The impossible case of pitching rust in a web dev shop
flakm.github.io
The impossible case of pitching rust in a web dev shop
1–10 of 17 posts
Re: The impossible case of pitching rust in a web dev shop
#2I keep looking for something really special for writing web applications using rust. Is there something on the horizon? It seems like rust on the server side is an easy decision but I'm unclear if there is a way rust could replace nodejs as the isomorphic application language on both server and browser.
Clojure and ClojureScript was such an interesting way to bring lisp into the browser and seemed really revolutionary. And without really limitations on what you could build. Where is the rust version of that?
Re: The impossible case of pitching rust in a web dev shop
#3I keep looking for something really special for writing web applications using rust. Is there something on the horizon? It seems like rust on the server side is an easy decision but I'm unclear if there is a way rust could replace nodejs as the isomorphic application language on both server and browser. Clojure and ClojureScript was such an interesting way to bring lisp into the browser and seemed really revolutionar…
Yew: https://yew.rs/
Re: The impossible case of pitching rust in a web dev shop
#4I would say "the irrelevant case.."
Where Rust excels is in safety when using concurrency. Most web apps and APIs have a single thread used per request. I've also used async/futures that work fine without ownership modelling.
As for performance switching from Ruby/Python to Java/C#/Go will pay off not so much to then go to Rust unless it's a particularly high throughput or latency zensitive app in which case you'd already know it matters.
Re: The impossible case of pitching rust in a web dev shop
#5Rust in a web dev shop would be a huge premature optimzation smell for me.
Most public web apps are expected to be up pretty much all of the time, which today, means cloud hosting and horizontally scalable architecture. In such a setup, I don't see how squeezing every last drop of performance is worth it. When I can just change a number and get more instances, choosing the language with higher cognitive load is going reduce any cost savings due to lower resource usage.
Doing something as mundane as writing text to a socket just doesn't need to perform all that well. Developer effort costs more than the CPU/memory.
Re: The impossible case of pitching rust in a web dev shop
#6I keep looking for something really special for writing web applications using rust. Is there something on the horizon? It seems like rust on the server side is an easy decision but I'm unclear if there is a way rust could replace nodejs as the isomorphic application language on both server and browser. Clojure and ClojureScript was such an interesting way to bring lisp into the browser and seemed really revolutionar…
Yew: https://yew.rs/
Thanks, this was very informative.
I found some other alternatives, like Perseus. Do you have thoughts on it?
https://framesurge.sh/perseus/en-US/comparisons
I am interested more in it because it mentions a lack of VirtualDOM, like SolidJS (so I assume it is more like Svelte than React).
Re: The impossible case of pitching rust in a web dev shop
#7The bummer is that the main reasons I would want my company to adopt Rust are 1) the toolchain and 2) the type system (we're on Python, which has the worst story imaginable for both of these). But most of the pain in learning and using Rust comes from its performance characteristics, which we mostly don't need
Such is life
Re: The impossible case of pitching rust in a web dev shop
#8The bummer is that the main reasons I would want my company to adopt Rust are 1) the toolchain and 2) the type system (we're on Python, which has the worst story imaginable for both of these). But most of the pain in learning and using Rust comes from its performance characteristics, which we mostly don't need Such is life
In that case I think Go is a better language for your shop.
It does not have the performance of rust, but on the other hand it is almost as easy to learn as python.
Re: The impossible case of pitching rust in a web dev shop
#9The bummer is that the main reasons I would want my company to adopt Rust are 1) the toolchain and 2) the type system (we're on Python, which has the worst story imaginable for both of these). But most of the pain in learning and using Rust comes from its performance characteristics, which we mostly don't need Such is life
In that case I think Go is a better language for your shop. It does not have the performance of rust, but on the other hand it is almost as easy to learn as python.
It has the tooling story figured out, but its type system leaves a lot to be desired
Re: The impossible case of pitching rust in a web dev shop
#10Earlier quoted context omitted.
In that case I think Go is a better language for your shop. It does not have the performance of rust, but on the other hand it is almost as easy to learn as python.
It has the tooling story figured out, but its type system leaves a lot to be desired
A lot? I mean, it's not rust but it's not that bad.