What kind of projects are a best fit for Rust, for those who are experienced with it? Any recommendations for projects to make Rust stick?
Thanks!
Find one of the most simple utility (probably a CLI) you have and rewrite it.
I have a lot, and find useful to turn some python stuff so is easier to ship (surely Go work here...)
That’s basically how I learned Go. The language is very simple coming from practically any imperative language, once I stopped hating on the directory structure imposed on me (Go module helped) and went through the short language tour, code basically flowed. Even got a nice productivity boost over Python when dealing with concurrency.
Rust is totally different. Want to deal with a few strings? Fight with borrow checker for an hour. Want some functionality outside the practically nonexistent standard library? Frantically try to find a library, turning up either none or a bunch without clear winner. Realizing that examples are outdated. Then fight with borrow checker for another hour. After a while I ask myself, why the hell should I subject myself to this torture when I don’t need to squeeze out every ounce of performance? Switch to Go or Python, be done in an hour.
esr had a similar experience: http://esr.ibiblio.org/?p=7294
> In practice, I found Rust painful to the point of unusability. The learning curve was far worse than I expected; it took me those four days of struggling with inadequate documentation to write 67 lines of wrapper code for the server.
> Even things that should be dirt-simple, like string concatenation, are unreasonably difficult. The language demands a huge amount of fussy, obscure ritual before you can get anything done.
> The contrast with Go is extreme. By four days in of exploring Go I had mastered most of the language, had a working program and tests, and was adding features to taste.
(Not saying Go doesn’t have pain points. Specifically for writing command line tools, the argument parsing story is garbage.)
Oh and I did end up learning Rust to a reasonable degree, one way or another.
Complaining about Rust's documentation is a bit weird because even back in 2015 there was quite a bit of good documentation, let alone 2017 and the language encourages good crate API documentation.
I also find it ironic that the author of The Cathedral and The Bazaar would say:
> The Rust community appears to have elected to use the decentralized nature of their crate system (which undeniably has some very nice technical properties) to execute a swarm attack on the design space around the language. Which is wonderful in theory but seems to be failing in practice.
There are certainly deficiencies in the language and the ecosystem, but to me they stem from immaturity, only 5 years out from 1.0, and not from any fundamental major mistake.
It always seemed to me that what happened here is what happens often with a lot of experienced people that try Rust and bounce, where they approach the endeavour as "I know how to program, this will be easy, I just have to map my already existing mental framework to the new language" and hit walls early on because Rust is different enough but they didn't anticipate it. Functional languages have similar issues, but the more "esoteric" syntax helps flip a switch in people's minds to learning mode. This happened to me back in May of 2015 and it wasn't until I picked it up a few months later that it clicked.
If it needs to run in the browser, or React native, I use TypeScript.
Else if it needs to run on the JVM I use Kotlin.
Else if performance doesn't matter and the project is small, I use Python.
Else if I can use a garbage collector, I use Go. Because why manually manage memory with the borrow checker in Rust if I can let the runtime handle it?
Else Rust. Because I don't want to mess with C++ anymore.
Sarcasm aside, choosing languages this way makes great sense. Though it can be hard if you don’t one that’s ideally suited to your problem set.
I don't know Rust, I've just read the book before and played around with it a little. So using C or C++ would have been a more natural choice - but I like having an opportunity to add another tool to my belt. Compared to learning Go it's been a slog.
Do you personally start writing Rust from the get-go or do you first write something in some scripting language and then transform it to rust once the basics work?
Eventually.
I struggle, a lot, for the first 3 months of Rust. Not each minute doing rust. I code in 2-3 other langs, but the slow was pretty obvious.
I think have worked in so many other langs (+12) actually hurt me most. I try to start with rust like all the rest: Go ahead coding, figure stuff along.
I stop doing like a idiot when I start looking at it as my FIRST lang ever: read a book, be slowly, PAY attention to what are the concepts "owned", "move", etc..
Also:
I try to learn with building a relational lang, and develop the routine to build very small rust projects of a single file. I have done dozens now, and I start to click to me, and can do a lot of stuff far faster and near python-speed (when simple and normal).
It really depends on your attitude toward type errors. I find rust great for exploration because when I change something, the compiler tells me how to fix everything up.
I also have a lot of bias here, of course.
[0] https://gist.github.com/siraben/5209bfc7165c45aec18eeae852ec...
I really would just like a $job working in it. But there's a bit of a chicken and egg problem there.
Got back a page with nothing but this:
We're Sorry...
Commure is unavailable in your location.
We’re required to comply with US sanctions laws that restrict the use of our site
in certain jurisdictions. Because of this, our services are not available in Crimea -
Region of Ukraine, Cuba, Iran, Iraq, Syria, North Korea and other US embargoed
countries.
(FWIW, I'm southern Ontario. Currently @ Google for last 8.5 years. I have had health care dev experience before, and yes, it was terrible. No React experience, but 20 years general software dev experience at basically every part of the stack. Currently doing embedded though.)
Yes, that works.
Anything that you're using C++ or Go for.
If you proceed anyway, be sure candidates crossed over from C++ because, underneath the syntax fluff, they will need to understand resource management and performance cost of architectural choices. Finally, if Rust doesn't work out quite as well as hoped, they can pivot to mature tooling without loss of momentum.