I've been a web developer since 2019. I used React.js and React-based frameworks like Gatsby, Next, Remix, Astro, and Hydrogen. I've never been ful...
For further actions, you may consider blocking this person and/or reporting abuse
Definitelya high value post. Finaly something to read. Big respect. Thanks for sharing
Thank you!
Great post Daniel. It's interesting how similar or journeys have been so far - except that I've never been at all tempted by Ruby! I'm firmly in the Remix camp at the moment, and am quite happy with it at the moment, however you have inspired me to look into Elixir and Phoenix in greater detail.
I'm not in total agreement with all your points, I personally don't think that co-location is a bad thing.
You're right though, there's no such thing as the perfect tech stack, it's fun to keep looking though!
I'm also in the Remix camp right now! I work primarily as a front-end developer, and my team uses Remix (Hydrogen to be more precise). And the developer experience from it is miles ahead of Next in my opinion. If Remix had batteries included (for the backend), I would still use it for my personal projects, but for my use cases I decided to move to Elixir + Phoenix
It's good to hear something so positive about Hydrogen. I've got a potential e-commerce freelance job coming up and was considering it
... all the great things for Hydrogen are coming from the 20 people who use it lol.
Thank you for sharing your thoughts!
Deploying Live view requires some additional caution, as restarting the server leads to state reload on the backend. There is built recovery solution for forms but custom components may require additional code. This is not a problem with Rails and Stimulus because state is stored in HTML.
Finally, someone did a lot of hard work to write this. Daniel, I did the same. I started my journey in 2018 with React but slowly by the time 2020 Ruby on Rails took the tide, because of the market demand. Elixir is good, might try it someday.
Fun Fact this website is on Ruby on Rails. 😆
Frameworks are a double edged sword. They're convenient because they provide a lot boilerplate out of the box. However as you've noted, the big downside is that you're at the mercy of what the framework provides. If you stray from it, be prepared for much pain.
There's a place for frameworks. While they can hide some complexities, that can only go so far and is detrimental sometimes when debugging. There's no substitute for fundamental software engineering knowledge and experience.
What about Blazor + .NET 8? I know you said no to C#, but the new version of Blazor blew me away. C# is improving drastically with every new version. The good thing about learning C# is that you can also make games with it, which is my hobby.
To be honest, I was just too lazy to try C#, I never looked too deep into the language
you are misjudging C# my friend :) it's not the old and ugly Windows-related thing it was many years ago. It has become one of the most modern and sexy languages. The syntax is pretty much the same as Typescript. For example, writing a fully type safe hello world API is just 4 lines of code now.
ASP.NET Core apps are now faster than Node.js apps and much much faster than PHP/Rails apps. Like 10x faster or more. They are pushing out new modern features quite often and the team has moved away from the clumsy old .NET platform to a new lightweight and easy to learn platform. It's like Node.js but easier, safer, and faster
I can never get over
MapGetin Microsoft related languages, looks so uglyKeeping up with the JS / React ecosystem is a huge challenge. Building a project and coming back to it after 2 months and nothing works anymore since there’s some new paradigm. Gave up on it entirely too! 😅
Exactly my feeling as well! I can get a Ruby on Rails codebase from 5 years ago and it still works. But If I go back to a 1-month-old JS codebase, nothing works
After trying Next (production level at work), SvelteKit, and NuxtJs.
I would say the best to use is Nuxt for 3 simple reasons :
I’ve heard good things about Nuxt!
Sorry, but based on my experience, your story doesn't end here. Once upon a time, I felt the same way. However, I came across something called "You Don't Know JS," which made me realize that I didn't know JS yet—where the actual problem lies. After exploring ecosystems of other languages, I returned to the JS/TS ecosystem. I believe the exposure I gained in this ecosystem is unparalleled, and I wouldn't find it in any other language. Restarting from scratch isn't something I can afford, and let's be practical—you just can't get rid of JavaScript. It's the most mature option available for working in the browser. Moreover, if you delve into it in-depth, it becomes the most suitable option for many use cases on the backend as well.
I'm not here to advocate for the JS/TS ecosystem, but mark my words, brother, you will come back one day.
Thanks for sharing your experience! But after my burnout on JS, it's extremely unlikely I'll ever go back to it. Especially because with Phoenix Live View I can write all my frontend code in Elixir, zero JS required
Great post. I agree with most of it. I am ironically moving in the opposite direction. (I have used almost every language you can think of in production including haskell)
(Also, I might recommend golang + elm SPA).
I love elixir. I have been working with it full time (backend only) for almost two years. The only downside that I know of with liveview is the lack of ready accessible headless UI component library. I am also hoping either roc or gleam eventually beat elixir, but right now it is a top lang + framework.
I am moving in the opposite direction for job interviews only. If more early stage startups used elixir, I would drop js in a second. (If you are looking to use elixir professionally, that will be an uphill battle currently since there are few job openings and they usually focus on experienced elixir devs)
OK, I actually didn't expect to see a post talking about Elixir coming from a frontend, specially talking about a journey coming from React.
It was, to say the least, interesting to know about your perspective in that journey.
I think more React developers should give Elixir a try! It fits our mental model much better than a traditional OOP framework
I've been working with svelte kit in the past two years and you can get a functioning project with a single "npm create svelte" command.
And the code you have to write to create a new page/api is pretty much the same as in your Ruby's example.
Plus the components use the same language of the server with some small additions to the html side, this fix two language issue you face when not implementing the UI in JS, at some point you have to add some custom script in the page and you find yourself do deal with two languages.
The alternative is to use something like Dart or Reason that doesn't require you to write JS because they are compiled to JS code, but they both are a downgrade from the simplicity and the flexibility that comes with JS/TS.
That's good to know!
At this point I'm so burned out from JS, that I just want to run away from it to be honest
Great post and this is exactly why I quit JS web development years ago. It really is a job for code monkeys that somehow pays well these days. There are so many other, a lot more interesting things going on in the world besides new JS frameworks.
Agreed 100%
Being a JS developer, I have felt this burnout a few times. These frameworks are simply too much. Only a few frameworks and libraries remain critically unchanged over time, and provide backward compatibility.
Sometimes I have this urge to simply move to another stack, but get hit by the team I work with being a JS team.
Sometimes I wonder if I should just go with PHP or Python or even Go, or Rust. With this new WASM thing popping up.
Tech is getting crazy, from WASM to LLM to WebWorker, Kubernetes and whatnot. Its simply too much to take in.
Yep, for sure. There is a lot of decision fatigue in the tech industry. Personally, I decided that I do not want to work professionally with JS long-term, so I'm planning to fully migrate to an Elixir job in the future. I'm burned out on React, Next, and all of these Frankenstein frameworks, it's just too much for me
Great article
For the problems that you faced while using Rails, you might want to look at Ruby-lsp extension. It adds Intellisense and other things to the editor that Rails have been lacking until now.
It's still a bit rought but much better than not using anything.
I tried it, but to me, it wasn't enough to have a good developer experience. Elixir LS fits my needs better currently
This is an excellent read - thank you.
As a side note: I wish to see more discussion about engineering - rather than framework x,y,z.
If you have 5 years + solid experience, it's so easy to move between React, Vue, Django, Rails, and whatever tool. What's difficult is engineering complex systems.
Interesting post,
I'm about to build website for a client, initially I was looking on Next.js and Vercel but now I'm eyeing on Astro.build and vanilla javascript + web components
Not only is a lot simpler but support markdown with no additional npm install, it also ships no javascript.
Payload CMS is also good alternative, but I wanted something more simple.
If I was building an web application I probably would have used something in the vain of rails
Very good approach! Astro for simple and static websites, and Rails for more complex backend stuff 🔥
Sounds like what you want is RedwoodJS. Heavily inspired by Ruby on Rails. Convention over configurations, generators, testing, component library, all configured and ready for ya with a single command. It's seriously underrated.
Right now I would rather use Ruby on Rails. I hate that it’s mandatory to use GraphQL with Redwood. Just a lot of overengineering in general
Really good Content as a react developer I can feel the same, it seems very breezy in start but you will find something is not alright, I wanted to switch to qwik, what is your thoughts?
I completely gave up on the JS ecosystem at this point, to be honest. Every year we keep saying the same thing "This is the perfect framework! We can finally stop the framework wars!" and then this "perfect framework" gets replaced by another one in the next 6 months. It's a never-ending cycle
Giving up on the entire JavaScript ecosystem sounds a bit dramatic tbh. It's not so bad if you ignore most of the hypes... React has been a pretty solid choice for me for the past 7 years and while I do keep a close eye on things like solidjs, I can also see that it is not mature enough that I would consider using it in anything except a personal hobby project. Just don't be scared to choose the mainstream for enterprise projects and keep the rebel attitude for personal projects or to further the ecosystem as a whole.
Thanks, nice article. Will try that sht Phoenix, which I read about for the last couple of years. Rails has autocomplete in VSCode and there is RubyMine by JetBrains
Very interesting post, thank you! You did not mention deployment with Elixir, how was you experience on that. Did you find a Vercel-like solution or was it more complex?
That’s my biggest struggle right now. It is a lot harder, you need to learn about docker, elixir releases etc. However I’m using Fly, which make this process a lot easier
I deploy using GH actions, no need for docker, just a simple elixir release that's copied to the server.
Great article, I think many developers will recognise that feeling of framework-fatigue that you describe so precisely.
Thanks!
Interesting read, thanks for sharing!
I'm glad you like it!
Great post and very interesting journey! 👏🙌
I would add lit and web components to the soup.
Thanks for sharing!
Definitely going to check elixir!
Have you tried AdonisJs? It's like Laravel but for Node.
For sure! I used Adonis for 2 projects, and I loved it! It's the closest thing we have to a Ruby on Rails for JavaScript. The only reason I stopped using it is because the community around it (unfortunately) is very small. So for example, to release a new version the core team takes multiple years.
But if you don't want to leave the JS ecosystem, it's a very good choice!
Read the entire article thinking "wow, that is great, I wonder if the author have any more content" then I realized IT'S FUCKING BERG FROM YOUTUBE LMAO
YOOOO IT'S MEEEEE, NOW IN ENGLISH 🤘
Thanks for the awesome comment!
Great read!
Thank you ... very nice observations. Convinced me to take up Elixir and Phoenix.
You lost out by ditching Laravel for small syntax issues :) Livewire 3 is also pretty neat.
It's just my personal preference, I could change my mind in the future
it's not just small syntax issues. PHP is a terribly designed language. It's verbose, far from being type safe, missing most of the type definition syntaxes that other modern languages have had for ages, and it's slow and inconsistent.
This is an old article and many issues have been fixed in version 7, but many are still there
eev.ee/blog/2012/04/09/php-a-fract...
Even basic features are missing. You have to use ugly workarounds like PHPDoc blocks for them. Type inference only supports very basic types, nothing like Typescript or any other modern language.
Laravel is awesome, it's a shame it is using PHP
Cool man
Any reason you didn't consider some go/rust based systems? Could you not find any that were mature enough, maybe?
Good question! I did consider trying golang, however, what annoyed me a little bit was the "fragmented" ecosystem, similar to JavaScript. There isn't a Ruby on Rails for it yet, just a bunch of libraries that I would have to "glue" together.
If someone builds a framework that includes an ORM, routing, authentication, testing, modern frontend and all the "batteries included" philosophy, then I'll definitely give it a chance!
Oh, and after being "bit" by the Functional Programming virus, I can no longer see the OOP languages in the same way
Nice post! I've also messed with Elixir and it is really cool!
Very good post, thanks!
As a developer with 15 years in the game, I can not more agree.
It's a mess. No MVC, bad developer experience without strict rules.
Could you share the reason to move from Next and Remix? What issues did you face exactly?
I actually had 0 reasons to do this switch, I only did it because of my day job 😅
Great post, exceptional work, thank you. (except the Java , C# comments 😊I am old school so these languages have earned my respect )
I may give a try to Elixir.
Hey, nice post, but for me react is best option.if you will want to extend from web app to mobile, then you will not be able share buisness logic with React Native
Phoenix has LiveView for the web and LiveView Native for mobile 👀
loved the writing style 🥹
This as an amazing read. I enjoy all of it. Great opinions on the madness going on in the JavaScript ecosystem.
What a fantastic post! I also gave up on the JavaScript ecosystem as my main stack for the same reasons you described. I've spent a ton of money on front-end courses, which I don't regret because I've gained valuable skills.
However, I recently discovered my preferred technology stack: C# / .NET and Dart / Flutter. I realize it might seem unusual to opt for both web backend and mobile dev technologies, but I believe this versatility will increase my chances of landing a job in either field.
Looking forward to reading more of your articles!
Finally I found an interesting article to read.
Definitely I have similar feel about Js. I get the burnout just from my short 3 years experience of its ecosystem. I am a Laravel user which I consider more complete and simpler than similar frameworks in Js ecosystem. I don't care if it is an "Old boring" tech stack, the only matter to me is its simplicity.
Now, I started learning Elixir/Phoenix since last week. I believe this is best the alternative of Laravel that I was looking for. Maybe doing some project while learning all about both the language and the framework. Let's see.
Another ecosystem I want to dive into is Dart. I worked once on a Flutter project, and it was a satisfying experience to work with. I love Dart as a language: simple, familiar and easy to learn. It also has complete tool chain and included by default, no need to learn every other tool chain just like in Js ecosystem.
That is a great post. I've been on the same journey until I settled with SvelteKit. The challenges with using something like Elixir, in my opinion, are two:
Dude, I watched your video on Youtube talking about this because I had kind of a similar experience. I've always loved Python and Django, but never had the opportunity to use it professionally, I ended up working with C# and .NET (by the way, you should give it a try).
I have been working with SAP and ABAP for 10 years but I hated it from day one and I've been trying to leave this shit for a long time now, so I decided to study something new and started studying Javascript. I felt overwhelmed by the amount of different libraries and frameworks that exist to solve the exact same problem and, like you said in your Youtube video, having options is good, but too many options can get in your way. So I kind of gave up on it and now I'm slowly learning Rails and Django, but I'm also very interested in Elixir.
I agree with most of your points, but let me tell you, there’s no perfect solution for everything and there's always a feature that sometimes I think "man, I'd really love if Django had that awesome Rails feature".
Brother, I am a beginner developer and currently learning javascript as well as trying to understand the node ecosystem. After reading this blog you have somehow put me into a dilemma.
Shall I continue this journey or do you have some better suggestions.