Rust devs twice as productive as C++. Who would have thought?
https://www.theregister.com/2024/03/31/rust_google_c/
Shopping Cart Returner Shirt $21.68 |
Rust devs twice as productive as C++. Who would have thought?
https://www.theregister.com/2024/03/31/rust_google_c/
Shopping Cart Returner Shirt $21.68 |
>posting el reg on IQfy
ask me how i know you've been inside for the last 15 years straight
>the register is based
>cope and seethe
> troonyshit
> the register
> "google reports", as they sit on the rust board and a major sponsor
> no conflict of interest here, everyone!
lmao. the absolute state of this forced meme.
How do they measure productivity? Number of commits? Number of lines changed? What about people who are always changing formatting? Also probably the Rust people are on Adderall and the C++ people aren't.
>How do they measure productivity?
The amount of work that gets done.
Great question. By any metric the number will be skewed since the majority of Rust projects are nothing but reproduced work rewriting things in Rust, or connecting c/c++ applications and libraries with rust. Wonder what mozilla thinks since they were the first to try anything major with rust and basically gave up
>How do they measure productivity?
They don't, they just lie about it.
Whoa, hold the phone! They figured out how to measure programmer productivity?? We've been trying to do that for 50+ years! This is an incredible breakthrough!
>Whoa, hold the phone! They figured out how to measure programmer productivity?? We've been trying to do that for 50+ years! This is an incredible breakthrough!
It's the same way you measure productivity with anything else.
They gave multiple teams the exact same project only different languages to use.
its a benchmark of the teams asmuch as a benchmark of the languages
dishonest rustrannies being dishonest. color me surprized...
link to the study?
it seems like one dude at a conference saying "decrease by more than 2x effort to build and maintain"
>dude trust me bro
trannies will always be better because they don't have lives. I support them though.
>C++ devs create new stuff
>Rustoids only do rewrites
Mystery solved
>add unsafe word to c++
>lock memory allocations/deallocations and direct pointer dereferences behind the unsafe block
it's that simple. we don't need a whole new language, just force them to use the safe constructs c++ already had for years
>I don't know how to program
why do you tards think you have the right to opine? Rust's safety extends beyond pointer deref.
what's the right way to in-place transform an array through a thread pool in rust without making a ton of copies?
https://docs.rs/rayon/latest/rayon/iter/index.html
Specifically with Rayon you can do
use rayon::prelude::*;
let mut data: Vec<i32> = /* ... */;
data.par_iter_mut().for_each(|n| *n = *n / 2 + 1);
Or if you want to create a new array instead of mutating in-place you can map and collect_into_vec() and it'll still pre-allocate a single array.
You can also iterate over contiguous windows and chunks and so on.
Good luck with that
WOAH THERE, BUDDY! IS THAT AN UNSAFE LINE OF CODE ? *wags finger* TSK TSK TSK... THAT'S A BIG YIKES, YANNOW! A BREACH OF THE CODE, DARE I SAY! *finger vibrates like a pole flipped suddenly up-right* THE CODE OF CONDUCT! *takes a quick sip of from his recyclabe straw connected to his backpack liquid śoy container* MMH! HR IS GONNA LOVE THIS!
OH MY SØYENCE! I'M SO SØYRRY MR MANAGER! I THINK I NEED A MENTAL HEALTH DAY!
that's cuz c++ people are boomers and they're all brain dead regardless.
worse than boomer, they're deranged. anyone following Boost dev should have learned this by now. I've been a knower for over a decade.
> Experience in programming is bad because -- it just is okay?
>Rust devs are twice as productive, says Schlomoberg. Developers over 30 years old typically do not improve much but continue to demand higher wages. This does not apply to rust devs.
>Developers over 30 years old typically do not improve much
this is true though. stop being a piece of shit and expand your skillset.
boomers literally deserve age discrimination tbh.
t. 30 year old boomer.
>rust devs do not make it past 30
good. we need Logan's run irl, fr fr.
your IQ starts to falls off a cliff with age. It will always be an uphill battle for boomers to compete, even if they tried
>Developers over 30 years old typically do not improve much but continue to demand higher wages. This does not apply to rust devs.
Rust devs keep improving at any age.
>devs using newer language have more commits because they're building from scratch
>devs using older language have fewer commits because they're modifying existing libraries
Wow!
They analyzed how long it took to do rewrites (of codebases that IIRC were themselves already rewrites, they do a lot of this at Google). No commit counting
They claim they did this a/b testing, but they didn't actually post any numbers or methodology.
The only data they actually showed is survey data, which unsurprisingly showed le heckin old==bad, new==good
This, also I don't think Rusts development speed is any greater than C++.
Just look at WebGPU, Googles dawn(C++) vs Mozilla's wgpu(rust). Where googles implementation is much more stable and usable, while wgpu is lagging behind.
Id put it down to employee number and skill more than anything, but from what I heard, Rust is a massive pain in the arse to fix stuff thats broke or needs changing because it’s very “stern” in how it’s written. A bit like how functional programming language basically need a full re-write just to change a part.
Sounds like the crystallisation problem of OOP where legacy code gets design patterns so every builds ontop of it rather than fixes underlying issues, so it’s most likely going to be slow as shit in prod as those cases arise
>Mozilla's wgpu
wgpu has nothing to do with Mozilla. wgpu also has nothing to do with WebGPU, it's only somewhat inspired by it. WebGPU is a Web API, wgpu is a Rust library.
>wgpu has nothing to do with Mozilla
>wgpu also has nothing to do with WebGPU
That seems a bit too strong. wgpu says in its intro blurb that it's
>the core of the WebGPU integration in Firefox and deno
and most of the currently active top contributors are Mozilla employees.
>nuh rust devsm nngh...
says fricking who?
ah, right
discount fricking willy wonka
lmfao
next time make it cote d'or or some fricking milka at least
you wont be credible either
but at least it wont come from literal who's in the chocolate industry...
don't give a frick still using C++
don't care since neither one is usable.
I was once a wagie at a domino's pizza. one day the manager decided to show my a trick to clean the dough lids. interrupt the dishwasher after the wash cycle, insert the next tray of lids, resume at the rinse cycle.
behold Productivity(tm)
Aren’t they just rewriting shit? No wonder they are more productive, the Rust team is just data entry while the C++ team are actually developing software
Maybe it's attracting all the massive autists and causing a brain drain from the C/C++ teams?
Citation? Replication? Furthermore, why should people not at google care that googlers, but presumably not anyone else, are twice as producitve using rust over c++.
>presumably not anyone else
Why would you presume this?
Why is ever rust library v.0.1.0? I thought this shit was ready for production. Even wrappers like openssl isn't even 1.0.0. I'm not using this alpha shitware.
I wonder if that's because there's no legacy code in Rust yet. They don't have mountains of old shit to deal shit.
Troonlang
A higher level, better designed language and modern language which incorporated lots of decent PL research (type classes, pattern matching, DSLs out of methods) and enabling actual non-bullshit innovations (lifetime checking, at most one exclusive mutable reference, etc) is more productive than a crap designed by an unqualified narcissist butthole.
Who would have thought lmao
https://schiptsov.github.io/programming/understanding_rust.html
How do they manage to magically figure that out?
>Most of development time is spent on debugging
>Use language that takes longer to write but decreases debugging time in return
Makes sense to me.
Why can't rustroons just learn to memory management?
The people who wrote the C software running on my computer can't seem to either
I haven't seen a segfault since the 90's, seems like a skill issue to me.
What are your systems?
Wtf are you talking about?
Why can't you educate yourself before making a stupid post?
Rust is not a managed language, it doesn't make memory management any easier. It's actually harder to do since allocators have to take lifetimes into account.
>said the hobbyist that's never worked on any serious large scale collaborative projects
Memory bugs aren't the only thing rust fixes. It also fixes data structure serialization so I don't have to tediously write a whole parser any time I need to read some json. Cniles don't value their time if they don't consider that an issue.
Coz they got more work to do?
That's not really surprising. I don't think we can tell much about quality of Google code, but Rust is probably being used for brand new projects and these usually move faster than maintaining legacy. But I can totally see Rust letting you be more productive in general as well. It's a comfy language.
I wonder why rustroon always treat their own language like it's some kind of E.coli
Does it remind them of their axe wound?
Is their goal to infect everything and dies out?
not surprising if you actually write code
Rust devs are younger and probably more passionate. That's really probably it. Nothing about the language itself I'd think.
But on the other hand, experience makes you better at avoiding dead ends. You have a greater repertoire of what works and what libraries are available.
Google’s C++ style guide forbids major language features such as exceptions, so their experience isn’t really representative.
exceptions considered harmful
Back in the day, there was a command called "GOTO." It was considered really bad and a major cause of spaghetti code. Someone joked that the reverse of GOTO would be even worse. The reverse is COMEFROM. Exceptions are COMEFROM.
C++ is full of moronic shit like this with implicit conversions and breaking RVO. It is one of the most error prone languages you could possibly be writing in
?t=142