News for hackers. Stuff that matters.
Re:A brief history of Rust at Facebook by bsder on Thursday April 29, 2021 @08:10PM (#26988421)
> I'm curious what people think about Rust as a long-term investment from an employment/job perspective. Is it going to pay the bills for the next 10 years?

The answer is almost certainly: no.

From a career standpoint, you would be better off learning C/C++ in much more detail before investing time in Rust.

If you want the "canary in the coal mine", watch what the game developers use for the game engines (they mostly use C++ right now). When Rust starts becoming common among game developers, it's got enough momentum for a job ("career" is probably a bit much).

However, even when Rust hits critical mass, it's unlikely to be a "career choice" programming language--you will have to know Rust AND <something else more important>. The world has changed, and "One Language To Rule Them All" is no longer a thing.

C, Java, Python, C++, Javascript, C# are going to hold the top 5 of programming languages for the conceivable future. Not matter how good Rust is it is not going to displace C and C++ on any reasonable timescale--inertia is a thing. Swift and Visual Basic have their niche and nothing is going to blast them out of it.

After that, it's pretty much a free-for-all in terms of language popularity, but everything is tiny at that point.

Re:A brief history of Rust at Facebook by pcwalton on Thursday April 29, 2021 @08:37PM (#26988611)
> If you want the "canary in the coal mine", watch what the game developers use for the game engines (they mostly use C++ right now). When Rust starts becoming common among game developers, it's got enough momentum for a job ("career" is probably a bit much).

I don't see why the entertainment industry would be any more important to pay attention to than any other tech-adjacent industry, such as finance, health care, transportation, etc. If anything, the short-term focus on the next title means that the games industry tends to be relatively late adopters of experimental core tech: witness how more quickly tech companies adopted C++11 as compared to game studios, for example. (This is not to criticize entertainment's conservatism with technology choices: it's a tough business and they are understandably risk-averse.)

Re:A brief history of Rust at Facebook by bsder on Thursday April 29, 2021 @09:26PM (#26988905)
> I don't see why the entertainment industry would be any more important to pay attention to than any other tech-adjacent industry, such as finance, health care, transportation, etc. If anything, the short-term focus on the next title means that the games industry tends to be relatively late adopters of experimental core tech

Um, precisely?

Game programming is hardware-bound and short-term focused. Consequently, if Rust is generally superior to C/C++, it will spread very quickly. This would be especially true in games that have a much higher networking component (where the memory and security guarantees start to matter more than just the raw engine performance).

Rust is a systems programming language. It is meant to extract efficiency/performance out of hardware. If that isn't your raison d'être, then using a systems programming language of any flavor is counterproductive.

So, the problem is that Rust doesn't really compete in "general programming"--it competes mostly in "systems programming". And it has to displace competitors that are mostly "good enough". And it can be done piecemeal so that you don't have to port entire codebases all at once.

None of that is a recipe for fast change. And THAT'S PERFECTLY OKAY.

Now, all that having been said, the wildcard in all of this is WASM. Rust support for WASM has been stellar when everybody else's has been hot garbage. WASM could be the thing that lets Rust break out of just the "systems programming" niche. That's a big hypothetical, though.

Re:A brief history of Rust at Facebook by follower on Friday April 30, 2021 @05:18AM (#26991321)
> the wildcard in all of this is WASM.

On a related note, Embark Studios who I mentioned above, just announced they've joined the Bytecode Alliance, an WebAssembly-focused non-profit: https://twitter.com/repi/status/1387393289847025669

The announcement mentioned the use of WASM to help build "...a future of software & creative game worlds that are not limited walled gardens where only we / the devs can add new capabilities or features, want to build an open ecosystem of interconnected software & game components that anyone can use & create with".

Essentially, Rust+WASM as a future for player-driven modding/customisation.

This is an area I've been exploring in my "WebAssembly Calling Card" project which enables people to create cross-platform personalized dynamic image avatars implemented in any language which can compile to WASM: https://wacc.rancidbacon.com

I've integrated WACC with the Godot game engine, Rust GUI viewers and the browser--they all use the same avatar ".wasm" file (with examples currently written in C & Rust).

Re:A brief history of Rust at Facebook by bsder on Friday April 30, 2021 @11:22PM (#27002748)
Neat!

I'm happy to see all of this, but, again, the question was about career in/for 10 years. And Rust just isn't going to be that.

Rust is, for example, far behind in terms of market penetration where Python was at a comparable age. This is unsurprising. First, Rust is nowhere near as broad spectrum as Python--so it's domain of applicability is much more constrained. Second, Rust is coming into a full ecosystem that Python didn't have to contend with.

Python was basically the only respectable person at the scripting language job interview in the late 1990s--Tcl and Perl both decided to start huffing glue in the middle of the DotBomb era while Sun tried to jam Java everywhere it didn't belong and left a lot of greenfield areas to Python. Rust doesn't have the "advantage" of that level of incompetence in its competitors/alternatives.

And, while I'm fully aware that Python and Rust aren't really in the same space, I simply chose Python because it's one of the only popular languages that is relatively "new" (if you can call something from somewhere in 1989-1994 "new") and didn't come upon its popularity as a mandated parasite attached to a much bigger host (looking at you: Swift/C#/Javascript).

Re:A brief history of Rust at Facebook by mistrial9 on Thursday April 29, 2021 @09:04PM (#26988781)
game programming is performance heavy, while other tech-adjacent industry may be requirements-driven. Also, C++ has existed for what? 30 years? t
Re:A brief history of Rust at Facebook by pcwalton on Thursday April 29, 2021 @09:07PM (#26988803)
Finance is performance critical too, just to name one. Games are just one small part of the high performance systems space.
Re:A brief history of Rust at Facebook by bsder on Thursday April 29, 2021 @09:29PM (#26988926)
Finance is performance heavy in very, very, very narrow points--and NO programming language cuts it in that spot. Those firms use custom silicon.

The vast majority of "programming" in finance is finding something to model that is relevant. And that's Python, R, Julia, etc.--all the exploratory tools. Nobody in that space has even heard of Rust.

Re:A brief history of Rust at Facebook by smabie on Sunday May 02, 2021 @06:15PM (#27019560)
There's a lot of C++ in financial/trading: custom silicon is only used for the most rudimentary/simple/rote part of the system (think like packet parsing or templating your messages you send to the exchange).

Many people are very interested in Rust right now for HFT: C++ is nothing but pain and causes tons of problems.

Re:A brief history of Rust at Facebook by follower on Friday April 30, 2021 @04:16AM (#26990954)
> If you want the "canary in the coal mine", watch what the game developers use for the game engines

On this front two relevant data points are:

* Embark Studios: https://medium.com/embarkstudios/inside-rust-at-embark-b82c0...

* Ready at Dawn Founder/CTO: https://twitter.com/AndreaPessino/status/1021532074153394176

Also, there is a specific Rust GameDev Working Group: https://www.rust-lang.org/governance/wgs/gamedev

Re:A brief history of Rust at Facebook by coldtea on Friday April 30, 2021 @06:41AM (#26991791)
How relevant are these data points though? Embark Studios e.g. appears to be some small company in Sweden (?), doing 3-4 different things (aside from games) to get by.

You can find all kinds of languages in game shops, including custom Lisps and the like, so finding Rust too is a given.

But I think the spirit of the parent was to check what game studios start adopting "en masse", not whether this or that studio has adopted a language.

Re:A brief history of Rust at Facebook by follower on Friday April 30, 2021 @08:33AM (#26992570)
> How relevant are these data points though?

Depends on who you are I guess. :)

> appears to be some small company in Sweden

According to[0] they're currently around 200 employees and the founding team includes former CEO of DICE (+ex-EA) & other experienced industry people.

They're funded in the large tens of millions range: https://www.pcgamesinsider.biz/news/69461/nexon-increases-st...

So, basically, they're a bunch of experienced game industry veterans funded with a bunch of money & a clean slate to choose their technologies--and have made Rust as one of their prime choices. (Although I note they're using Unreal for their 1st game, so not sure how much of the Unreal-related dev they'll be doing in Rust.)

> doing 3-4 different things (aside from games) to get by.

Curious what gave you that impression?

[0] https://medium.com/embarkstudios/our-continued-journey-89dad...

Re:A brief history of Rust at Facebook by bsder on Friday April 30, 2021 @11:08PM (#27002646)
Take a look at the Embark jobs.

C++, C++, C++, C++, C++, C++, Rust (for security), (Go, JavaScript, Python and Rust).

The Rust stuff is all confined to the back end where security matters--no engine stuff.

Ready At Dawn got bought, so it would be interesting to know if that still holds as they are owned by Facebook who is Rust-positive.