8 ms·
Rust 2017 Survey Results
- ewillbefull 9y agoThe next survey should ask which nightly-only features are keeping them on nightly. :)
- Manishearth 9y agoFrom asking around (totally unscientific, but I've been asking this question for a while now), it seems like most folks develop for stable but use nightly for incremental compilation, RLS, and/or clippy. The first two are on the path to stable in the next few releases, and the last one is starting to go in that direction too, but will take longer.
- gcp 9y agoclippy for me, yes.
- ekidd 9y agoYes, this is exactly how we do it: Our build systems all run stable, but developers use nightly for RLS and clippy. We did have one developer who was excited about using Rocket (which he said was much better than the other Rust web frameworks he had tried), but we decided that the grief of building production code on unstable wasn't worth it. (We're also holding off on directly using futures in any major way until the developer ergonomics and especially error messages get better.)
- JoshTriplett 9y agoAmong other things, embedded support; if you want to compile out the standard library, no_std effectively requires nightly right now. Many embedded developers are stuck on nightly for that reason. I've seen some proposed solutions in progress for that, at the moment, such as standardizing panic_fmt (needed to write no_std applications).
- dbcurtis 9y agoInteresting. I haven't tried Rust yet, but it is on my "to learn what all the fuss is about" list. The main reason I'm interested is for embedded.
- JoshTriplett 9y agoIt works very well for embedded; the only problem is that to build a complete OS-style image, you tend to need nightly, either for no_std or to build your own std/core for the target platform.
- bsder 9y agoDo the Rust embedded developers all hang on an IRC room or something?
- steveklabnik 9y agoThere's #rust-osdev (which has been pretty quiet lately) and #rust-embedded at least; I think Tock has their own channel too?
- pc2g4d 9y agoFor me it's usually little things like `TryFrom`. Bits of API that aren't stabilized.
- david-cako 9y agoAnd the one after on what percentage of software developers feel a disturbing lack of politics in programming language development. I love rust, but the community's politics is very odd to me.
- kbenson 9y agoPolitics as in national politics, or politics as in bureaucracy and people working the system for benefit (e.g. company politics)?
- david-cako 9y agoSocial politics. Just odd socially-charged community management and documentation. I get a strong impression that the lead contributors (perhaps it's a mozilla cultural affect?) are very interested in pushing their equality agenda. Not that I don't subscribe to it in theory, it's just very tacky to me.
- gcp 9y agoGot any examples where you think this shines through from the docs? Never noticed it. (I could easily imagine it in the community management so I'll take your word for that, but I've only really interacted with the IRC channels where this wasn't an issue)
- Ruud-v-A 9y agoNot from the docs, but a pretty clear community example is that tickets for RustFest are more expensive for people who don’t belong to an underrepresented group. Edit: replaced “minority” with “underrepresented group”, as that is the language used on the website [1]. [1]: http://blog.rustfest.eu/this-week-in-rustfest-5-tickets-diversity http://blog.rustfest.eu/this-week-in-rustfest-5-tickets-dive...
- kbenson 9y ago> Not from the docs, but a pretty clear community example is that tickets for RustFest are more expensive for people who don’t belong to a minority. Having just looked this up because I wasn't aware of it, I wouldn't classify that statement as true. They allow/ask for donations to help people from diverse locations attend. https://ti.to/asquera-event-ug/rustfest-zurich/ https://ti.to/asquera-event-ug/rustfest-zurich/ NOTE: THIS IS NOT A CONFERENCE TICKET. Support us bringing the opportunity to attend RustFest to more people around the globe who otherwise couldn't afford it. Anything you offer here will go into our diversity pool with the sole purpose of helping people from underrepresented groups attend RustFest. That doesn't mean minorities necessarily, and even if it did, it doesn't mean everyone in that group gets a cheaper ticket. From the first sentence though, it seems obvious they are trying to get diverse nationalities, not necessarily diverse ethnicity from the hosting country.
- erickt 9y agoCore/Community team that helped design the survey here. We actually did ask about this, but unfortunately it's quite time consuming to process through all the free text entry like this one. Some of the popular things people want from nightly are Clippy, impl traits, RLS, SIMD, and compiler plugins, which isn't surprising. It'll take some time to dig through the rest of the comments to see if there is anything unexpected. I'm hoping to do follow on posts as we go, but no promises :)
- pc2g4d 9y agoFinding most frequent words/phrases will tell you a lot. There's also topic modeling if you have enough text.
- Perceptes 9y agoThe requirement of a nightly compiler for many of my projects has been my biggest problem with Rust for both surveys so far. I recently made an issue for my largest Rust project, Ruma, that enumerates all the things we're waiting on from the nightly compiler, as well as other library/ecosystem maturity issues. Just in case anyone is curious what some specific problems are for a real project: https://github.com/ruma/ruma/issues/189 https://github.com/ruma/ruma/issues/189
- frankmcsherry 9y agoThat Rust binaries and Rust shared libraries use different allocators on stable, making FFI .. awkward.
- JoshTriplett 9y ago> Among editors, vim remains king, though we see healthy growth in VSCode adoption at 34.1% (up from last year’s 3.8%). This growth no doubt has been helped by VSCode being one of the first platforms to get support for the Rust Language Server. This suggests that it would help to have better getting-started guides for vim, and for using it with RLS in particular. > 10% responded that tools aren’t use mature enough. I've seen more than one person mention that the fact that RLS doesn't work on stable Rust is a sign of immaturity. Also, in the context of people who had stopped using Rust: > The rest of users mentioned [...] were turned away by Rust’s syntax, [...] or had a bad interaction with the Rust community. Is anyone systematically looking over each of the responses that said either of these, in detail, in order to address them as well as possible?
- steveklabnik 9y ago> This suggests that it would help to have better getting-started guides for vim, and for using it with RLS in particular. Yeah I'm not sure that there's classic Vim support, langserver.org has two things that are both for NeoVim. > Is anyone systematically looking over each of the responses Someone (jntrnr) did read every single response, that's part of why this post took so long to produce. I just asked him about it, basically, he's going to be giving feedback to each of the teams about the stuff that was in their area, and they can act on it that way; so that'd be the language and community teams in these two cases.
- digitalzombie 9y ago> This suggests that it would help to have better getting-started guides for vim, and for using it with RLS in particular. Oh man yeah I'm using vim to play around with Rust, would like to see a tutorial in term of usage of the plugins. Like a tutorial with vim and the guess number example would be nice while using the vim plugins. Idea whatever and VScode is meh. VScode is on the cloud iirc meeeeh.
- gcp 9y agoVScode is on the cloud iirc meeeeh It's based on Electron/Chromium and uses JS for extensions, but it's a local program in every other way. It's also fairly snappy, maybe much more so than the usual Electron app. It does have the usual broken font rendering on Windows, as does everything that is derived from Chrome. But that's no an issue on Linux or macOS.
- squiguy7 9y agoWhen they ask if an update to a stable compiler has broken their code, what does this mean exactly? I had thought that a goal of Rust was to always maintain backwards compatibility? I know how hard this is to pull off and I am not trying to say they have done a poor job; I genuinely want to know what others consider a breaking change.
- steveklabnik 9y agoIn a language that's as strongly typed as Rust, any change can be breaking. So for example, let's say that you had code that looks like this: use std::cell::*; struct MoveCell; We then add a new type to std::cell, also named MoveCell. Your code will now fail to compile, as the two structures conflict. This kind of breakage can still happen, even though we care a lot about stability.
- gcp 9y agousing namespace always backfires.
- squiguy7 9y agoAh that makes sense. I suppose you can't predict naming in everyone's source for their projects. Thanks for the answer!
- steveklabnik 9y agoNo worries. There is also other things as well, that's just one of the easiest ones to explain. For example, in 1.20 last week, we fixed a bug with include! in documentation tests behaving strangely; that's technically a change in behavior, and therefore, a breaking change. Two crates were the only users, apparently, and so they just wrote a patch and fixed it. Still technically breakage. The guiding principle is "it should not be painful to update the compiler", and I think we do as good a job as we can making sure that that's true.
- nickm12 9y ago
- y7 9y agoI wonder if the results for "I don't use Rust because..." would change if the answer "It would slow development down too much", or something along the same lines, would be there. At least that's my personal reason. I like Rust and I've played around with it, but it's just not the right tool for the job if you need Python-style rapid development.
- Can_Not 9y agoCould they make something like nim for rust?
- dom96 9y agoManaging memory manually doesn't sound like it would promote rapid development. So you may as well just use Nim :)
- steveklabnik 9y agoYou don't really "manage memory manually" in Rust either, generally.
- erickt 9y agoGood idea. I noted that for next year's survey [1]. One of our big initiatives this year is working on our ergonomics [2]. I'm not sure if we can ever get as approachable as Python, but we've been working hard on trying to make the language easier. Results are starting to come out over the next couple releases. Things will hopefully be a bit better if/when you try the language. Please let us know if you need any help! [1]: https://github.com/rust-community/team/issues/160#issuecomment-327282002 https://github.com/rust-community/team/issues/160#issuecomme... [2]: https://blog.rust-lang.org/2017/03/02/lang-ergonomics.html https://blog.rust-lang.org/2017/03/02/lang-ergonomics.html
- rsynnott 9y agoNice to see only 7.5% have had stuff broken by an upgrade. Last time I messed around with it, it was still changing rapidly, which is part of what put me off. May give it another go.
- steveklabnik 9y agoWas that before or after 1.0? We do time-based releases, every six weeks. New stuff is still being worked on all the time, like most languages.
- rsynnott 9y agoBefore 1.0 :) Just installed 1.20 now, giving it another try. IntelliJ support helps.
- steveklabnik 9y agoAh yeah, that makes perfect sense. :) IntelliJ's plugin just became officially supported by the company, so let's hope it gets even better from here!
- danjoc 9y agoOnly 75% of respondents feel welcome in the community. That seems bad, considering those who felt unwelcome probably left already and didn't participate in the survey. I would have expected closer to 100%.
- beefsack 9y ago23.6% of the rest of the respondents chose "not sure", I wonder what amount of these people just didn't interact with the community at all and chose that response. Only 1.3% specifically responded with not feeling welcome, which seems quite low to me.
- divs1210 9y agoWell, rust+emacs is pretty easy to setup and provides syntax-highlighting, indentation, documentation, jump-to-definition, etc. out of the box.