@@t3dotgg me leaving and never coming back to this video why did you do that to astro Astro is so much better from how it looks from outside rank astro top tier plssssssssss
I was about to say the same, simultaneously working on different projects right now and unfortunately all of them use a different UI, React, Svelte, Vue, Next and Angular, and for some reason I like angular the most because I understand it the best, find stuff in it the fastest and generally speaking I have less things to worry about because a lot of things are already in pace and defined, and documented very clearly. I don't know... it just feels right for a big application, the start is rough but once you are on the way its just so good to use. Unlike React. Started with that, and grow to hate it the most, I was fine for my home apps but I'm hating it more and more every time I get a ticket on that project. Not pleasant at all. (But that might just be due to the codebase, but the state o the codebase might just be because of React, who knows, not me.)
Vue/Nuxt way is the only abstraction that feels natural to me. When I started exploring frontend world i started with React, and felt like I could get used to it. But then I had to do project in work using Vue Composition API and things immediately clicked. It gives perfect balance between ease of use and how much you can do with it without using tricks. And Nuxt gives the real benefit of frontend and backend being separated without PITA of having to keep backend and frontend version of responses.
I went from Angular.js so the template syntax to me wasn't a thing. And take a HTML template and just paste it in a vue component and just change what you want is priceless
I couldn't agree more! I worked with all the major frameworks (Angular, React, Solid and Vue) but Vue was the most joyful experience I've ever had! We're using React on job, atm but I'll try being able to write my own automation project for health sector in Vue. Long live, Vue!
React is just faster, I guess it is preference, you can write performant code in Vue too. React is more innovative and follows better design principles as well, but with Nuxt, it is definitely "easier" to train devs and build apps using it, that honestly works just fine. But if we are talking on a technical level, React is simply superior, especially combined with Astro for static code generation, damn.
One of my favourite things about Nuxt is that it's not a one huge monolith, but rather a bunch of things composed together to create what is Nuxt. Everyone can use those parts in whatever they're building, as almost all of it is available as its own package. Want a server they use? It's nitro(pack). Want auto imports? It's unimport (which is an unplugin, means it works in Vite, Webpack and others). Want load configs like Nuxt? Look at c12. And so so much more. Nuxt existence benefits even the people who don't directly use it (and Nuxt is benefitted by this as well). And you are totally right about the community around Vue, it's a lot of amazing chill people from all around the globe just making neat things without drama or some hardcore competition. Truly lovely stuff.
it is defnitely a monolith, what do you mean? React is just a library, it's not even a framework, has better design philogosophy and runs faster. You can write performant nuxt apps too ofc and theoratically it will be easier to train a group of devs in nuxt faster and ship applications. But on technical side of things, React is just so much better.
@@ahmedifhaam7266 as long as react continues to go literally and directly against what native html provides so no, not a better design philosophy; in fact, react is the antithesis to good design philosophy, just like tailwind. as long as react refuses to align with how native html works in the dom, it will always be the outlier when it comes to progressive enhancements for production builds. both html and css continue to evolve at a rapid pace now and it continues to become quite clear that in the very near future of our developer careers, teams will be able to build robust online apps without leaning so heavily on javascript and/or any lib or framework in order to have a viable existence.
Alpine is a framework to complement HTMX, they're not a replacement for each other. With HTMX you can't do client interactivity that is not bound to the server data, things like modal windows, accordions etc. Alpine and Hyperscript add interactivity on the client side.
You can't say that AlpineJS and HTMX are a single framework (it's cheating) and combining HTMX with AlpineJS does not make HTMX better. Not to mention that you have 2 behaviors working in the same way and at the same time that they don't talk to each other. HTMX is so focused on request-based behavior that I don't understand how they call AlpineJS a complement to HTMX.
@@foot8036eh, I don’t think alpine complements HTMX, but using both complements the UX and developer experience. I rather reach for an alpine function for like toggles and pop ups instead of making a round trip to the server. They complement each other well just by their nature of simplicity in a world of JS bloat.
Ya Alpine does way more than HTMX. As a long time user of Alpine, HTMX doesn't come close to the experience of Alpine and I think shouldnt be mentioned in the same sentence.
@@Nehpets94 The main point is that many comments try to rank AlpineJS and HTMX together as a single framework, which is wrong. "HTMX deserved a better rank because combining with AlpineJS..." In my opinion, as a long-time full-stack programmer and AlpineJS enthusiast since 2020, if you don't work directly with programming, HTMX could be a good choice (for you).
Vue had a lot of ties from the Laravel community. I started on react class component as a junior/intermediate dev, hated the function component at first, switched to vue for SFC.... never came back. Altough I've recently nedded Vue's JSX support for some edge cases. (Sorry for my english... i'm not a native speaker!!)
I feel vue is kind of simple, the "separation of concern" in a Single File component makes it, in my opinion pretty clear of what is what. I come from a website building background (altough I've evolved) and I know my css prettymuch flawlessly. So having a , a and a tag makes the code clearer in my opinion (Its much like a Codepen project in your app) than an mix of HTML in JS (which can be powerful from time to time I must admit). The css scoping in vue make managing style easy, doesn't relly on CSS Modules or the use of tailwind (which I still use for most of my components, thanks to Daisy UI) . Scss works out of the bat if you want it. I've tried svelte, and at first I tought the syntax was like Vue but even more clear. But in the end, it's reliance on a compile step make it such as you must even make your code more complex to some area where other framework make them direct. Solid could be a Solid (!) contender if you don't mind the JSX syntax, which in my opinion solves a lot of React problems.
Disappointed Angular isn't higher. With the recent advancements (e.g. signals, standalone components, deferred loading, etc), it's come a long way in terms of DX.
this is a tier list in terms of what javascript framework to use when you follow easy todo app tutorials not what framework to use when building complex enterprise ready applications that require to be stable, reliable and maintainable. i dare everyone to build a complex web application in jquery it will fail after some years when the workers shifted and no one has a clue what's going on anymore. the same with every newly hyped javascript framework that literally changes it's interfaces more often that i change my underwear.
@@amilismurfs Dude nobodys going to create a web framework with JQuery but for its impact towards the web its S tier is deserved. Jquery is still a better API than vanilla JS imo
@@amilismurfs The best frameworks make the hard things easy and the impossible things possible. JQuery did the first, and Angular did the second, and they both deserve credit for it, but neither of them are perfect. I personally still stand behind React and Angular being the best frameworks on this list simply because they adhere to general programming principles rather than the smaller set of "frontend web development" principles a lot of frameworks like Svelte exclusively stick to, but I also don't think they are better for building small client-focused apps like todo lists than Svelte.
I think he doesn't understand that Alpine isn't a variant of htmx it's a technology that compliments htmx. Also that Lit example shown doesn't work in a browser because of the decorators. It needs TS.
Yea, would've liked if he took a bit more than 40 seconds to look into tools he hasn't really checked out before. I'm also confused why he says Astro isn't a framework, because "it doesn't have routing"? It has as much of routing SvelteKit has, for example. Astro can also do dynamic routing, redirects, paginating, etc. It's a tier list video, but a lot of Theos viewers will take his word as truth and assume Astro isn't a framework and that Alpine is a HTMX like framework. Even if it's a silly bait video to get interactions with the placements, I'm confident at least a good few viewers will now confidently think these things because Theo said so :p
Angular at D? Lol I’ll take Angular any day over React. Maybe it’s not as easy to dive into from scratch but once you’ve got it figured out it is extremely well organized and for that I can navigate working through an Angular repo with my eyes closed.
@@geniusmaker6438 I stopped watching as soon as he ranked Angular at D lol. I should’ve at least panned through to the end to see the end product though.
Astro is an ssg or a ssr framework like nextJs with html component. It is best for getting granular control like vanilla and jquery while maintaining component based architecture and ssr ssg with other frameworks like react. S tier for me (personal opinion 😅)
HTMX and Alpine are very different overall. I don’t think they necessarily “compete” in the same arena. Ok, there's some similarity in that they are both based on composing behavior directly into markup via attributes. And there is a bit of overlap in functionality (e.g. x-on:click vs hx-on:click). However, HTMX is focused on interactions with the server via HTML. When using them together, we will primarily rely on HTMX to communicate with the server while using Alpine to provide richer client-side interactions and (if necessary) manage some transient state related to just the UI.
HTMX + Alpine can be a very attractive combination for backend developers to provide a high-quality UX while using their preferred languages on the server or even for those frontend developers who are no longer so excited about current trends and just want the simplest stack to build on their web apps. There's room for all tools. New tools are always welcome so that the industry can find new paths and improve the state of the art. We don't need to censor people for using any technology. Let people use what they like and be happy :-)
You don't need to create refs for v-for in vue, if you don't need reactivity, it can use plain javascript arrays. we've dropped next entirely for nuxt, the dx is way better for full stack.
@@carldrogo9492 "we" = who the framework is aimed at: developers. If you don't use reactivity with Vue you're doing it wrong. By asking this question I assume that you are not a developer.
Alpine.js is great for having reactivity with javascript expressions in HTML; for more complex web apps it can be used outside of HTML but it hasn't a native router and the debug experience with errors is not good so probably it has not the same goals than others SPA javascript frameworks, so in a "competition" I understand it can't be at top.
I think of Alpine as Tailwind for JavaScript. No more writing separate JavaScript files every time you need simple user interactivity. Although there is some crossover between Alpine and HTMX as far as fetching data, they really serve two different purposes and I think go great hand in hand, especially when the file size for Alpine is like 7k. If you had to write separate JavaScript files for every user interaction you needed, the file size would grow exponentially and you would have way more server requests.
I think you should really look deeper into Qwik (maybe you have overlooked something there), do you really think React Server Components are a better way. I would be curious to see a comparison of them from your point of view.
As a junior dev who is still in university I for sure don't have enough experience to determine the best framework, but I honestly don't understand why Angular is always rated so low. Recent updates move Angular in a great direction and I kinda prefer separating the html from the js compared to React and such approaches. Gonna check out some of the others in this video aswell :)
yeah, Angular got better recently, that's so true, as well as it still has its fundamental problems, such as basic code complexity (boilerplate) in all kind of things, starting from basic syntax to any structures. Any way react sucks ) Suck a 🐕 sh*t library
I don't think a lot of people have used the new angular they're hyping right now, but are kinda basing it on past experiences. Framework popularity isn't entirely based on how good things are. It's based on social, economic, and ideological factors. I mean, React is not particularly good, but it's the most popular one in the west.
He has strong negative bias towards angular which is baseless. There is a reason why angular is #1 big corporate framework. 1) It has everything you need built in, 100% of your knowledge is transferable between projects. You dont have re-learn 20 libraries for each project or argue with the team which one to use. 2) its always strongly typed, you barely need to debug your code, it just works. No random runtime errors. 3) its not magical. There are some boilerplate exactly so you can see what is called, why and with which types so you know exactly how framework does things. I've worked with former react devs and they write most disgusting code I've ever seen because they got used to "it just works" mentality. It can be annoying to tinker with typescript for days to make your stuff work but in my experience, every type this happend it was due flaw in the design itself. 4) Angular is opinionated, framework tells you how to write code and how to structure your app. So you dont have to lose your mind reading other people's code. React is fun only if you are the only dev in the team.
Angular is scalable, well-architected, and focused on each and every aspect html,js,css separately. It is integrated, and disciplined. I think the only down side to it, is relatively long learning curve. That's why people struggle with it. I used it for 7 years now, and I never gonna stop. It is way better than shitty, mixed code which is not html nor JS called JSX.
I actually did relatively large projects, with large set of components and modules, and a well-define architect. The life span of projects are more than 3 years by now. So, I know what is component, and how you can develop one in a clean, readable code , with actual separation of the concern for each section of a component. @@yajirushik2871
Some good points, and glad to see the acknowledgment of being a react fanboy. The only thing I would change is bump Angular to A, and put JQuery at D. JQuery shouldn’t get a pass because of its usage and history. It’s awful to work with. As for Angular, the framework has a lot of great positives which we just glossed over because Theo is so biased against it.
I am a React dev, never used anything else. Recently I went over Vue's docs and thought to myself 'Vue is actually better than React'. I also went over Angular's docs and thought 'I would never use this'
@@draganaleksic6071 I can definitely appreciate that something like Angular is not for everyone and it doesn't work well for every use case. The are so many factors that go into choosing a framework for a given project or environment. It's not realistic that React is always the right answer. Theo is just so hardline biased that he doesn't really ever acknowledge the strong points of Angular. All frameworks have their strong and weak points.
I can’t believe in defending Angular in what is almost 2024… That being said, Angular + standalone components and Tailwind is a great DX. Angular nailed state management but there was so much extra bloat that just slowed down the entire process. I personally like Solid the most, and I think the Angular ecosystem is a bit lacking, but it’s gotten a lot better in the last three years, especially with the new signals.
also if you want to loop over some array, you can just create a {} expression which will evaluate js inside it, so you can use array.map etc, if you don't want to use #each syntax
@@hamm8934 Well exactly, he isn't judging the demand. He judged Angular on how horrible it feels to use it, which is accurate imo. God forbid people actually care about enjoying what they do and not being in it solely for the money.
@@Todiros ive worked on everything from visual basic, to c, to go. You get used to whatever you use. Preferences are silly. Just focus on the power and demand of the tool. The rest is a distraction for devs that are customers for saas companies like vercel that prey on new devs fumbling through coding based on preference.
I see Qwik kinda of as htmx + react, it has the same performance of htmx, where interactions are encoded in html, but you get to write it in a simple composable way like you do react, which is pretty sweet. so if you like htmx you like Qwik. with htmx you still need a backend framework to render html for different components on different api endpoints.
Completely different. Htmx is a frontend helper framework for writing les Javascript. Qwik is a ssr framework which sends small chucks of js on every request
Flutter... F? How so? On some metrics maybe, maybe not performance, but the fact that you don't have to touch JavaScript alone makes the dev experience better.
I clicked the video because I saw flutter in S in the thumbnail. I was curious since I love flutter and honestly have done web much recently besides flutter web, but I would have expected like B teir at best for a 'web framework' tier list. Because it's a bit janky. I think it's a great decision when you are making a cross platform mobile app and want web too and are willing to compromise a bit on the web side. Seo is harder, it has quirks you have to work around. But it's also web for nearly free rather than maintaining two code bases. That gives it value I'd say, but if we ignore that, I could see D tier. It depends how you feel about JavaScript/typescript and what you're trying to do. B for situational and pros/cons D for being a poor decision in the wrong situation
@@ChaseAnderson5k Because websites are supposed to be accessible and index-able but that's not possible with Flutter. It is a great framework for mobile apps but we aren't making mobile apps on the web. We are making websites and the goals are different. It almost feels unfair to put it on this list, different goals and problems to solve.
I'm pretty "vanilla" when it comes to these frameworks, because for most of my use-cases, I just don't have any need for the brand new functionality. Usually I just enjoy either React or Vue in their simple form, but lately I've been learning Next, and it's quite excellent at what it offers
Honestly, this video seems like a troll video, I didn't follow the whole video, I skipped to the end and saw JQuery & Solid on S, Angular on D, Astro & React on F... If this is serious... lads, this video is why you shouldn't ever base anything on the "youtube programmers".
@@jasinrefiku2914you didn’t listen to the intro 🤦🏻♂️
Год назад+15
I'm using Remix professionally and there is no ambiguity about how the components are mounted and what data goes where. I think the confusion here comes from mixing "routes" with "components". As they look very similar. Main difference is - routes are the routing mechanism and have nothing to do with reusable components. It's by convention similar to Next. If you need to re-use component you do it the same as anywhere in React, just extract it from the route first - copy/pasta in to a new file, define the props and re-use it as much as you want. No ambiguity or inconsistently behavior.
Very new to Remix here! But i think he meant that the could not be extracted to a component as it is dependent on the actions function which is only defined on the router "component". Or is it possible to have a "UserCard" component for example that fetches its own data and can do its own mutations?
Год назад+1
@@MrRandomgamerdkHD yes it is possible and even in severally ways I'd say. I prefer using routes In the way Theo showd most of the time. Makes it real easy to see what's going on in the page. For reusable bits there are several options. You of course do norml fetch... But I prefer a "headless" route. So a route that won't render any component, but I can connect to a form. So for instance I have an address that I can add in several places if the app, same form, same be logic. That's connected in this way. The one difference is - using useFetcher form instead of normal form. As it's easier in the code to connect the form with the route. Makes a bit more sense in this kind of scenario.
@@MrRandomgamerdkHD can't share code from work, but there are some examples on using fetcher in Remix RUclips channel. As for an exact example of what I described. I might have a moment to put something together in coming days
It's funny how Theo calls you "Remix Brained" if god forbid you try to embrace web standards like every other fullstack engineer out there who's not on React juice, but goes ahead fanboys about Next every 2 days with all the shit that App Router.
You really should give Angular another shot. D is ridiculous, the developer experience is SO great nowadays - will get even better with broad adaptation of signals. I guess you only have time to scratch the surface to even have a chance of knowing all those frameworks, but in your usual real world project Angular is brilliant for many reason. One is high quality component libraries that actually work flawless, unless most libraries for other frameworks. Another is the build process just works perfectly out of the box. RENDERING performance is sooooo useless for 99,9% of applications - web applications are nearly always bottlenecked by network calls anyways and after that 300ms call it does not matter if your framework renders in 1ms or 2 (and this is already a long stretch as difference between a "slow" and a "fast" framework is certainly not 100%....
I picked react as my first framework/library thinking about going for Angular next since I want to at least know one stable language suitable for big projects
As an avid Svelte fanboy I agree with a lot of your critiques of Svelte. Components are easy to write at first but it can get complicated quickly. However, don’t underestimate the readability of Svelte once you’re in a Svelte codebase for a while and learn the patterns. Supposedly Svelte 5 is supposed to address a lot of these concerns so I’m excited to see what Rich and the team cook up. Also this video makes me want to code a small web app in jquery just to try it out lol
While I do think Svelte is improving, I unfortunately still feel like 90% of the reason behind the hype of Svelte was something along the lines of "wow you need to import useState and write 300 lines of code in React just to get a reactive variable?? Gross! Svelte does it in one line!", so I'm interested in what the response will be to Svelte introducing a more predictable system with their "runes" thing. I do agree that it's a better system, but I also think that once Svelte 5 fully releases, a lot of the initial hype which came from the "I Hate Everything Related To Javascript And The Web" crowd will disappear.
@@okie9025 Svelte 5 introduces similar design principles anyways at least when it comes to dx (under the hood runes still use the same reactivity). They really had a good product and went "Remember the whole compiler we made ? Let's not take advantage of it, that'll show them!" 😅
@@charlesm.2604 I kinda understand why they didn't want to go with the custom JS compiler though, because the last thing the web needs is "hey, here's this framework that's 'just javascript!' but in reality it has a bunch of quirks compared to normal js that you have to look out for". The web is stuck between a rock and a hard place - modifying the existing JS syntax for your own needs is either frowned upon (eg. coffeescript) or straight up impossible (backwards compatibility), while improving JS in secondary ways like type safety and tooling can only go so far before plateauing. I think the only solution is to have more advanced languages be supported on the web in the same accessible way as JS. WASM is cool until you realize the build system is slower and more complicated than existing JS tools.
@@okie9025 Of course but if not for the abstraction possibilities that the compiler can offer then what is the point of Svelte's existence? You get better performances and reactivity with vapor, a better ecosystem of integrations and libraries with react, a better collaborative experience and the ability to ship more robust applications with Angular. Not only is Svelte not solving any problem but it makes the development process more tedious because there is barely any community effort to write and maintain *much needed* libraries. Svelte was an amateur's best friend, I'm not sure it's still the case...
react in the list is like the best friend you love so much that if you express it, it will bring out too many emotions, so its just easier to bash it and be playful
I am at 3:36, this tier list has to be an joke, hasn't it? You placed jQuery at the top with the reasoning of being impactful and widely used. With the same reasoning, but the only addition that you personally don't like it, you place angular at D tier. Alpine goes, after looking half a minute at the docs, into D, too. Astro goes into F you because of being an impostor. I mean don't include it in the list, if it doesn't belong there :D I like your other content, but you have to be trolling here. I'm out.
Alpine is to Javascript what Tailwind is to CSS: a way to have LOB and to get rid of extra files. Comparing it to HTMX is just missing the point completely.
Worth mentioning that preact also has better webcomponent support than react and is smaller than lit. If you want a small framework just to make webcomponents with no build step it is a very good option. Compared to lit, lit has first class CSS while preact has a better reactivity system. You're not realistically going to notice perf differences between them, but you will notice preact webcomponents working well compared to react. And unlike Lit, it isn't maintained by google so you don't have to deal with a new incompatible version of it each year with dropped support for the old one (polymer -> lit-element -> lit -> lit major version numbers). It's basically a more stable version of react and it takes backwards compatibility seriously while still being easy to hire for due to its similarity to react
23:31 I've been using Svelte but I've never seen @const current I'd rather extract the @const's into reactive variables ($:) or something Or, even better, simply create a new list item component and put the row in it
I have no doubt that Theo had a rotten time with a particular job that happened to use Ember, and a job can make or break ones opinion of any framework. :(
His way of deciding if a framework is good or bad by simply Googling it without really use it deeper and further understanding speaks volumes about how unfair the rating is to an unfamiliar framework.
I use svelte and at this point syntax is just a matter of preference. What I would say though, is that with svelte you don't really worry about writing it the "wrong" way. With react you always have to make sure you are writing proper code otherwise you'll be writing slow code or code that just doesn't do what you think it will. If you understand reach really well then obviously you have less of a problem. But with so much to keep in mind at all times you have less mental capacity to think about the actual problem you want to solve. I don't think svelte is ever going to get close to react but like Theo said, I do love that it pushes other frameworks from so many angles. And I especially love that Rich is not afraid to go places other frameworks can't/won't go.
No hate but you’re off when you say you can’t isolate away code from Vue components in the same way you can in react. I’d argue Vue with the composition API is way more flexible than React in that sense. You infact can have reactivity outside of Composables (Vue hooks) and components if you really wanted to and import it into composables and components as needed I like react, don’t get me wrong, and I even think it’s probably easier to learn than Vue is. I do think Vue templating is leagues upon leagues better than templating in JSX, that always feels like a much less pleasant experience / less readable (in my opinion)
i find it really helpful to learn by drawing on other peoples conclusions and summaries. (because i can skip rabbitholes) thanks for a really genuine and justified opinion.
RE: Lit "I don't think any of the other frameworks here have that capability where you've now created an element you can access in the browser". Really makes me think he doesn't know enough about Angular to grade it after that.
as someone who has desperately been seeking a true successor to knockoutjs thank you for this bc I had never heard of solidjs and I think it's the framework I've been looking for
Having started with React, moving to Svelte and Vue felt infinitely better from a developer experience pov. The React ecosystem’s its biggest pain point.
I'm yet to learn react. I have built a hand gesture recognition. Webapp using only vanilla js. Should I rewrite it using a framework(code is clunky) so which one should I learn for 2025, react or solid?
You can reuse any component logic across components in Svelte by making those properties global or by using binding methods available, Svelte is easily the best one in this list, solid is also good but different
10:02 hydration isn't a too big deal? I thought that's why you were very excited about server component, if it's not a big deal, what's the advantage of server component then?
I don't understand why people can't get over the Vue and Svelte templating. By that logic, all templating in every programming language is bad, but this is just a bad take. I hate JSX because I hate mixing JS with my markup. I'd rather have attribute based templating instead. It's so much easier to read! Also, hooks will forever suck!
dude alpine is sooooo convenient. I used to use jquery on sites I was making, when I want to just use a little bit of interactivity.. .now it's always alpine for those use cases... so nice.
Would love to hear your thoughts on Storybook -- yes, please make a video on that! Also, I'm getting started with the RedwoodJS metaframework, would love your thoughts on it as a whole as well. (it uses Storybook, also Prisma/GraphQL and some funky CLI tools to create data "cells"... super curious if you think this is useful or a bad idea.) [I'm working on a new side project, very experienced backend engineer but first time using JS. I'm building something with a lot of data entry components, complex UI, and strong security needs since it's FinTech, so I'm a little paranoid about getting things right from the start.]
Jquery at S tier is wild. As someone who started as a Flash developer before Jquery was a thing, I just don’t get it. Just because lots of sites use it is like saying most of the web is Wordpress, so Wordpress is S tier.
@@evancombs5159 just create your some utility functions.. then you don't need to include a painful JQuery CDN.. in last 3-4 years. almost no body using it.
I am a backend developer and I use jQuery in my current project. Not in all projects sophisticated frameworks are needed. If you know how to use vanillajs, jQuery and html templates, you have the freedom to create components and to each thing what you want.
@@hugochavez6170 of course you can use Jquery but you can use Handlebars too but why would you? I do frontend work & used Jquery for years I just have no desire to go back to it.
How did Astro get categorised in F and jQuery S??? And why the hell is Flutter in this list? Since when is Flutter a JavaScript framework? Oh, I completely forgot that he really wanted to give a side kick. But it's good that the F category is highlighted in green. Maybe you have to read the list from bottom to top.
* *makes videos about frameworks* * * *almost everything is not actually a framework* * that being out of the way, thanks for making such a cool video :)
"Angular just follows what others are doing" that's a cute idea.. except the fact that they basically invented the idea of components (called directives at the time), two-way data binding, shadow DOM, and were the first major framework to adopt TS out of the box (which pissed a lot of people off at the time, but eventually the majority of the web dev community ended up following suit). Since then, they have become more of a mature, slow-moving framework (which has massive advantages when building enterprise software), but let's not forget Angular's important role in web dev history
Just the moment I saw he put Angular in D tier, I lost all the hope for people who're watching this video. "It is one of the most painful frameworks", were you talking about React Library? Angular is built for enterprise applications, not a todo app or an alarm timer. Even if you have 10k pages, you can easily navigate through the project. This is why every company out there, have problems with their frontend. This is the reason, you as viewer won't grow. Your tiny fingers can't write Typescript? You call your React function a component? Aww.
By the time a JS person can finish telling you about all the different frameworks they've used recently, they will have already used another dozen frameworks that they will now need to tell you about.
Finally someone else shits on Ember. My first frontend job used ember, and having come already from Angular.JS 1.x, I could not make heads or tails out of what the fuck was supposed to happen at any given time, the documentation was approximately on par with Apple's (which inspired Ember) documentation (bad or missing), and it relied heavily on some arcane convention over configuration. I believe my first burnout happened when I just spent too many damn days trying to build some relatively simple stuff but getting nowhere.
Solid in S tier. Well deserved. I'm using it for everything I do now. I cant' wait for it to hit mainstream. Not a fan how solidstart is also using "use client"/"use server". I preferred the server$ and generally functionThatRunsOnServerEndsInDollar$. Much better than string statements.
I used Ember between 2014-2019. Your review of it was pretty fair. In it's heyday it was an excellent choice: - Batteries included / convention over configuration (vs the disaster of React at the time... this was before create-react-app) - Ember Data (although frustrating) provided a comprehensive data story that other frameworks were lacking - Ember CLI was a breath of fresh air compared with early Webpack - Ember Addons extended your app functionality and were as easy to install as Rails gems - Nested layouts/routes akin to Next.js 13 app folder - IMO the class-based components with the @tracked decorator is cleaner than React's use function mess (although I still love writing React) I think Ember really lost momentum around 2017 after React and Vue took over the ecosystem. Some things I think they could have done better: - Official TypeScript support sooner (it took them FOREVER) - Abandoned Handlebars for JSX
It's funny how I use Vue because of SFC and in the end, they look more like vanilla HTML than when using JSX, which ends up looking like a mess of JS, HTML, and CSS simultaneously.
Everything up high on your list is based on frameworks not able to scale well yet Angular which is the best at that is at D… You mention you don’t like the design decisions behind Angular but once you build a huge project with it you’ll see why they did it like that
0:41 vanilla js
1:05 jquery
1:24 alpinejs
2:35 angular
3:19 astro
3:45 emberjs
6:28 handlebars
7:53 qwik
11:27 lit
12:50 preact
14:52 react/hardly know her
14:57 htmx/not a javascript framework (same thing)
16:37 solidjs
19:11 svelte
24:24 vue
27:35 nextjs
28:54 remix
32:51 gatsby
34:24 nuxt
36:27 flutter/instant S tier (same thing)
37:06 elm/haskell (same thing)
38:26 flash/based (same thing)
39:15 react again
Legend. Thanks.
Seeing flutter as S rank is a kick on the balls.
🫡
ah thanks, I found nuxt 💚
@@t3dotgg if you edit this into the description it will replace the bad (autogenerated?) chapters
So Theo decided to start a war for Christmas...
Ho ho ho!
@@t3dotgg h0h0h0
@@t3dotgg me leaving and never coming back to this video why did you do that to astro
Astro is so much better from how it looks from outside
rank astro top tier plssssssssss
Next but no Nuxt?
@@rajikkali2381 it's added in towards the end (34:26)
Why does modern angular still gets judged on AngularJS experience? I find it easy to maintain large codebases in Angular than in React's ecosystem.
Angular infact is the only Framework the rest are just UI libraries.
I saw him saying "It's been 3 years since I used Classes, I don't use them, it's only for library maintainers". That alone shows why he is bias.
Next is a UI library?@@carldrogo9492
@@carldrogo9492 i think Nextjs is a framework too
I was about to say the same, simultaneously working on different projects right now and unfortunately all of them use a different UI, React, Svelte, Vue, Next and Angular, and for some reason I like angular the most because I understand it the best, find stuff in it the fastest and generally speaking I have less things to worry about because a lot of things are already in pace and defined, and documented very clearly. I don't know... it just feels right for a big application, the start is rough but once you are on the way its just so good to use. Unlike React. Started with that, and grow to hate it the most, I was fine for my home apps but I'm hating it more and more every time I get a ticket on that project. Not pleasant at all. (But that might just be due to the codebase, but the state o the codebase might just be because of React, who knows, not me.)
Vue/Nuxt way is the only abstraction that feels natural to me. When I started exploring frontend world i started with React, and felt like I could get used to it. But then I had to do project in work using Vue Composition API and things immediately clicked.
It gives perfect balance between ease of use and how much you can do with it without using tricks.
And Nuxt gives the real benefit of frontend and backend being separated without PITA of having to keep backend and frontend version of responses.
U mean next combines backend, I.e server components?
@@ib_conceptNuxt
I went from Angular.js so the template syntax to me wasn't a thing. And take a HTML template and just paste it in a vue component and just change what you want is priceless
I couldn't agree more! I worked with all the major frameworks (Angular, React, Solid and Vue) but Vue was the most joyful experience I've ever had! We're using React on job, atm but I'll try being able to write my own automation project for health sector in Vue.
Long live, Vue!
React is just faster, I guess it is preference, you can write performant code in Vue too. React is more innovative and follows better design principles as well, but with Nuxt, it is definitely "easier" to train devs and build apps using it, that honestly works just fine. But if we are talking on a technical level, React is simply superior, especially combined with Astro for static code generation, damn.
So cool to see the Nuxt team get the credits they deserve 👍
I'm shocked.
They deserve to be S tier. And Vue right next to it
One of my favourite things about Nuxt is that it's not a one huge monolith, but rather a bunch of things composed together to create what is Nuxt. Everyone can use those parts in whatever they're building, as almost all of it is available as its own package. Want a server they use? It's nitro(pack). Want auto imports? It's unimport (which is an unplugin, means it works in Vite, Webpack and others). Want load configs like Nuxt? Look at c12. And so so much more. Nuxt existence benefits even the people who don't directly use it (and Nuxt is benefitted by this as well). And you are totally right about the community around Vue, it's a lot of amazing chill people from all around the globe just making neat things without drama or some hardcore competition. Truly lovely stuff.
it is defnitely a monolith, what do you mean? React is just a library, it's not even a framework, has better design philogosophy and runs faster. You can write performant nuxt apps too ofc and theoratically it will be easier to train a group of devs in nuxt faster and ship applications. But on technical side of things, React is just so much better.
@@ahmedifhaam7266 as long as react continues to go literally and directly against what native html provides so no, not a better design philosophy; in fact, react is the antithesis to good design philosophy, just like tailwind. as long as react refuses to align with how native html works in the dom, it will always be the outlier when it comes to progressive enhancements for production builds. both html and css continue to evolve at a rapid pace now and it continues to become quite clear that in the very near future of our developer careers, teams will be able to build robust online apps without leaning so heavily on javascript and/or any lib or framework in order to have a viable existence.
Alpine is a framework to complement HTMX, they're not a replacement for each other. With HTMX you can't do client interactivity that is not bound to the server data, things like modal windows, accordions etc. Alpine and Hyperscript add interactivity on the client side.
You can't say that AlpineJS and HTMX are a single framework (it's cheating) and combining HTMX with AlpineJS does not make HTMX better.
Not to mention that you have 2 behaviors working in the same way and at the same time that they don't talk to each other.
HTMX is so focused on request-based behavior that I don't understand how they call AlpineJS a complement to HTMX.
@@foot8036 ok junior
@@foot8036eh, I don’t think alpine complements HTMX, but using both complements the UX and developer experience. I rather reach for an alpine function for like toggles and pop ups instead of making a round trip to the server. They complement each other well just by their nature of simplicity in a world of JS bloat.
Ya Alpine does way more than HTMX. As a long time user of Alpine, HTMX doesn't come close to the experience of Alpine and I think shouldnt be mentioned in the same sentence.
@@Nehpets94 The main point is that many comments try to rank AlpineJS and HTMX together as a single framework, which is wrong.
"HTMX deserved a better rank because combining with AlpineJS..."
In my opinion, as a long-time full-stack programmer and AlpineJS enthusiast since 2020, if you don't work directly with programming, HTMX could be a good choice (for you).
Vue had a lot of ties from the Laravel community. I started on react class component as a junior/intermediate dev, hated the function component at first, switched to vue for SFC.... never came back. Altough I've recently nedded Vue's JSX support for some edge cases. (Sorry for my english... i'm not a native speaker!!)
I am a junior react dev trying to pick up a new framework how hard would vue be??
@@federicodiaz7280You know that Europe is a continent, right?
I feel vue is kind of simple, the "separation of concern" in a Single File component makes it, in my opinion pretty clear of what is what. I come from a website building background (altough I've evolved) and I know my css prettymuch flawlessly. So having a , a and a tag makes the code clearer in my opinion (Its much like a Codepen project in your app) than an mix of HTML in JS (which can be powerful from time to time I must admit).
The css scoping in vue make managing style easy, doesn't relly on CSS Modules or the use of tailwind (which I still use for most of my components, thanks to Daisy UI) . Scss works out of the bat if you want it.
I've tried svelte, and at first I tought the syntax was like Vue but even more clear. But in the end, it's reliance on a compile step make it such as you must even make your code more complex to some area where other framework make them direct.
Solid could be a Solid (!) contender if you don't mind the JSX syntax, which in my opinion solves a lot of React problems.
For my job, we are currently using Laravel Inertia with vue
Disappointed Angular isn't higher. With the recent advancements (e.g. signals, standalone components, deferred loading, etc), it's come a long way in terms of DX.
this is a tier list in terms of what javascript framework to use when you follow easy todo app tutorials not what framework to use when building complex enterprise ready applications that require to be stable, reliable and maintainable. i dare everyone to build a complex web application in jquery it will fail after some years when the workers shifted and no one has a clue what's going on anymore. the same with every newly hyped javascript framework that literally changes it's interfaces more often that i change my underwear.
Can i get you OF h o e?
@@amilismurfs every simpleton is after the next shinny new flavour of the month Framework and disrespect Angular.
@@amilismurfs Dude nobodys going to create a web framework with JQuery but for its impact towards the web its S tier is deserved. Jquery is still a better API than vanilla JS imo
@@amilismurfs The best frameworks make the hard things easy and the impossible things possible. JQuery did the first, and Angular did the second, and they both deserve credit for it, but neither of them are perfect.
I personally still stand behind React and Angular being the best frameworks on this list simply because they adhere to general programming principles rather than the smaller set of "frontend web development" principles a lot of frameworks like Svelte exclusively stick to, but I also don't think they are better for building small client-focused apps like todo lists than Svelte.
I think he doesn't understand that Alpine isn't a variant of htmx it's a technology that compliments htmx. Also that Lit example shown doesn't work in a browser because of the decorators. It needs TS.
Yea, would've liked if he took a bit more than 40 seconds to look into tools he hasn't really checked out before.
I'm also confused why he says Astro isn't a framework, because "it doesn't have routing"? It has as much of routing SvelteKit has, for example. Astro can also do dynamic routing, redirects, paginating, etc.
It's a tier list video, but a lot of Theos viewers will take his word as truth and assume Astro isn't a framework and that Alpine is a HTMX like framework. Even if it's a silly bait video to get interactions with the placements, I'm confident at least a good few viewers will now confidently think these things because Theo said so :p
I think you didn't understand the purpose of the video.
I also think you don't know the meaning of the word variant.
@@foot8036 Could you explain how Alpine and HTMX are variants of each other for us?
@@cptive Does Astro work well for SPAs? I've only ever used it as a SSG.
He’s a sr developer not a staff arch like primagen give him a break
12:00 Angular has this feature called Angular elements.
I've never seen Theo as angry as when he talked about Ember
You haven't seen him talk about Robert Martin in his video about fake guests in a tech conference.
@@ithinkimhipster502 You're right, I guess I'm going to have to see for myself...
Angular at D? Lol I’ll take Angular any day over React. Maybe it’s not as easy to dive into from scratch but once you’ve got it figured out it is extremely well organized and for that I can navigate working through an Angular repo with my eyes closed.
Yep I am mad he can't even get the new logo.
@@dgdev69 He’s pretty much your meme JS dev who has a new tech stack every other week.
He put Angular above React though
@@geniusmaker6438 I stopped watching as soon as he ranked Angular at D lol. I should’ve at least panned through to the end to see the end product though.
@@DannyMcPfisterMee too
Astro is an ssg or a ssr framework like nextJs with html component.
It is best for getting granular control like vanilla and jquery while maintaining component based architecture and ssr ssg with other frameworks like react.
S tier for me (personal opinion 😅)
Me as well. ASTRO is S tier for me.
S tier for me
+1
I've only ever used Astro as a SSG but all of these positive comments make me want to use it to build an app.
@@semyaza555 yes the islands architecture allows you to choose which component should interactive and which should be static
@debug in Svelte is not a console.log, it is invoking the javascript debugger on the variable, or function that you have specified
HTMX and Alpine are very different overall. I don’t think they necessarily “compete” in the same arena.
Ok, there's some similarity in that they are both based on composing behavior directly into markup via attributes. And there is a bit of overlap in functionality (e.g. x-on:click vs hx-on:click).
However, HTMX is focused on interactions with the server via HTML. When using them together, we will primarily rely on HTMX to communicate with the server while using Alpine to provide richer client-side interactions and (if necessary) manage some transient state related to just the UI.
HTMX + Alpine can be a very attractive combination for backend developers to provide a high-quality UX while using their preferred languages on the server or even for those frontend developers who are no longer so excited about current trends and just want the simplest stack to build on their web apps.
There's room for all tools. New tools are always welcome so that the industry can find new paths and improve the state of the art.
We don't need to censor people for using any technology. Let people use what they like and be happy :-)
You don't need to create refs for v-for in vue, if you don't need reactivity, it can use plain javascript arrays. we've dropped next entirely for nuxt, the dx is way better for full stack.
Yea but reactivity is one of the main things we deal with in an application that uses (for example) Vue.
Who's we?
@@carldrogo9492 "we" = who the framework is aimed at: developers.
If you don't use reactivity with Vue you're doing it wrong.
By asking this question I assume that you are not a developer.
Alpine.js deserves a higher tier.
For backend engineers, it is easier&smoother for them to adapt Alpine.js than completely switching to HTMX.
Arent they supposed to be used together like Alpine for interactivity n htmx for server requests
you're 100% right!
Alpine.js is great for having reactivity with javascript expressions in HTML; for more complex web apps it can be used outside of HTML but it hasn't a native router and the debug experience with errors is not good so probably it has not the same goals than others SPA javascript frameworks, so in a "competition" I understand it can't be at top.
Using HTMX + Alpine has been a nice experience so far for me.
They complement each other very well.
I think of Alpine as Tailwind for JavaScript. No more writing separate JavaScript files every time you need simple user interactivity. Although there is some crossover between Alpine and HTMX as far as fetching data, they really serve two different purposes and I think go great hand in hand, especially when the file size for Alpine is like 7k. If you had to write separate JavaScript files for every user interaction you needed, the file size would grow exponentially and you would have way more server requests.
I think you should really look deeper into Qwik (maybe you have overlooked something there),
do you really think React Server Components are a better way.
I would be curious to see a comparison of them from your point of view.
As a junior dev who is still in university I for sure don't have enough experience to determine the best framework, but I honestly don't understand why Angular is always rated so low. Recent updates move Angular in a great direction and I kinda prefer separating the html from the js compared to React and such approaches.
Gonna check out some of the others in this video aswell :)
yeah, Angular got better recently, that's so true, as well as it still has its fundamental problems, such as basic code complexity (boilerplate) in all kind of things, starting from basic syntax to any structures.
Any way react sucks ) Suck a 🐕 sh*t library
I don't think a lot of people have used the new angular they're hyping right now, but are kinda basing it on past experiences.
Framework popularity isn't entirely based on how good things are. It's based on social, economic, and ideological factors. I mean, React is not particularly good, but it's the most popular one in the west.
He has strong negative bias towards angular which is baseless. There is a reason why angular is #1 big corporate framework.
1) It has everything you need built in, 100% of your knowledge is transferable between projects. You dont have re-learn 20 libraries for each project or argue with the team which one to use.
2) its always strongly typed, you barely need to debug your code, it just works. No random runtime errors.
3) its not magical. There are some boilerplate exactly so you can see what is called, why and with which types so you know exactly how framework does things. I've worked with former react devs and they write most disgusting code I've ever seen because they got used to "it just works" mentality. It can be annoying to tinker with typescript for days to make your stuff work but in my experience, every type this happend it was due flaw in the design itself.
4) Angular is opinionated, framework tells you how to write code and how to structure your app. So you dont have to lose your mind reading other people's code.
React is fun only if you are the only dev in the team.
@@yankotliarov9239nice answer
“Baseless”? Angular was shut before the new version lol
Lit and Vue being so high up 💪 My favorite frameworks
Modern Angular is pretty good as for me. And it's much closer to web standards than React.
Angular is scalable, well-architected, and focused on each and every aspect html,js,css separately. It is integrated, and disciplined. I think the only down side to it, is relatively long learning curve. That's why people struggle with it. I used it for 7 years now, and I never gonna stop. It is way better than shitty, mixed code which is not html nor JS called JSX.
@@arash_vazirilooks like You never builded a components
@@yajirushik2871 Angular is a component based framework, you don't know what you're talking about.
I actually did relatively large projects, with large set of components and modules, and a well-define architect. The life span of projects are more than 3 years by now. So, I know what is component, and how you can develop one in a clean, readable code , with actual separation of the concern for each section of a component. @@yajirushik2871
@@yajirushik2871 how about you component deez nuts
Some good points, and glad to see the acknowledgment of being a react fanboy. The only thing I would change is bump Angular to A, and put JQuery at D. JQuery shouldn’t get a pass because of its usage and history. It’s awful to work with. As for Angular, the framework has a lot of great positives which we just glossed over because Theo is so biased against it.
I am a React dev, never used anything else. Recently I went over Vue's docs and thought to myself 'Vue is actually better than React'. I also went over Angular's docs and thought 'I would never use this'
@@draganaleksic6071 until you actually use it man :) I used vue 2 and I think it sucked
JQuery is legend for simple things, and much of websites is not required more.
@@draganaleksic6071 I can definitely appreciate that something like Angular is not for everyone and it doesn't work well for every use case. The are so many factors that go into choosing a framework for a given project or environment. It's not realistic that React is always the right answer. Theo is just so hardline biased that he doesn't really ever acknowledge the strong points of Angular. All frameworks have their strong and weak points.
I can’t believe in defending Angular in what is almost 2024…
That being said, Angular + standalone components and Tailwind is a great DX. Angular nailed state management but there was so much extra bloat that just slowed down the entire process. I personally like Solid the most, and I think the Angular ecosystem is a bit lacking, but it’s gotten a lot better in the last three years, especially with the new signals.
Yes, Angular might seem to a bit slow when you write it. But these drawback will help you clarify what the hell you are writing in 5 or 10 years.
Svelte 5 has snippets, which allows you to have a similar experience to multiple components per file.
also if you want to loop over some array, you can just create a {} expression which will evaluate js inside it, so you can use array.map etc, if you don't want to use #each syntax
@@phoneywheeze Can you show an example?
Theo: shitting on Angular => noobs not learning it.
Meanwhile Experienced Angilar Dev: 💰 🤑 💸
I swear people that crap on angular like to code, they don't like to make money
@@hamm8934but react is more in demand than angular, it's the first library framework I picked I'm considering going for Angular next
@@hamm8934 Well exactly, he isn't judging the demand. He judged Angular on how horrible it feels to use it, which is accurate imo. God forbid people actually care about enjoying what they do and not being in it solely for the money.
@@Todiros ive worked on everything from visual basic, to c, to go. You get used to whatever you use. Preferences are silly. Just focus on the power and demand of the tool. The rest is a distraction for devs that are customers for saas companies like vercel that prey on new devs fumbling through coding based on preference.
I see Qwik kinda of as htmx + react, it has the same performance of htmx, where interactions are encoded in html, but you get to write it in a simple composable way like you do react, which is pretty sweet. so if you like htmx you like Qwik. with htmx you still need a backend framework to render html for different components on different api endpoints.
Completely different. Htmx is a frontend helper framework for writing les Javascript. Qwik is a ssr framework which sends small chucks of js on every request
Flutter... F? How so? On some metrics maybe, maybe not performance, but the fact that you don't have to touch JavaScript alone makes the dev experience better.
I clicked the video because I saw flutter in S in the thumbnail. I was curious since I love flutter and honestly have done web much recently besides flutter web, but I would have expected like B teir at best for a 'web framework' tier list. Because it's a bit janky.
I think it's a great decision when you are making a cross platform mobile app and want web too and are willing to compromise a bit on the web side. Seo is harder, it has quirks you have to work around. But it's also web for nearly free rather than maintaining two code bases.
That gives it value I'd say, but if we ignore that, I could see D tier. It depends how you feel about JavaScript/typescript and what you're trying to do.
B for situational and pros/cons
D for being a poor decision in the wrong situation
@@ChaseAnderson5k Because websites are supposed to be accessible and index-able but that's not possible with Flutter. It is a great framework for mobile apps but we aren't making mobile apps on the web. We are making websites and the goals are different. It almost feels unfair to put it on this list, different goals and problems to solve.
I'm pretty "vanilla" when it comes to these frameworks, because for most of my use-cases, I just don't have any need for the brand new functionality. Usually I just enjoy either React or Vue in their simple form, but lately I've been learning Next, and it's quite excellent at what it offers
Putting Angular on D tier is criminal
If it wasn't so boilerplatey and constantly 5 years behind on innovation, it might actually be competitive with the others.
Z
Honestly, this video seems like a troll video, I didn't follow the whole video, I skipped to the end and saw JQuery & Solid on S, Angular on D, Astro & React on F...
If this is serious... lads, this video is why you shouldn't ever base anything on the "youtube programmers".
Yeah it should have been in the F tier
@@jasinrefiku2914you didn’t listen to the intro 🤦🏻♂️
I'm using Remix professionally and there is no ambiguity about how the components are mounted and what data goes where.
I think the confusion here comes from mixing "routes" with "components". As they look very similar. Main difference is - routes are the routing mechanism and have nothing to do with reusable components. It's by convention similar to Next.
If you need to re-use component you do it the same as anywhere in React, just extract it from the route first - copy/pasta in to a new file, define the props and re-use it as much as you want. No ambiguity or inconsistently behavior.
Very new to Remix here! But i think he meant that the could not be extracted to a component as it is dependent on the actions function which is only defined on the router "component". Or is it possible to have a "UserCard" component for example that fetches its own data and can do its own mutations?
@@MrRandomgamerdkHD yes it is possible and even in severally ways I'd say.
I prefer using routes In the way Theo showd most of the time. Makes it real easy to see what's going on in the page. For reusable bits there are several options.
You of course do norml fetch...
But I prefer a "headless" route. So a route that won't render any component, but I can connect to a form. So for instance I have an address that I can add in several places if the app, same form, same be logic. That's connected in this way. The one difference is - using useFetcher form instead of normal form. As it's easier in the code to connect the form with the route. Makes a bit more sense in this kind of scenario.
@ Nice, do you have a example of this?
@@MrRandomgamerdkHD can't share code from work, but there are some examples on using fetcher in Remix RUclips channel.
As for an exact example of what I described. I might have a moment to put something together in coming days
It's funny how Theo calls you "Remix Brained" if god forbid you try to embrace web standards like every other fullstack engineer out there who's not on React juice, but goes ahead fanboys about Next every 2 days with all the shit that App Router.
lol, Merry Ragebait! You got me for a while there, Mr. “Flash and jQuery are S-Tier” Theo 😅
Y’all are being _trolled_ for Christmas. 🎁
You really should give Angular another shot. D is ridiculous, the developer experience is SO great nowadays - will get even better with broad adaptation of signals. I guess you only have time to scratch the surface to even have a chance of knowing all those frameworks, but in your usual real world project Angular is brilliant for many reason. One is high quality component libraries that actually work flawless, unless most libraries for other frameworks. Another is the build process just works perfectly out of the box.
RENDERING performance is sooooo useless for 99,9% of applications - web applications are nearly always bottlenecked by network calls anyways and after that 300ms call it does not matter if your framework renders in 1ms or 2 (and this is already a long stretch as difference between a "slow" and a "fast" framework is certainly not 100%....
I picked react as my first framework/library thinking about going for Angular next since I want to at least know one stable language suitable for big projects
As an avid Svelte fanboy I agree with a lot of your critiques of Svelte. Components are easy to write at first but it can get complicated quickly. However, don’t underestimate the readability of Svelte once you’re in a Svelte codebase for a while and learn the patterns.
Supposedly Svelte 5 is supposed to address a lot of these concerns so I’m excited to see what Rich and the team cook up. Also this video makes me want to code a small web app in jquery just to try it out lol
While I do think Svelte is improving, I unfortunately still feel like 90% of the reason behind the hype of Svelte was something along the lines of "wow you need to import useState and write 300 lines of code in React just to get a reactive variable?? Gross! Svelte does it in one line!", so I'm interested in what the response will be to Svelte introducing a more predictable system with their "runes" thing.
I do agree that it's a better system, but I also think that once Svelte 5 fully releases, a lot of the initial hype which came from the "I Hate Everything Related To Javascript And The Web" crowd will disappear.
jQuery is garbage this dude is stupid
@@okie9025 Svelte 5 introduces similar design principles anyways at least when it comes to dx (under the hood runes still use the same reactivity). They really had a good product and went "Remember the whole compiler we made ? Let's not take advantage of it, that'll show them!" 😅
@@charlesm.2604 I kinda understand why they didn't want to go with the custom JS compiler though, because the last thing the web needs is "hey, here's this framework that's 'just javascript!' but in reality it has a bunch of quirks compared to normal js that you have to look out for". The web is stuck between a rock and a hard place - modifying the existing JS syntax for your own needs is either frowned upon (eg. coffeescript) or straight up impossible (backwards compatibility), while improving JS in secondary ways like type safety and tooling can only go so far before plateauing. I think the only solution is to have more advanced languages be supported on the web in the same accessible way as JS. WASM is cool until you realize the build system is slower and more complicated than existing JS tools.
@@okie9025 Of course but if not for the abstraction possibilities that the compiler can offer then what is the point of Svelte's existence? You get better performances and reactivity with vapor, a better ecosystem of integrations and libraries with react, a better collaborative experience and the ability to ship more robust applications with Angular. Not only is Svelte not solving any problem but it makes the development process more tedious because there is barely any community effort to write and maintain *much needed* libraries. Svelte was an amateur's best friend, I'm not sure it's still the case...
react in the list is like the best friend you love so much that if you express it, it will bring out too many emotions, so its just easier to bash it and be playful
Nice to see Nuxt getting the love it deserves
I am at 3:36, this tier list has to be an joke, hasn't it? You placed jQuery at the top with the reasoning of being impactful and widely used. With the same reasoning, but the only addition that you personally don't like it, you place angular at D tier. Alpine goes, after looking half a minute at the docs, into D, too. Astro goes into F you because of being an impostor. I mean don't include it in the list, if it doesn't belong there :D I like your other content, but you have to be trolling here. I'm out.
Alpine is to Javascript what Tailwind is to CSS: a way to have LOB and to get rid of extra files.
Comparing it to HTMX is just missing the point completely.
My thoughts exactly.
Worth mentioning that preact also has better webcomponent support than react and is smaller than lit. If you want a small framework just to make webcomponents with no build step it is a very good option. Compared to lit, lit has first class CSS while preact has a better reactivity system. You're not realistically going to notice perf differences between them, but you will notice preact webcomponents working well compared to react.
And unlike Lit, it isn't maintained by google so you don't have to deal with a new incompatible version of it each year with dropped support for the old one (polymer -> lit-element -> lit -> lit major version numbers). It's basically a more stable version of react and it takes backwards compatibility seriously while still being easy to hire for due to its similarity to react
Don't mind me, just curious where Svelte and Angular fall
That Thumbnail made my Jump 10 Meters
Is different from final result no ?
Angular was doing great in the thumbnail 😂
So glad when you highlighted Nuxt. The nuxt team is doing a great job.
When I first look at elm I was also a bit confused because of my biases based on the languages I coded with before.
But then, it's just pure love
angular should be higher
23:31 I've been using Svelte but I've never seen @const current
I'd rather extract the @const's into reactive variables ($:) or something
Or, even better, simply create a new list item component and put the row in it
I have no doubt that Theo had a rotten time with a particular job that happened to use Ember, and a job can make or break ones opinion of any framework. :(
14:10 i literally just found that out on this angular js project haha, awesome video!
His way of deciding if a framework is good or bad by simply Googling it without really use it deeper and further understanding speaks volumes about how unfair the rating is to an unfamiliar framework.
Its just a tier (opinion) list, he could have ranked them off their color scheme, no need to take it personally.
I use svelte and at this point syntax is just a matter of preference. What I would say though, is that with svelte you don't really worry about writing it the "wrong" way. With react you always have to make sure you are writing proper code otherwise you'll be writing slow code or code that just doesn't do what you think it will. If you understand reach really well then obviously you have less of a problem. But with so much to keep in mind at all times you have less mental capacity to think about the actual problem you want to solve. I don't think svelte is ever going to get close to react but like Theo said, I do love that it pushes other frameworks from so many angles. And I especially love that Rich is not afraid to go places other frameworks can't/won't go.
can i get your only fans h o e?
Yeh agree. This tier list is a bit of a joke tbh. Which is fine as I like a laugh
No hate but you’re off when you say you can’t isolate away code from Vue components in the same way you can in react. I’d argue Vue with the composition API is way more flexible than React in that sense. You infact can have reactivity outside of Composables (Vue hooks) and components if you really wanted to and import it into composables and components as needed
I like react, don’t get me wrong, and I even think it’s probably easier to learn than Vue is. I do think Vue templating is leagues upon leagues better than templating in JSX, that always feels like a much less pleasant experience / less readable (in my opinion)
Exactly!
That's a 40 minutes of my life that I'll never get back
i find it really helpful to learn by drawing on other peoples conclusions and summaries. (because i can skip rabbitholes)
thanks for a really genuine and justified opinion.
Ah yes, flutter is javascript ✅️
So true 😂
dart2js bro
You thought I was Dart there all this time, but it was just js in disguise..
Actually it is transpiled to JS and compiled Webassembly, which can make it faster than every other JS application.
RE: Lit "I don't think any of the other frameworks here have that capability where you've now created an element you can access in the browser". Really makes me think he doesn't know enough about Angular to grade it after that.
Exactly lol. Angularjs had that with directives in version 1.0 in 2013
You dont really need to write single file component in Vue 3. Its composable
This should be higher
as someone who has desperately been seeking a true successor to knockoutjs thank you for this bc I had never heard of solidjs and I think it's the framework I've been looking for
Having started with React, moving to Svelte and Vue felt infinitely better from a developer experience pov. The React ecosystem’s its biggest pain point.
I'm yet to learn react. I have built a hand gesture recognition. Webapp using only vanilla js. Should I rewrite it using a framework(code is clunky) so which one should I learn for 2025, react or solid?
You can reuse any component logic across components in Svelte by making those properties global or by using binding methods available, Svelte is easily the best one in this list, solid is also good but different
10:02 hydration isn't a too big deal? I thought that's why you were very excited about server component, if it's not a big deal, what's the advantage of server component then?
Díky!
I don't understand why people can't get over the Vue and Svelte templating. By that logic, all templating in every programming language is bad, but this is just a bad take. I hate JSX because I hate mixing JS with my markup. I'd rather have attribute based templating instead. It's so much easier to read!
Also, hooks will forever suck!
Would have been nice to see Quasar Framework in this. Might have been a "Never Heard", but it is a real hidden gem.
he talked about it in his video about the developer survey
What did he say? @@SandraWantsCoke
@@axetroll I remember he said it's cr@p. However from my experience you should do your own tests. I've never touched it.
@@SandraWantsCoke I never saw a negative comment about it. I did try sometime ago
My experience with React Native as a Junior dev:
Me: can you please build this app for iOS?
React Native: naw 🖕🗿🖕🗿🖕🗿🖕🗿
I swear Theo put flutter on the tier list just to shit on it
1:05
"jQuery on S tier" trigger "do not recommend this channel"
I needed to use with a Java backend and it was messy and verbose
Rage baiting to get everyone in the Christmas spirit! Theo might be getting coal this year 😂
dude alpine is sooooo convenient. I used to use jquery on sites I was making, when I want to just use a little bit of interactivity.. .now it's always alpine for those use cases... so nice.
You are brilliant! That was so fun. Cheers.
7:30 jQuery had jQuery Templates. Small neat rendering thing. I used it to create an online shop for a major clothing retailer back in 2010-11.
As a gatsby developer I agree 100% with this placement. It’s a terrible development experience and has been a nightmare to use for a large project
Would love to hear your thoughts on Storybook -- yes, please make a video on that! Also, I'm getting started with the RedwoodJS metaframework, would love your thoughts on it as a whole as well. (it uses Storybook, also Prisma/GraphQL and some funky CLI tools to create data "cells"... super curious if you think this is useful or a bad idea.)
[I'm working on a new side project, very experienced backend engineer but first time using JS. I'm building something with a lot of data entry components, complex UI, and strong security needs since it's FinTech, so I'm a little paranoid about getting things right from the start.]
Jquery at S tier is wild. As someone who started as a Flash developer before Jquery was a thing, I just don’t get it. Just because lots of sites use it is like saying most of the web is Wordpress, so Wordpress is S tier.
JQuery was the React of its day, not very good, but still popular because it is better than using nothing.
@@evancombs5159 just create your some utility functions.. then you don't need to include a painful JQuery CDN.. in last 3-4 years. almost no body using it.
I am a backend developer and I use jQuery in my current project. Not in all projects sophisticated frameworks are needed. If you know how to use vanillajs, jQuery and html templates, you have the freedom to create components and to each thing what you want.
@@hugochavez6170 of course you can use Jquery but you can use Handlebars too but why would you? I do frontend work & used Jquery for years I just have no desire to go back to it.
had a wonderful time watching this with your quick run downs on each framework 👍
How did Astro get categorised in F and jQuery S??? And why the hell is Flutter in this list? Since when is Flutter a JavaScript framework? Oh, I completely forgot that he really wanted to give a side kick. But it's good that the F category is highlighted in green. Maybe you have to read the list from bottom to top.
* *makes videos about frameworks* *
* *almost everything is not actually a framework* *
that being out of the way, thanks for making such a cool video :)
Vue3 for SPA, AlpineJS to inject some reactivity into a server side rendered page. Anything else I'd rather work in retail than code in them.
"Angular just follows what others are doing" that's a cute idea.. except the fact that they basically invented the idea of components (called directives at the time), two-way data binding, shadow DOM, and were the first major framework to adopt TS out of the box (which pissed a lot of people off at the time, but eventually the majority of the web dev community ended up following suit). Since then, they have become more of a mature, slow-moving framework (which has massive advantages when building enterprise software), but let's not forget Angular's important role in web dev history
3:16 😆 " for impact, its an A, but dont make me use it." 🤣🤣
>didnt differentiate between angularjs and angular
entire fucking opinion discarded lmao
Just the moment I saw he put Angular in D tier, I lost all the hope for people who're watching this video.
"It is one of the most painful frameworks", were you talking about React Library?
Angular is built for enterprise applications, not a todo app or an alarm timer.
Even if you have 10k pages, you can easily navigate through the project.
This is why every company out there, have problems with their frontend.
This is the reason, you as viewer won't grow.
Your tiny fingers can't write Typescript? You call your React function a component? Aww.
💯
By the time a JS person can finish telling you about all the different frameworks they've used recently, they will have already used another dozen frameworks that they will now need to tell you about.
Finally someone else shits on Ember. My first frontend job used ember, and having come already from Angular.JS 1.x, I could not make heads or tails out of what the fuck was supposed to happen at any given time, the documentation was approximately on par with Apple's (which inspired Ember) documentation (bad or missing), and it relied heavily on some arcane convention over configuration. I believe my first burnout happened when I just spent too many damn days trying to build some relatively simple stuff but getting nowhere.
As a previous Flash and jQuery developer I guess I should try out SolidJS instead of being pissed off at how shitty NextJS is.
Flutter! my favourite JS framework 😍
for web he mean ig
Solid in S tier. Well deserved. I'm using it for everything I do now. I cant' wait for it to hit mainstream.
Not a fan how solidstart is also using "use client"/"use server". I preferred the server$ and generally functionThatRunsOnServerEndsInDollar$. Much better than string statements.
Do you use typescript
AlpineJS at D is such a disservice lol. It's basically the jQuery of reactivity.
Great video, made me laugh out loud multiple times. 😆
Да, Nuxt и Vue - великолепны❤❤❤
I used Ember between 2014-2019. Your review of it was pretty fair. In it's heyday it was an excellent choice:
- Batteries included / convention over configuration (vs the disaster of React at the time... this was before create-react-app)
- Ember Data (although frustrating) provided a comprehensive data story that other frameworks were lacking
- Ember CLI was a breath of fresh air compared with early Webpack
- Ember Addons extended your app functionality and were as easy to install as Rails gems
- Nested layouts/routes akin to Next.js 13 app folder
- IMO the class-based components with the @tracked decorator is cleaner than React's use function mess (although I still love writing React)
I think Ember really lost momentum around 2017 after React and Vue took over the ecosystem. Some things I think they could have done better:
- Official TypeScript support sooner (it took them FOREVER)
- Abandoned Handlebars for JSX
It's funny how I use Vue because of SFC and in the end, they look more like vanilla HTML than when using JSX, which ends up looking like a mess of JS, HTML, and CSS simultaneously.
Nuxt is underated
Giving F for htmx was infuriating. I can't handle it
astro does have a router, import ViewTransitions and put it inside head tag and your app becomes an SPA
HTMX connects you to server state over ajax, alpine gives you client state, both by extending HTML declaratively
I remember people correcting me that react is a library rather than a framework.
The best thing to watch during the Christmas
Great vid can’t say I understand enough about nuxt to see why it went higher than next. I get the file base routing is a bit of a force
Props for putting jquery S tiers right away without questionning, it gets too much hate because not fancy and new enough but it just gets the job done
Mark my words, the next big thing will be built with jQuery, ushering in a new era of simplicity and efficiency over complexity for its own sake
Are you nuts
@@PanosPitsi ( I'm building it )
Everything up high on your list is based on frameworks not able to scale well yet Angular which is the best at that is at D…
You mention you don’t like the design decisions behind Angular but once you build a huge project with it you’ll see why they did it like that