux creators said that when Redux came out React didn't have many features like context API and hooks so back then Redux had to be used. But since new features have been introduced he says that you don't have to use redux 99% of the time. Only for some super complex state management.
what turned me off from becoming front-end developer is the constant re-invention of the wheel (framework hell) and the non-stop tools and packages to stay on top of / maintain ... if I were to just focus on front-end dev only it would be word-press + php
SAME. I have Web dev skills. which really is just html css and javascript. but I don't wanna do this shit professionally. I just wanna write scripts and pipelines
I still make a good living doing simple html, CSS and JS websites. I've become quite proficient in GSAP and the sites do everything the client wants. There's plenty of small businesses out there who want a simple classy web presence and don't want the hassle of learning something like Wix. There's a ceiling to that business for sure, but as a simple small business, I'm never without work or short of money.
Well the good companies will make changes when it matters. That constant learning crap is not my cup of tea, either... i learn when i have too. Other than that i don't bother with the hundrerds of libraries and stuff... only the basics matter if you're good at that, that's great
IMHO, I found projects implementing React a total mess because of the unopinionated nature. After multiple nightmares, I now consider Angular as a better option for various reasons.
I'm currently using Blazor with Mudblazor, believe me when I say, it is a dream come true not having to deal with the mess that JS has become. And experiencing the structural guidance of a language like C# helped realize that.
I'm getting into web dev now and I hope that this trend continues. Given wasm's momentum and wasi being a probable thing, I don't think it's crazy to assume that future web developers will have options.
I think vanilla js is gonna be a thing. Now after IE11 end of life (IE11 did not support classes) you can use more new features, e.g. ES6 with classes. You can write your own components that are not locked into react easier. Everyone will do components their own way but code will be easier to understand because people will use more vanilla js functions instead of the framework ones. When querySelector/querySelectorAll was added to js, jquery became obsolete. IE6 (IE6 could not use querySelector) and windows XP had to die and now you do not need to use jquery at all. Similar thing will be waiting for React
I think the answer is 42. :) So in this era where architectural decision depends on tweets and trends, I'd suggest to invest in learning architecture. Something like clean architecture, hexagonal architecture, domain driven development even for the frontend (I am talking about mid/big project), and try to push frameworks and libraries calls as far as possible from the core business logic. React should be considered just a delivery media and should be switchable with anything else like Vue
"Jquery dead", WTF!? Not all projects need humongous frameworks, such as React! To many, the jQuery is enough! "where is Perl today. It's gone", WTF again!? Just because you have stopped using it, doesn't mean, it is gone! The community is not diminished! The new version came out! The current version is greatly improved!
I value the elegance and simplicity of svelte, given it's a transpiler vs framework like in the case of react, which allows it to eliminate complexity of generated code as the dom and ecmascript evolves.
As a former JQuery user, and an early adopters of Angular, I have to agree with this video. All these JS frame-works don't offer enough over the native code library. In my early JS days, JQuery was the thing to know, to the point of ignoring what it actually did. Half of it was wrapper code for finding elements, and the rest a wrapper for accessing them. I would have been better off learning the native JS ways to do those things, which I eventually did. For my latest project, I looked into React and Vue, and ended up using native Web Components. There might be a few things to learn about how they comunicate, but everything has it's own learning curve. I've been totally pleased with nesting my Web Components into more complex componenets, and I am always supprised how well the expected behaivor tends to be. Everything works with just a refresh since there are no non-native languages to compile. If I want to waste time compiling code, I can always make some of my slower JS into a faster WASM library file through Assembly Script, which is basicly a version of TypeScript.
I think the thing you are missing is all the work being done to improve DX for react. Vercel is a whole company built around React now. It isn’t just a tool anymore
React will go the way of the dodo when developers stop taking react jobs. As long as we keep accepting roles as full stack developers using react it will stay dominate. When businesses can’t find new react devs they will look at moving to something else. It’s a losing battle trying to change direction from within most companies as they want the most out of existing investments. If you want to ditch it, find a new job that doesn’t use it. When you leave tell your boss it’s because of the tech stack. Over time they will get the picture.
I started as web dev in 2000 with ASP and I stopped there. Had to change my carrier path. I said that just to set some background in Webdev and programming. Now, I'm learning reactJS. Seriously, it amazes me how "new" technologies just try reinvent the wheel and complicates everything. Back to 2000's, we were heading to a future where everything would be render on server side. They shut that down because it would overload servers. Now we have clouds and SSR.
I come from the same path as you and I notice the same as you how they try to reinvent everything. Id recommend if youre not that far on your react journey to try first svelte. This framework feels like a step in the right direction, it is actually a compiler very intuitive and with a low learning curve.
@@HeyDan1983 guess what? I was thinking about svelte few days ago. I won't stop learning react cuz I spent too much time on it, but will start svelte as soon as possible. Thanks for your tip. Good luck!
I think there are pros and cons with React etc vs jQuery. The complexity has gotten insane when it comes to configuration and dependencies, but the code quality (at least for larger applications) is generally better. The structure that frameworks provide for component bases architecture has helped us. I could write a well structured app with just JS today, but not so sure I would know how if I never worked with React or something similar.
We use the jsx/virtual dom engine in a non-React environment at my work and write in Typescript. No frameworks attached or anything - that's about as good as it gets in my opinion.
Custom Elements have been around for 3 years, and nobody cares, or wraps them in a higher level library. Fun fact: you cannot de-register a component, which makes it super fun to make dynamic UIs :)
I had to switch from React to Angular when I got hired and I have to say, I fuckin love Angular. Obviously this is my opinion, but I wish there was more coverage online as to the benefits of it, I love the "we got pretty much all you need" situation out the box, the structure and layout. But, and it's as big of a but that I'm sitting on, I am still considered a junior, only been doing this professionally for just over a year now, so hey, who knows what my opinion will be in a few years.
Chris is always speaking facts. Web components are amazing because it's native, easy to use and doesn't involve all the configuration you get with React, which in the long term make you a better programmer that's not dependent on a framework/lib. All people know is React, Vue, Svelte, etc. It's crazy to think we looked at those examples on the React site and thought those were cool even though one could literally build the counter in less than 5 mins with like 10 lines of JavaScript. One could possibly say JSX is cleaner than having a bunch of querySelector, setAttribute, createElement, and append methods calls.
If and when react falls out of fashion fwiw I do v much want something like JSX to live on Some people don't like mixing logic and markup together but I've had a blast with it personally
Web components are those simple classes with html written inside them right ? I started those but then I figured I should just learn react at that point , after taking an internship using plain ol js css and html . I could find more info on react so I convinced the guy in developing for to let me convert it, but it’s been a challenge, especially with no direction or guidance except these kind of videos.
Shadow DOM still has a lot of bugs in various browsers. For example, submit buttons not being recognized or password managers not finding inputs. As of right now, web components are too much of a hassle to maintain with all the extra bugs compared to vanilla DOM.
@W true but you can't ignore some of the more annoying design pitfalls react has by it's implementation it's better than vanilla js but you can't ignore how sometimes the lifecycle kinda sucks with useEffect hook and how much you need states everywhere
aren't examples meant to be just examples? They're not meant to be cool, they're meant to show the basic principles of library/framework on a simple and popular case. Every framework shows its principles based on counters. The more advanced examples are shown using usually a Todo List. Nobody claims you can't do a todo list in vanilla js just because somebody used it as an example of how a framework works. It's just an example, it doesn't have to be cool and was never meant to be cool.
7 years as a self taught fullstack dev and frontend is tiring... The good part for us is you learn the new shiny thing and companies pay more. Is this an inside job? The most fun I had with Javascript was writing a framework with vanilla javascript.
I think that the React docs are really meant more for devs like me, who are just starting out. I read them and try the examples because I'm starting from a blank slate, and the docs answer the question of "what does React even do?" You have a lot more experience so it makes sense that you will look at the docs with a more critical eye, thinking "but it's possible to do this without React, so..."
I don't see WASM replacing UI frameworks any time sson. JS is more than fast enough for 99.9% of UIs and gives you a lot more flexibility. There are other places where WASM can shine a lot more.
@W While I agree with you that WASM will probably not replace JS, I don't know what security issues you are talking about. Load time is also not an issue, it can even be faster than JS because WASM can be executed while it's still loading and there is only a very small warmup period.
WASM is really oriented for non-reload “as close to real time as possible in a browser context” calculations. Games, audio/video DSP, etc. Other comments in this video echo that someone/we need to push a better spec for web UI experience, whether that’s ground up/user first or clustered/networked server control of experience. Do we go back to server first? What control does the client have? Yadda… We already live in a dystopia in the free world; might as well cyberpunk the f outta these stacks!
For example imagine if in general our ISP digital networks were 10 times faster. Our current stack would have to catch up not only because we’ve written client code based on principles one to two decades old but also because our timers wouldn’t make sense anymore, as just a foundation. It’s like someone else above said; imagine if the whole technology behind the web changed within a year or two, for some freak reason. I guarantee you you’d have lobbyists not wanting that news to get out too fast. But also you have a whole new market. Those last two statements can’t coexist… We can’t be that far from the edge, right?
This would make a great debate on a live video/podcast. So many tech podcasts suck to listen to. It’s like watching paint dry, you def make it interesting.
I'm a newly educated web developer, and while I still see projects now being converted from early versions of Angular to a new version of React. When I was in an internship, the company I was working for, was actually converting from React to Vue. So it does happen, apparently. However, as it is right now, the company I currently work for, which is a consultancy, wants me to learn React (both JS and TS + materialUI) and redux (thunk and saga).
Also here since the beginning, redux was meant to scale but when we reached 100+ routes in React we started to learn it doesn't at all complexity doesn't scale. Recoil on the other hand does; The main things preventing React scaling today are the main libraries styled-components, Redux, Antd, Material, etc.
When I was first learning how to program, I learned react. Earlier this year, I got a new role and had to learn Angular. In my opinion, Angular makes more sense in my head and I like that it provides everything out of the box (routing, http, typescript, testing, etc.). Additionally, Angular encourages having different files for the html, typescript, and css so you’re not writing your html in your JavaScript files, though you can if you want to.
I personally think that, while I personally have a better DevExperience with frameworks like Vue and Svelte, most developers working for companies don't get to choose their technology. Even if they're in a CTO position, they still have to justify their choices to business people. If non-tech people don't see enough value, then mass adoption will only happen after a major incident or groundbreaking innovation. I think React still has another 7+ years at the top, and then follow how jQuery with how it is still omnipresent on anything deployed to the web.
Meh, I don't think React is going anywhere. It will just evolve and get better. Hooks are a great example of it evolving. Best and worst part of the framework is it is not opinionated.
One word... Svelte... There is something ganing traction in the UK and thats D.E Developer Experance, Svelte has a high rate of D.E its simple to use its comunity is passionate and I feel that it will become the go-to-framework in the next five years
For react to go away there needs to come something very revolutionary and that changes completely how we write web apps, it's not something that comes easily and react is really growing well, it is really implementing every necessary feature and idk but as I said just something very revolutionary could kick react out of the game
I went to a meetup where somebody said functional programming was cool so everything should be functional because everything should be cool. I don't actually know what functional programming is though so here's this weird useEffect thing that can only be called from one special place in one special function even though absolutely nothing enforces that constraint and it does 30 completely different things depending on where you pass the empty array to it but I don't feel like documenting all of them because I already quit Facebook after I invented a new best practice that I could put on my resume.
@@JohnSmith-yr4vi totally agree, useEffect must be only used for effects (sync with external things) but then you see people using it as event handlers and things that doesn't make sense xd
The biggest issue I see with WebComponents is their insistence on making you use classes. I wish there was a way to do Web Components in a functional programming way.
I was a backend dev that had to learn react for fullstack at a job. React dev experience is a giant headache. I'm excited to move to nextjs or even vue &nuxt 3
"Will new programmers want to write what programmers were writing 10 years ago" The thing is though, that React today is very different from React 10 years ago. I tried to get into React a few times in the past but it just wasn't making my life easier. Sure, it was providing a pattern that was easy to follow and structure my application with, but I didn't see the benefit. That all changed when I tried again a couple years ago and learned about functional components and hooks. It solved so many problems that class components had and actually did make things easier than trying to write separate HTML and JS. So to me, this new React is still relatively new (hooks came out in 2019). Of course I learned this because I made an effort to look into it and see how it was to use React. Your point may still stand because it kind of takes some amount of experience with it to know that it isn't the same. So they may see that it's 10 years old and immediately write it off. They do at least know that the landscape has a history of changing rather dramatically, and quite quickly. But, on the other hand, there are still so many companies out there hiring React developers. Part of the market is driven by that as well. Only time will tell with this one. Personally, I hope it sticks around for awhile, I quite like it.
According to Igor Minar React appeared at the beginning of "Generation 2" of Web technology. That generation of solutions was designed "client-side-first". There is every indication that Gen 3 (nothing to do with Web 3) solutions will be designed "server-side-first" in order to make SSR more efficient and minimize client side hydration costs. I encourage everybody to look up "Why Efficient Hydration in JavaScript Frameworks is so Challenging" (Debunking Server Rendering as the Silver Bullet). If anybody is looking for the next shiny object, Astro.build is the safest pick right now and you can even keep using React (until you pick Preact for a smaller bundle size and then perhaps Solid JS for even better performance). Web Components are a Gen 2 technology as SSR wasn't even under consideration. And in many cases web components are implemented with "web component frameworks" (like Lit) which adds dependencies and bloat if WCs from various sources use different WC frameworks . Astro.build does use custom elements to implement its islands, so there are valuable bits and pieces in the Web Component API. When I looked into WCs in 2019 I really wanted to like them but like many others I had to come to the conclusion that they missed their mark. "It's almost as if congealing 2010-era best practices in the platform before we'd finished exploring this territory was a mistake" - Rich Harris
@@jw4659 It's early days. Astro was introduced as an SSG in June 2021. It just went 1.0 beta April supporting SSR as an experimental feature. But it looks promising. More to the point: This March 2022 the Thoughtworks Technology Radar put a Hold on "SPA by default". That means that leading tech consultancies are noticing that their clients are often using SPAs where the are not appropriate. Astro isn't intended for building SPA's but to be used in the marketing, publishing, and eCommerce space for building performant server routed solutions (perhaps with nested routes on the page) that are a better fit than SPAs. And that seems to be the theme of Gen 3 - to move beyond SPAs.
This is great! I’ve been doing Angular and React the past three years, but before that, I dug into the web component documentation and wrote my own little library to get my head wrapped around this new functionality in the browser. It gave me my start in the above, but I’ve not heard anyone taking about it since then. I believe you that this is the immediate future, and also that the native APIs needs some help to compete and get there. I dig your tutorial videos, but moreso your videos that do philosophical breakdowns of technology.
The next framework from React will be the next version from React. The advantage of having such technologies be unstable is that innovation can be introduced without changing the names, so people can still keep looking for react developers and innovation can be brought to those business which are reluctant to change their stack. It happened already when functional components were introduced: you do not really use the same framework now as you did when classes were the only way to write components. The technology changed but the name stayed.
Playing devil's advocate, I believe there is a reason that react is the most used JavaScript framework. As complicated as it is, it is simple enough to setup routing, spas and bunch of stuff. Are there other options that are just as good? Sure, vue looks pretty good, angular is decent (I rather react but I like angular too), svelte seems to be good. In all honesty, if we learn javascript properly we can pick up any framework and I think that's where the focus should be at. I don't see javascript going anywhere so you future prove yourself learning JS and just pick up which ever framework is in demand.
@@titokris5162 The only reason is beacause is backed up by a major company (unlike vue) and because it overtook angularJs which was bad. After that people got used to it and company / developers as they got older and more structured does not want to change or take the risk to change, especially when the clients know react and not the others frameworks even if they are better in every aspect.
Yea react is cool but every 6 months new shit gets added for no reason. Personally ive been saying for years we should use native web components. Maybe even polymer Lit html
At work we use stencilJS, mainly for building React / Angular components. But we have web-component compatibility. Also, for react developers, there are many similarities so jumping ship isn't that bad.
Yet another brain-vomit episode about X JavaScript framework 🤣 We use both React and Vue at work, class components are a heaven-sent. Gives it that streamlined feel that doesn't hurt your brain when you switch between those 2 on a daily basis.
I remember back in 2007, we had GWT. Now, Jquery, react, svelte, vue... are just hell to maintain . Only Angular knows how to manage big projects, but no one knows how to use it correctly :'(
Yup that’s true. I stay with Angular. Only negative I can find is the hassle it is to have lazy loaded modules but I hear that might be faxed out in Angular 14 (they might be ditching ngModule)
@@CodingAbroad The "standalone" component ? Yeah, should be nice feature :) But you'll need to declare providers directly on component... I kind of get use to ngModules haha
Yes!!! Frameworks come and go, runtimes stays! With customElements, shadowDOM, slot, cssPart, the native component model is very powerful. Just need a minimalistic pub/sub/hub utility (for data/app event for data/app state management). For UI state events use the DOM event model.
3 years. Im glad someone else is onboard with LIT. I have always kind of avoided react its just my personality I sort of always look for alternative simpler options..react just seemed so unecessarily complex ..so when I found LIT it clicked with me. But I was concerned that maybe it would be another "failed" google project...and maybe I should suck it up and just learn react and stop being so anti conformity lol. But Im glad you put it into perspective, really need to think about the future rather than whats hot today. So im glad I am on the right path with LIT and it makes me confident in investing my energy to learn it
"really need to think about the future rather than whats hot today"... you cannot really predict the future and betting on future technology won't pay your bills today. Learn the fundamentals such as HTML/CSS/JS and then whatever future throws at you on the web, you'll be able to adapt. It might be React for the next 5 years, it might be whatever else.
"Tech have become obsolete before" is not an argument, that some tech WILL become obsolete, because it largely depends on circumstance. I can argue that Windows platform didn't become obsolete, but it won't be a fair comparison, because circumstances are different. Is Perl situation the same with React? If so, why?
I'm a backend developer, and i opt to focus my learn of frontend into Web Components, basiclly web basics and standards, since i don't want to work with frontend for a company and i just write frontend code for my projects or freelancing, i feel much more free to choose what i will use and i do think that web components is a great tecnology that could have some more focus. Just sharing the idea, has i'm not a frontend developer maybe this will not help to much, but i really liked the video
Have built a saas company using react. Has enabled us to create some fairly complex UIs. That being said, the state management is still a huge problem. Flux/Redux are too complex for small components. Forcing the developer to use a mix of custom hooks and redux or all redux. On top of that other concerns such as routing, animation, styling, async and failure modes is an afterthought. Another point to consider is cross-platform development - iOS, Android and Desktop. JSX can be very useful for that. Personally, will still stick with React or a JSX like implementation for a while. But I do think a solution that incorporates JSX, SOLID state, routing (web, mobile), async out of the box, JSX-less animation and a solve for styling will win. That could be react still.
Web components are getting more attention, in my org we have a decent sized application which we are moving out of old angular, instead of going with a newer framework like react etc. we went for web components based on LIT & it's working great till now. There are challanges but in long term it might work out better from maintenance perspective.
Im a firm beliver that the Dom is the final refference and can be relied upon... as long as browser implementers 'agree' on function and scope....your away...alll my projects arr vanilla js...css and axios for REST....thats enough
The worst part about it that I notice being new to react is that in a working environment you probably should know classes and hooks. Redux further complicates things because state sucks so bad. Finally the very worst part of react especially for beginners is that in many ways it operates like magic. For example if you want to decrement state you can define a function that does this.state - 1. Thats pretty intuitive. The magic comes in when you realize you have to do something like prevState => prevState and this is what is considered proper. Its like an obfuscation of pointers from other languages or something then you realize that js doesn’t “technically” have classes like you would expect. Personally Im starting to see its a javascript problem. I think we should use a language that operates closer to C++ or even Python for the web. Why? Well we don’t have to rely on frameworks at all in that case. We only have to rely on fundamental programming concepts like OOP, pointers, structs, defines etc. Build the ecosystem by vetting open sourced built in functions for the language. Have a sort of staging branch where people can build useful built in functions, that you can import from that branch and use on the web, but have only the best ones succeed to being part of the core language after optimization. We dont need frameworks, we need to make our core languages more powerful and abide to programming fundamentals so that 100 years from now the ecosystem is pure, not muddy and jumbled up like the mess the web is in these days.
Javascript wants to be a functional language but doesn't have the right design for implementing this well. It also doesn't have a good design for being OOP, and most damning of all it does not have the right features for web development which is the primary use of the language. Javascript I think illustrates a wider problem with language design, the people with the skills to create or fix languages don't add functionality that helps users deal with what a language tends to be used for into the language, everything is delegated to libraries and frameworks that turn into languages unto themselves if they become big enough.
You make some valid points on React but to me the issue is the DOM itself. React offers a consistent unidirectional flow of javascript generated DOM elements. State -> updated UI But why do we have to work with the DOM restriction? Why is it not intuitive to animate and position things. Flash gave a different paradigm with the MovieClip object that resembled more of setting up a movie scene with actors. Unity3D and Unreal follow a similar style if I'm not wrong. The DOM with it's restrictions offers a more structured UX, but it was never designed for fast and many content changes. Maybe the next thing is a Canvas UI framework or maybe the AI Code would evolve to a point that we merely describe the UI and it is auto generated. Who knows but for now the functional react way is a solid paradigm.
I think the switch to web components has already started at places like Google. I was thinking about web components for my next project (having used React, Vue, Angular, etc. before but more recently just plain HTML and JavaScript usually), but the available components registry/site that I saw kind of seemed like a garbage dump. Then I was thinking about what I want, such as a resizable panel, a drop down menu, and draggable windows maybe. Its honestly easiest for me to just use regular HTML with something like split.js and plain-draggable.js and then the web component that I saw for the menu actually said to put 'summary' and 'details' in the markup.. and since I already have that working without any component.. I think I am just going to skip the component stuff it and stick with my old-fashioned ways. But anyway I think that within two years, React will be uncool. Web components will have their day in the sun. But within another two years, it will all be about web assembly and actually Chrome and Firefox will start to be uncool. Instead we will separate out the information browser from the application delivery, which will be handled by web assembly which will get some UI stuff added so its not tied to the rest. And if they have any sense at all then they will start unbundling all of the other device APIs that are in the browser so that people can implement pieces of it and we aren't all globally beholden to one company on their implementation of a monolithic overlay operating system.
Agreed Chris I am doing react currently on several projects and I am the a hole in the room that brings up the question of why react? I could see if I had hundreds of things being simultaneously updated but 99% of the time I don’t and it’s just stuff that I can easily do with vanilla JavaScript I said when I learned react 4 1/2 years ago what happens when they just update how the DOM currently works?
Because web browsers were never supposed to run dynamic apps loke mobiles do. Web browser is just a static html viewer. And browser vendors are very slow to change because they are afraid that drastic changes will make them lose their users.
no, it's bc changing takes time and effort. a browser is a very complex software. you wouldn't be able to write one entirely on ur own even of u wanted to. also even if they changed all of a sudden the adoption of those new standards takes even more time to catch on
probably? all major mobile OSes are unix based. iPhones use Darwin (old fork of freebsd kernel) and Android uses a patched version of Linux. All the tooling needed to interface with the underlying hardware are either C or C++ (nowadays Rust is getting in on the fun as well but not as widespread) and the app talks to the native API in some way (GraalVM on android, idk if apple does native on iphone) so it can do something in there while being sandboxed
I tried to use web components, but it was an absolute pain. Biggest selling point (shadow dom separation) became a biggest problem, because I struggled to manage ui consistency. The API gives you nothing useful and you need a framework over it to make it useful. So how's that better than react?
I mean, as bloated as Angular is, people still use it; so I definitely don't see React going anywhere anytime soon. When it comes right down to it, make sure you learn your fundamentals and understand how to code in vanilla. Frameworks should be tools and scaffolds to use in a language, not a language into and of themselves.
I think react has 10 more years. It's the jQuery of our time. No framework is perfect but for now it's the King. I feel it needs a rewrite though. They need to go back on the drawing board and bring in everything, routing, fullstack, state management, and steal some ideas from angular,
Last time I worked in PERL, it was for a site that was updating to MOO when other sites were ditching PERL entirely (such as Amazon and Zappos). I didn't work there for long. I loved the company, loved the co-workers, but we were trying to adapt old tech to new paradigms that were dysfunctional together. We should have been rewriting it in Python or splitting out the backend from the site application, and it was constantly frustrating.
To be completely honest, coming from vue (v2), I gave react a chance and so far I'm loving it. Probably because I went with Nextjs and the sheer amount of libraries react have is excellent. I'm yet to try react-query + Zustand but I'm hearing great things regarding these libraries.
HTML Re-usability was a solved problem since the XHTML/XML/XSLT days. Everything that came after is people thinking the status quo is too complicated, making something simple, then realizing that there's a need for complexity after all and bloating their framework until it's more complicated than the previous status quo.
I've never understood why frameworks are so popular, especially nowadays. If it takes years to learn the framework then why not just learn vanilla JS? Is vanilla JS really that bad to work with?
React is just a sign that Javascript is fundamentally unfit for what needs to be used. Web Components mostly suffers from the same problem since it doesn't actually fix the underlying problem which is Javascript.
I have a dilemma/catch 22 with coding/Programmng I do it as a hobbie. I currently work as an operations technician on an government controlled businesses. Background as acarpenter. But always found the joy snd peace with computers, everything self thougt, I am google certified Data analyst an it-support pro. I am afraid that if I switck from current work I will loose the fun wiith computers and data processing. But in the same tkme i rather work with something more computerrelated. Computers has been a part of my life snce mid 90's
Here it goes again the React mood swing
🤣 lol
🤣🤣🤣🤣🤣🤣🤣
It's hilarious. We live in a time where you can choose whatever library or framework you prefer. I don't like Flutter but I won't go online and rant about it, cause some people find it useful for their needs.
@@neneodonkor in this situation, Facebook is the reason react sux
😂😂😂😂
I think you've answered your own question. The "new thing" 10 years from now is a new version of React, because of how it constantly evolve itself towards something better. People will only move away if React stops innovating and adapting, then the replacement will be a new thing that's willing to adapt.
I've been using React in production since 2015 and went through all the migration, so I actually appreciate all the updates :)
Chris my boi! In SolidJS there's no need to deal with stale closures or useEffect infinite loop headaches, plus handles global state out of the box, just put createStore hook inside it's own file or outside of component (in SolidJS, hooks don't need to be declared inside components). One last thing, JSX elements are just native elements so if you console.log(hi) it will actually log native div element.
In answer to your question at the end of the video, I think React has another 5-10 years.
Solid makes React look like when there's a good, well written book, and then it get's adapted and the history being told takes several weird moves hehe
if you have problem with useEffect, you are doing it wrong.
I don't know why people keep invoking the useEffect loop as an issue. I mean there's a thousand issues with React but this isn't one. If you're getting loops caused by useEffect the problem is not the hook but you who didn't understand how it works or how reference equality works. I haven't had that issue since the first time it happened to me because I actually took my time to read how it works and understand it. It's really simple
Skill issue git gud with react. It's designed to make you shoot yourself in the foot
React code is unreadable. Complete mess. I can't believe that Svelte or Vue are not in top when it comes to jobs. Framework should be simple to learn and use. "It's not real Javascript" argument is stupid, no one cares. Just make product in simple way.. React is preventing us from that.
Great video.
I’m a newbie developer and I have not learned React. Instead I looked at my options and decided to use Svelte(kit).
The only downside is that it will take longer to find a job. But meantime I’m building basic projects with Sveltekit for freelance clients whilst working my finance job, no regrets so far.
My friend just landed a job with React and the over complexity for the simplest things is throwing him off. I showed him some svelte over the weekend and his reply was ‘fuck you’ :)
I did the same.
It feels like react is exactly a mirror of the human behavior towards "an issue/error" in real life.
Instead of seeking simplicity and seeking the root of the issue, it implement more layers to the issue to "fix" something.
It's what you can see exactly in psychology.
But like you i'm a junior dev (in my 30's) so i'm new in mostly anything towards that.
and i'm like you (haha) it does take longer to find a job with Svelte.
Why not learn React and get some work experience while learning your preferred framework? In my opinion, it could be easier for you to get that new job with that framework you like with some work experience under your belt.
@@jmse37 I was pondering how to write this without sounding bigheaded or just plain ignorant, but i'll just go for it. Overall, thanks for the comment as I think for the most part, for most people, you are 100% correct. I just don't think I can apply that for myself (and you're not the first person to suggest just learning React).
Ultimately React is fine to use and I could land a job faster, you're right. It would probably take me 2-3 weeks to get used to it to an OK level.
But I have always benefited from doing things my own way, it helped me stand out. I got a finance analyst job because of sales & graphic design experience (plus finance knowledge) as an example. I usually find a way to mix my skills and knowledge when i apply for very specific jobs.
So again, you're right generally speaking. I just never personally benefited from copying/following what other people did, it does not help me stand out and usually slows me down.
And my ultimate goal is to build a small agency that does things its own way but lots to learn till then. I'll definitely land a job, at worst I'll pick up Vue as NuxtJS 3 is looking very interesting too!
@@semikolondev We think alike! Whether we're right or not I ultimately think that whatever works for you as an individual is best to do.
Following the masses never worked for me personally. I just find myself getting annoyed at everything over time.
I also think Svelte will pick up. Don't think Vercel would have hired Rich Harris to work on it full time if they didn't see potential.
Building everything faster and having more readable code is amazing for me personally. OF COURSE I wish Svelte had a bigger ecosystem (lets say TailwindUI for Svelte) but that will all come with time. I don't think Svelte is going anywhere but up.
And React is amazing in its own right anyway, it paved the way for things like Vue and Svelte. All of these frameworks learned from each other after all :).
The friend i mentioned who is learning React is just weirded out by how complex some simple things are in NextJS. I was too when he was showing me and we were comparing by building identical components in React and Svelte. Fun little exercise that was :)
svelte is great. A compilation step instead of a runtime is a step in the right direction.
SolidJS, Svelte and the like are the future. Use compiler to abstract away the underlying complexities, present a simple surface dev friendly API.
Exactly. If you like Vue the natural way in the future is Svelte, if you like React you go to SolidJs. They both are excellent (I personally prefer solid because I have always liked react). Solid builds upon the foundation of React and smooth its pain points. IMHO, we components is a dead fish
@@victorlongon interestingly enough I don't enjoy writing Vue at all but have had a good time writing Svelte so far.
Haven't tried SolidJS yet, but if it's more geared towards the React side of things I'm really curious to check it out.
@@baka_baca i tried just a bit. It is very similar to React but better
The problem with web components is that it only addresses one issue, namely encapsulation. SSR is out of the scope, API is needlessly complex, doesn't address state management and data binding, etc.
I ditched React for Vue3 mostly because of its reactive system and composable model (vs. Vue2), but having said that I'm quite confident that Svelte and SvelteKit will rule this and next decade. Why? This:
let count = 1;
const inc = () => count += 1;
{count}
inc
.count { font-weight: bold; }
It just CAN NOT get any simpler. Do that in React, Vue, Angular or WebComponents. When compared to anything Svelte is just superior in every aspect.
Oh yeah, svelte is looking great, it just needs more PR
To be honest, you just boiled down everything that is to be said.
Most of my projects are just written in Svelte, Svelte-Kit. Because it just make a lot of sense with it simplicity.
Am not batching any frameworks here, but there is a saying that:
Svelte is fun, Vue is all rounded, and React just works.😄
I kinda preffer this:
It's Solid, and the state can be anywhere, it uses JSX which makes it really easy to learn if you come from React and you don't get "why the f do I need to write so much extra code just to read a value?" moments, add to that the fact that is probably unbeatable in performance and size, yes, Svelte is small and performant too but Solid is small and performant even when there'sm a big ammount of components and Svelte can't handle that too well
vue 3 script setup looks pretty similar to that
@@elultimopujilense It does, but Svelte made it even simpler (no reference vs. reactive objects), TS integration is far better when defining component properties and slots. Also templating is a bit more powerful with "awaited" keyword, so there's no need to wrap things inside AsyncComponents. Exposing properties is also simpler (you just "export" them)...and probably a lot more (I'm not Svelte expert). No VDOM. All in all, it's Vue3 on steroids and better TS integration which is very crucial for me.
This is where Vue3 is probably the worst option, TS support. While it's way better then Vue2, it's still behind Svelte and especially React which has great support for TS (since it has no templating and it's all function anyways).
I've only been writing React for about 3 years.
And the answer is yes.
Also because you touched on it, it blows my mind how terrible their docs are.
Like this is React. Made by Facebook.
And their docs are UTTER garbage for nearly everything.
you are just an incompetent programmer ....
Facebook ui sorta blows too
@@realchrishawkes large yes
Never found the docs being terrible, the only terrible part out of it is navigating
@@JEsterCW They're a lot better now.
But for a while, they were laughably bad and incomplete.
React is a dream for developers. At least in UI functional programming won and became commercially viable. It's amazing.
Svelte is a treat. It can take us 2 hour to learn and it’s very close to native
I was watching a senior dev writing react and he got annoyed of how bad it is.
If it wasn't for Nextjs i would've ditched react long time ago. Hopefully SvelteKit catches up quickly.
ah yes redux. 99.9% of the time is like bringing a rocket launcher to a fist fight.
So true
Even Redux creator said that he does not use redux rofl
@@IvanRandomDude I had a job interview the other day and he gave feedback that he found it “weird” I don’t use redux libs. It’s because I don’t have to!
@@CodingAbroad From What I remember one of the redux creators said that when Redux came out React didn't have many features like context API and hooks so back then Redux had to be used. But since new features have been introduced he says that you don't have to use redux 99% of the time. Only for some super complex state management.
Try svelte, its like the dose of fresh air we all need after this react fatigue.
I've never liked React. Kind of just got forced into using it by all my employers. I'd love to see it go bye bye.
what turned me off from becoming front-end developer is the constant re-invention of the wheel (framework hell) and the non-stop tools and packages to stay on top of / maintain ... if I were to just focus on front-end dev only it would be word-press + php
SAME. I have Web dev skills. which really is just html css and javascript.
but I don't wanna do this shit professionally. I just wanna write scripts and pipelines
I still make a good living doing simple html, CSS and JS websites. I've become quite proficient in GSAP and the sites do everything the client wants. There's plenty of small businesses out there who want a simple classy web presence and don't want the hassle of learning something like Wix. There's a ceiling to that business for sure, but as a simple small business, I'm never without work or short of money.
Well the good companies will make changes when it matters. That constant learning crap is not my cup of tea, either... i learn when i have too. Other than that i don't bother with the hundrerds of libraries and stuff... only the basics matter if you're good at that, that's great
IMHO, I found projects implementing React a total mess because of the unopinionated nature. After multiple nightmares, I now consider Angular as a better option for various reasons.
I'm currently using Blazor with Mudblazor, believe me when I say, it is a dream come true not having to deal with the mess that JS has become. And experiencing the structural guidance of a language like C# helped realize that.
I'm getting into web dev now and I hope that this trend continues.
Given wasm's momentum and wasi being a probable thing, I don't think it's crazy to assume that future web developers will have options.
Agree 💯 Blazor with Mudblazor really is a dream come true. Its almost fuckin perfect
I think vanilla js is gonna be a thing. Now after IE11 end of life (IE11 did not support classes) you can use more new features, e.g. ES6 with classes.
You can write your own components that are not locked into react easier. Everyone will do components their own way but code will be easier to understand because people will use more vanilla js functions instead of the framework ones.
When querySelector/querySelectorAll was added to js, jquery became obsolete. IE6 (IE6 could not use querySelector) and windows XP had to die and now you do not need to use jquery at all. Similar thing will be waiting for React
@W actually i can write a todo list with vanilla js in 100 lines of code while in react i should write more , what is that?
@@mahdisalmanizadegan5595 if you need to write more than 50 lines of code for a todo app in react, something is wrong.
@@kristiansmith5298 not only a simple todo list, a todo app with delete, add and edit capability
@@kristiansmith5298 imagine you want to create a component for every part of your UI, it become so huge code base
@@mahdisalmanizadegan5595 🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣
I think the answer is 42. :) So in this era where architectural decision depends on tweets and trends, I'd suggest to invest in learning architecture. Something like clean architecture, hexagonal architecture, domain driven development even for the frontend (I am talking about mid/big project), and try to push frameworks and libraries calls as far as possible from the core business logic. React should be considered just a delivery media and should be switchable with anything else like Vue
Hey Filippo, do you know any good resources on those architectural topics? Books, courses, etc. Thanks in advance!
"Jquery dead", WTF!?
Not all projects need humongous frameworks, such as React!
To many, the jQuery is enough!
"where is Perl today. It's gone", WTF again!? Just because you have stopped using it, doesn't mean, it is gone! The community is not diminished! The new version came out! The current version is greatly improved!
I value the elegance and simplicity of svelte, given it's a transpiler vs framework like in the case of react, which allows it to eliminate complexity of generated code as the dom and ecmascript evolves.
As a former JQuery user, and an early adopters of Angular, I have to agree with this video. All these JS frame-works don't offer enough over the native code library. In my early JS days, JQuery was the thing to know, to the point of ignoring what it actually did. Half of it was wrapper code for finding elements, and the rest a wrapper for accessing them. I would have been better off learning the native JS ways to do those things, which I eventually did.
For my latest project, I looked into React and Vue, and ended up using native Web Components. There might be a few things to learn about how they comunicate, but everything has it's own learning curve. I've been totally pleased with nesting my Web Components into more complex componenets, and I am always supprised how well the expected behaivor tends to be.
Everything works with just a refresh since there are no non-native languages to compile. If I want to waste time compiling code, I can always make some of my slower JS into a faster WASM library file through Assembly Script, which is basicly a version of TypeScript.
I think the thing you are missing is all the work being done to improve DX for react. Vercel is a whole company built around React now. It isn’t just a tool anymore
I believe once Svelte lands libraries for accessibility like React-Aria and several others that are common in React, the decline will come faster
Svelte ftw.
svelte stonks on the rise
React will go the way of the dodo when developers stop taking react jobs. As long as we keep accepting roles as full stack developers using react it will stay dominate. When businesses can’t find new react devs they will look at moving to something else.
It’s a losing battle trying to change direction from within most companies as they want the most out of existing investments. If you want to ditch it, find a new job that doesn’t use it. When you leave tell your boss it’s because of the tech stack. Over time they will get the picture.
I started as web dev in 2000 with ASP and I stopped there. Had to change my carrier path.
I said that just to set some background in Webdev and programming. Now, I'm learning reactJS.
Seriously, it amazes me how "new" technologies just try reinvent the wheel and complicates
everything. Back to 2000's, we were heading to a future where everything would be render on
server side. They shut that down because it would overload servers. Now we have clouds and
SSR.
I come from the same path as you and I notice the same as you how they try to reinvent everything. Id recommend if youre not that far on your react journey to try first svelte. This framework feels like a step in the right direction, it is actually a compiler very intuitive and with a low learning curve.
Ditto
First time I heard SSR and understood what it is I shouted "WTF, this is literally where we were coming from originally"
@@HeyDan1983 guess what? I was thinking about svelte few days ago. I won't stop learning react cuz I spent too much time on it, but will start svelte as soon as possible. Thanks for your tip. Good luck!
Of course this video comes out when I just start my Angular internship
And before I've even gotten to JS 🥲
Good luck and welcome to the space ☺️ I think you’ll have a more straight forward introduction to it (I use Angular)
I think there are pros and cons with React etc vs jQuery. The complexity has gotten insane when it comes to configuration and dependencies, but the code quality (at least for larger applications) is generally better. The structure that frameworks provide for component bases architecture has helped us. I could write a well structured app with just JS today, but not so sure I would know how if I never worked with React or something similar.
We use the jsx/virtual dom engine in a non-React environment at my work and write in Typescript. No frameworks attached or anything - that's about as good as it gets in my opinion.
Custom Elements have been around for 3 years, and nobody cares, or wraps them in a higher level library. Fun fact: you cannot de-register a component, which makes it super fun to make dynamic UIs :)
I had to switch from React to Angular when I got hired and I have to say, I fuckin love Angular. Obviously this is my opinion, but I wish there was more coverage online as to the benefits of it, I love the "we got pretty much all you need" situation out the box, the structure and layout. But, and it's as big of a but that I'm sitting on, I am still considered a junior, only been doing this professionally for just over a year now, so hey, who knows what my opinion will be in a few years.
Plus complex forms can be done so much more easily in angular vs react and vue (and I do all 3)
Chris is always speaking facts. Web components are amazing because it's native, easy to use and doesn't involve all the configuration you get with React, which in the long term make you a better programmer that's not dependent on a framework/lib. All people know is React, Vue, Svelte, etc. It's crazy to think we looked at those examples on the React site and thought those were cool even though one could literally build the counter in less than 5 mins with like 10 lines of JavaScript. One could possibly say JSX is cleaner than having a bunch of querySelector, setAttribute, createElement, and append methods calls.
If and when react falls out of fashion fwiw I do v much want something like JSX to live on
Some people don't like mixing logic and markup together but I've had a blast with it personally
Web components are those simple classes with html written inside them right ? I started those but then I figured I should just learn react at that point , after taking an internship using plain ol js css and html . I could find more info on react so I convinced the guy in developing for to let me convert it, but it’s been a challenge, especially with no direction or guidance except these kind of videos.
Shadow DOM still has a lot of bugs in various browsers. For example, submit buttons not being recognized or password managers not finding inputs. As of right now, web components are too much of a hassle to maintain with all the extra bugs compared to vanilla DOM.
@W true but you can't ignore some of the more annoying design pitfalls react has by it's implementation it's better than vanilla js but you can't ignore how sometimes the lifecycle kinda sucks with useEffect hook and how much you need states everywhere
aren't examples meant to be just examples? They're not meant to be cool, they're meant to show the basic principles of library/framework on a simple and popular case. Every framework shows its principles based on counters. The more advanced examples are shown using usually a Todo List. Nobody claims you can't do a todo list in vanilla js just because somebody used it as an example of how a framework works. It's just an example, it doesn't have to be cool and was never meant to be cool.
7 years as a self taught fullstack dev and frontend is tiring...
The good part for us is you learn the new shiny thing and companies pay more.
Is this an inside job?
The most fun I had with Javascript was writing a framework with vanilla javascript.
I think that the React docs are really meant more for devs like me, who are just starting out. I read them and try the examples because I'm starting from a blank slate, and the docs answer the question of "what does React even do?" You have a lot more experience so it makes sense that you will look at the docs with a more critical eye, thinking "but it's possible to do this without React, so..."
It could something like web assembly that could kill React, something like Blazor. You have a valid point.
I don't see WASM replacing UI frameworks any time sson. JS is more than fast enough for 99.9% of UIs and gives you a lot more flexibility. There are other places where WASM can shine a lot more.
@W While I agree with you that WASM will probably not replace JS, I don't know what security issues you are talking about. Load time is also not an issue, it can even be faster than JS because WASM can be executed while it's still loading and there is only a very small warmup period.
WASM is really oriented for non-reload “as close to real time as possible in a browser context” calculations. Games, audio/video DSP, etc.
Other comments in this video echo that someone/we need to push a better spec for web UI experience, whether that’s ground up/user first or clustered/networked server control of experience. Do we go back to server first? What control does the client have? Yadda…
We already live in a dystopia in the free world; might as well cyberpunk the f outta these stacks!
For example imagine if in general our ISP digital networks were 10 times faster. Our current stack would have to catch up not only because we’ve written client code based on principles one to two decades old but also because our timers wouldn’t make sense anymore, as just a foundation. It’s like someone else above said; imagine if the whole technology behind the web changed within a year or two, for some freak reason. I guarantee you you’d have lobbyists not wanting that news to get out too fast. But also you have a whole new market. Those last two statements can’t coexist…
We can’t be that far from the edge, right?
Add on top of that the lag AND innovation going on around web3 and blockchain data. Keep a substantial part of your grey matter fresh, choom! 😁
This would make a great debate on a live video/podcast. So many tech podcasts suck to listen to. It’s like watching paint dry, you def make it interesting.
I'm a newly educated web developer, and while I still see projects now being converted from early versions of Angular to a new version of React. When I was in an internship, the company I was working for, was actually converting from React to Vue.
So it does happen, apparently.
However, as it is right now, the company I currently work for, which is a consultancy, wants me to learn React (both JS and TS + materialUI) and redux (thunk and saga).
Also here since the beginning, redux was meant to scale but when we reached 100+ routes in React we started to learn it doesn't at all complexity doesn't scale. Recoil on the other hand does; The main things preventing React scaling today are the main libraries styled-components, Redux, Antd, Material, etc.
When I was first learning how to program, I learned react. Earlier this year, I got a new role and had to learn Angular. In my opinion, Angular makes more sense in my head and I like that it provides everything out of the box (routing, http, typescript, testing, etc.). Additionally, Angular encourages having different files for the html, typescript, and css so you’re not writing your html in your JavaScript files, though you can if you want to.
Most of the reasons I just like Angular because, of its SoC(Separation of Concern) it gives on how to effectively and efficiently manage once project
I hate Angular so badly, the worst experience thst ive experienced in my entire web development life, i dont like anything about it 🤣
Guys switch to elm. You'll love it, that's the true game changer. Four words:
Guaranteed no runtime errors
I personally think that, while I personally have a better DevExperience with frameworks like Vue and Svelte, most developers working for companies don't get to choose their technology.
Even if they're in a CTO position, they still have to justify their choices to business people. If non-tech people don't see enough value, then mass adoption will only happen after a major incident or groundbreaking innovation.
I think React still has another 7+ years at the top, and then follow how jQuery with how it is still omnipresent on anything deployed to the web.
Agreed, have to just accept what we can't control. Go with the flow as best as possible.
Meh, I don't think React is going anywhere. It will just evolve and get better. Hooks are a great example of it evolving. Best and worst part of the framework is it is not opinionated.
I agree it’s the best of the top three.
One word... Svelte... There is something ganing traction in the UK and thats D.E Developer Experance, Svelte has a high rate of D.E its simple to use its comunity is passionate and I feel that it will become the go-to-framework in the next five years
For react to go away there needs to come something very revolutionary and that changes completely how we write web apps, it's not something that comes easily and react is really growing well, it is really implementing every necessary feature and idk but as I said just something very revolutionary could kick react out of the game
I went to a meetup where somebody said functional programming was cool so everything should be functional because everything should be cool. I don't actually know what functional programming is though so here's this weird useEffect thing that can only be called from one special place in one special function even though absolutely nothing enforces that constraint and it does 30 completely different things depending on where you pass the empty array to it but I don't feel like documenting all of them because I already quit Facebook after I invented a new best practice that I could put on my resume.
🤣🤣
what? useEffect doesn't do 30 different things, it's really not a complicated function. You're just confused.
@@JohnSmith-yr4vi totally agree, useEffect must be only used for effects (sync with external things) but then you see people using it as event handlers and things that doesn't make sense xd
Instantly distrust any opinion beginning with the words "everything should"
The biggest issue I see with WebComponents is their insistence on making you use classes. I wish there was a way to do Web Components in a functional programming way.
I was a backend dev that had to learn react for fullstack at a job. React dev experience is a giant headache. I'm excited to move to nextjs or even vue &nuxt 3
"Will new programmers want to write what programmers were writing 10 years ago"
The thing is though, that React today is very different from React 10 years ago. I tried to get into React a few times in the past but it just wasn't making my life easier. Sure, it was providing a pattern that was easy to follow and structure my application with, but I didn't see the benefit. That all changed when I tried again a couple years ago and learned about functional components and hooks. It solved so many problems that class components had and actually did make things easier than trying to write separate HTML and JS. So to me, this new React is still relatively new (hooks came out in 2019).
Of course I learned this because I made an effort to look into it and see how it was to use React. Your point may still stand because it kind of takes some amount of experience with it to know that it isn't the same. So they may see that it's 10 years old and immediately write it off. They do at least know that the landscape has a history of changing rather dramatically, and quite quickly. But, on the other hand, there are still so many companies out there hiring React developers. Part of the market is driven by that as well. Only time will tell with this one. Personally, I hope it sticks around for awhile, I quite like it.
According to Igor Minar React appeared at the beginning of "Generation 2" of Web technology. That generation of solutions was designed "client-side-first".
There is every indication that Gen 3 (nothing to do with Web 3) solutions will be designed "server-side-first" in order to make SSR more efficient and minimize client side hydration costs.
I encourage everybody to look up "Why Efficient Hydration in JavaScript Frameworks is so Challenging" (Debunking Server Rendering as the Silver Bullet).
If anybody is looking for the next shiny object, Astro.build is the safest pick right now and you can even keep using React (until you pick Preact for a smaller bundle size and then perhaps Solid JS for even better performance).
Web Components are a Gen 2 technology as SSR wasn't even under consideration. And in many cases web components are implemented with "web component frameworks" (like Lit) which adds dependencies and bloat if WCs from various sources use different WC frameworks .
Astro.build does use custom elements to implement its islands, so there are valuable bits and pieces in the Web Component API.
When I looked into WCs in 2019 I really wanted to like them but like many others I had to come to the conclusion that they missed their mark.
"It's almost as if congealing 2010-era best practices in the platform before we'd finished exploring this territory was a mistake" - Rich Harris
This Astro.Build sounds very cool, does it really work?
@@jw4659 It's early days. Astro was introduced as an SSG in June 2021. It just went 1.0 beta April supporting SSR as an experimental feature. But it looks promising.
More to the point: This March 2022 the Thoughtworks Technology Radar put a Hold on "SPA by default". That means that leading tech consultancies are noticing that their clients are often using SPAs where the are not appropriate.
Astro isn't intended for building SPA's but to be used in the marketing, publishing, and eCommerce space for building performant server routed solutions (perhaps with nested routes on the page) that are a better fit than SPAs.
And that seems to be the theme of Gen 3 - to move beyond SPAs.
@@PeerReynders Astro sounds very good to me, I will use it with React to experiment with how it goes. This will be fun!
This is great!
I’ve been doing Angular and React the past three years, but before that, I dug into the web component documentation and wrote my own little library to get my head wrapped around this new functionality in the browser. It gave me my start in the above, but I’ve not heard anyone taking about it since then. I believe you that this is the immediate future, and also that the native APIs needs some help to compete and get there.
I dig your tutorial videos, but moreso your videos that do philosophical breakdowns of technology.
The next framework from React will be the next version from React. The advantage of having such technologies be unstable is that innovation can be introduced without changing the names, so people can still keep looking for react developers and innovation can be brought to those business which are reluctant to change their stack. It happened already when functional components were introduced: you do not really use the same framework now as you did when classes were the only way to write components. The technology changed but the name stayed.
Playing devil's advocate, I believe there is a reason that react is the most used JavaScript framework. As complicated as it is, it is simple enough to setup routing, spas and bunch of stuff. Are there other options that are just as good? Sure, vue looks pretty good, angular is decent (I rather react but I like angular too), svelte seems to be good. In all honesty, if we learn javascript properly we can pick up any framework and I think that's where the focus should be at. I don't see javascript going anywhere so you future prove yourself learning JS and just pick up which ever framework is in demand.
That was deaf on point !!!
@@titokris5162 The only reason is beacause is backed up by a major company (unlike vue) and because it overtook angularJs which was bad. After that people got used to it and company / developers as they got older and more structured does not want to change or take the risk to change, especially when the clients know react and not the others frameworks even if they are better in every aspect.
Yea react is cool but every 6 months new shit gets added for no reason. Personally ive been saying for years we should use native web components. Maybe even polymer Lit html
I'll jump ship when it affects my wallet.
At work we use stencilJS, mainly for building React / Angular components. But we have web-component compatibility. Also, for react developers, there are many similarities so jumping ship isn't that bad.
I want to thumbs up this like 45 more times than I can
WEB COMPONENTS AND JSDOC FOR THE WIN
Ditch React, Ditch Redux, Ditch all upcoming JS Frameworks, Ditch Companies that use React
Ditch them all
Let Jquery Rule once again !!!
Fascinating walk over the past ten react years placed in the perspective of where next. Must view.
Yet another brain-vomit episode about X JavaScript framework 🤣
We use both React and Vue at work, class components are a heaven-sent. Gives it that streamlined feel that doesn't hurt your brain when you switch between those 2 on a daily basis.
I remember back in 2007, we had GWT. Now, Jquery, react, svelte, vue... are just hell to maintain . Only Angular knows how to manage big projects, but no one knows how to use it correctly :'(
Yup that’s true. I stay with Angular. Only negative I can find is the hassle it is to have lazy loaded modules but I hear that might be faxed out in Angular 14 (they might be ditching ngModule)
@@CodingAbroad The "standalone" component ? Yeah, should be nice feature :) But you'll need to declare providers directly on component... I kind of get use to ngModules haha
ODIN is now my favorite language. Data oriented is the future
Yes!!! Frameworks come and go, runtimes stays!
With customElements, shadowDOM, slot, cssPart, the native component model is very powerful. Just need a minimalistic pub/sub/hub utility (for data/app event for data/app state management). For UI state events use the DOM event model.
3 years. Im glad someone else is onboard with LIT. I have always kind of avoided react its just my personality I sort of always look for alternative simpler options..react just seemed so unecessarily complex ..so when I found LIT it clicked with me. But I was concerned that maybe it would be another "failed" google project...and maybe I should suck it up and just learn react and stop being so anti conformity lol.
But Im glad you put it into perspective, really need to think about the future rather than whats hot today. So im glad I am on the right path with LIT and it makes me confident in investing my energy to learn it
"really need to think about the future rather than whats hot today"... you cannot really predict the future and betting on future technology won't pay your bills today. Learn the fundamentals such as HTML/CSS/JS and then whatever future throws at you on the web, you'll be able to adapt. It might be React for the next 5 years, it might be whatever else.
"Tech have become obsolete before" is not an argument, that some tech WILL become obsolete, because it largely depends on circumstance. I can argue that Windows platform didn't become obsolete, but it won't be a fair comparison, because circumstances are different. Is Perl situation the same with React? If so, why?
Totally agree. 90% of web applications don’t need a framework. Learn to interact with the DOM and URL params
solid js. react like syntax, no vdom, awesome state management built in
How am I gonna ditch react when my company uses it?
You're using Camtasia for your videos?
Yes
I'm a backend developer, and i opt to focus my learn of frontend into Web Components, basiclly web basics and standards, since i don't want to work with frontend for a company and i just write frontend code for my projects or freelancing, i feel much more free to choose what i will use and i do think that web components is a great tecnology that could have some more focus. Just sharing the idea, has i'm not a frontend developer maybe this will not help to much, but i really liked the video
Have built a saas company using react. Has enabled us to create some fairly complex UIs. That being said, the state management is still a huge problem. Flux/Redux are too complex for small components. Forcing the developer to use a mix of custom hooks and redux or all redux. On top of that other concerns such as routing, animation, styling, async and failure modes is an afterthought. Another point to consider is cross-platform development - iOS, Android and Desktop. JSX can be very useful for that.
Personally, will still stick with React or a JSX like implementation for a while. But I do think a solution that incorporates JSX, SOLID state, routing (web, mobile), async out of the box, JSX-less animation and a solve for styling will win. That could be react still.
Web components are getting more attention, in my org we have a decent sized application which we are moving out of old angular, instead of going with a newer framework like react etc. we went for web components based on LIT & it's working great till now.
There are challanges but in long term it might work out better from maintenance perspective.
Im a firm beliver that the Dom is the final refference and can be relied upon... as long as browser implementers 'agree' on function and scope....your away...alll my projects arr vanilla js...css and axios for REST....thats enough
The worst part about it that I notice being new to react is that in a working environment you probably should know classes and hooks. Redux further complicates things because state sucks so bad. Finally the very worst part of react especially for beginners is that in many ways it operates like magic. For example if you want to decrement state you can define a function that does this.state - 1. Thats pretty intuitive. The magic comes in when you realize you have to do something like prevState => prevState and this is what is considered proper. Its like an obfuscation of pointers from other languages or something then you realize that js doesn’t “technically” have classes like you would expect. Personally Im starting to see its a javascript problem. I think we should use a language that operates closer to C++ or even Python for the web. Why? Well we don’t have to rely on frameworks at all in that case. We only have to rely on fundamental programming concepts like OOP, pointers, structs, defines etc. Build the ecosystem by vetting open sourced built in functions for the language. Have a sort of staging branch where people can build useful built in functions, that you can import from that branch and use on the web, but have only the best ones succeed to being part of the core language after optimization. We dont need frameworks, we need to make our core languages more powerful and abide to programming fundamentals so that 100 years from now the ecosystem is pure, not muddy and jumbled up like the mess the web is in these days.
It werks ?!
Javascript wants to be a functional language but doesn't have the right design for implementing this well. It also doesn't have a good design for being OOP, and most damning of all it does not have the right features for web development which is the primary use of the language. Javascript I think illustrates a wider problem with language design, the people with the skills to create or fix languages don't add functionality that helps users deal with what a language tends to be used for into the language, everything is delegated to libraries and frameworks that turn into languages unto themselves if they become big enough.
SoilidJS is…… pretty solid. It’s if React were created with state management as a primary concern.
I'm hearing good things about it. I've been meaning to add a course on it.
What do you think should the starters start learning by this point.
You make some valid points on React but to me the issue is the DOM itself.
React offers a consistent unidirectional flow of javascript generated DOM elements.
State -> updated UI
But why do we have to work with the DOM restriction? Why is it not intuitive to animate and position things.
Flash gave a different paradigm with the MovieClip object that resembled more of setting up a movie scene with actors.
Unity3D and Unreal follow a similar style if I'm not wrong.
The DOM with it's restrictions offers a more structured UX, but it was never designed for fast and many content changes.
Maybe the next thing is a Canvas UI framework or maybe the AI Code would evolve to a point that we merely describe the UI and it is auto generated.
Who knows but for now the functional react way is a solid paradigm.
I think the switch to web components has already started at places like Google. I was thinking about web components for my next project (having used React, Vue, Angular, etc. before but more recently just plain HTML and JavaScript usually), but the available components registry/site that I saw kind of seemed like a garbage dump. Then I was thinking about what I want, such as a resizable panel, a drop down menu, and draggable windows maybe. Its honestly easiest for me to just use regular HTML with something like split.js and plain-draggable.js and then the web component that I saw for the menu actually said to put 'summary' and 'details' in the markup.. and since I already have that working without any component.. I think I am just going to skip the component stuff it and stick with my old-fashioned ways.
But anyway I think that within two years, React will be uncool. Web components will have their day in the sun. But within another two years, it will all be about web assembly and actually Chrome and Firefox will start to be uncool. Instead we will separate out the information browser from the application delivery, which will be handled by web assembly which will get some UI stuff added so its not tied to the rest. And if they have any sense at all then they will start unbundling all of the other device APIs that are in the browser so that people can implement pieces of it and we aren't all globally beholden to one company on their implementation of a monolithic overlay operating system.
Agreed Chris I am doing react currently on several projects and I am the a hole in the room that brings up the question of why react? I could see if I had hundreds of things being simultaneously updated but 99% of the time I don’t and it’s just stuff that I can easily do with vanilla JavaScript I said when I learned react 4 1/2 years ago what happens when they just update how the DOM currently works?
yes, it's time to switch back to Angular 1.
the greatest framework known to man.
Because web browsers were never supposed to run dynamic apps loke mobiles do. Web browser is just a static html viewer.
And browser vendors are very slow to change because they are afraid that drastic changes will make them lose their users.
no, it's bc changing takes time and effort. a browser is a very complex software. you wouldn't be able to write one entirely on ur own even of u wanted to.
also even if they changed all of a sudden the adoption of those new standards takes even more time to catch on
Imo the biggest barrier will be replacing internal design systems.
1 month ago, "Why you should learn React". That was an eventful month it seems 😂😂
C is like 40-50 years old and your phone probably have Linux/FreeBSD based OS that is 90% written in C.
probably? all major mobile OSes are unix based. iPhones use Darwin (old fork of freebsd kernel) and Android uses a patched version of Linux. All the tooling needed to interface with the underlying hardware are either C or C++ (nowadays Rust is getting in on the fun as well but not as widespread) and the app talks to the native API in some way (GraalVM on android, idk if apple does native on iphone) so it can do something in there while being sandboxed
I tried to use web components, but it was an absolute pain. Biggest selling point (shadow dom separation) became a biggest problem, because I struggled to manage ui consistency. The API gives you nothing useful and you need a framework over it to make it useful. So how's that better than react?
I mean, as bloated as Angular is, people still use it; so I definitely don't see React going anywhere anytime soon. When it comes right down to it, make sure you learn your fundamentals and understand how to code in vanilla.
Frameworks should be tools and scaffolds to use in a language, not a language into and of themselves.
What viscose theme is that
Hi Chris, do you have link to your Web Components course or any plans for one? I'm hoping to skip the React stack.
I think react has 10 more years. It's the jQuery of our time. No framework is perfect but for now it's the King. I feel it needs a rewrite though. They need to go back on the drawing board and bring in everything, routing, fullstack, state management, and steal some ideas from angular,
What's wrong with regular ol javascript? For 95% of projects, standard js, html and css will get the job done.
Last time I worked in PERL, it was for a site that was updating to MOO when other sites were ditching PERL entirely (such as Amazon and Zappos). I didn't work there for long. I loved the company, loved the co-workers, but we were trying to adapt old tech to new paradigms that were dysfunctional together. We should have been rewriting it in Python or splitting out the backend from the site application, and it was constantly frustrating.
It's been time ditch React for the past 3 years
To be completely honest, coming from vue (v2), I gave react a chance and so far I'm loving it. Probably because I went with Nextjs and the sheer amount of libraries react have is excellent. I'm yet to try react-query + Zustand but I'm hearing great things regarding these libraries.
I opened up a nextjs project for saleor and it is just so unnecessarily complex, you can do the exact same thing in vue and it’s 10 times cleaner
HTML Re-usability was a solved problem since the XHTML/XML/XSLT days.
Everything that came after is people thinking the status quo is too complicated, making something simple, then realizing that there's a need for complexity after all and bloating their framework until it's more complicated than the previous status quo.
Just started learning React 3 days ago. How is React now in comparison to today? How improved is React?
I've never understood why frameworks are so popular, especially nowadays. If it takes years to learn the framework then why not just learn vanilla JS? Is vanilla JS really that bad to work with?
You end up basically rewriting your own framework that is assuredly worse than React
React is just a sign that Javascript is fundamentally unfit for what needs to be used. Web Components mostly suffers from the same problem since it doesn't actually fix the underlying problem which is Javascript.
Kinda why I prefer the backend, the front end seems so bloated nowadays with all the stuff that gets jammed into it.
I have a dilemma/catch 22 with coding/Programmng I do it as a hobbie. I currently work as an operations technician on an government controlled businesses. Background as acarpenter. But always found the joy snd peace with computers, everything self thougt, I am google certified Data analyst an it-support pro. I am afraid that if I switck from current work I will loose the fun wiith computers and data processing. But in the same tkme i rather work with something more computerrelated. Computers has been a part of my life snce mid 90's
I would say knowing two different industries is quite helpful given the probable recession we're headed into.