Survey results post #196
Survey results post #196
Conversation
| author: Jonathan Turner | ||
| --- | ||
|
|
||
| It's that time of the year, where we take a good look at how things are going by asking the community at large -- both Rust users and non-users -- how things are going. And wow, did you respond! |
There was a problem hiding this comment.
Repetition of "how things are going" reads a bit awkwardly
|  | ||
| *(hover for more info)* | ||
|
|
||
| People are working ever-larger amounts of Rust, with medium- and large-scale lines of code totals both nearly doubling since last year as a percentage of the whole, now making up 16% of respondents (up from last year's 8.9%). This shows a growing interest in using Rust in ever-larger projects, and a growing need for tools to support this growth. |
|
|
||
|  | ||
|
|
||
| Despite the rising amount of code developers are working with, we're seeing a small downtick in both daily and weekly users. Daily users have fallen from 19% to 17.5%, and weekly users have fallen from 48.8% to 43.3%. This could be a natural transition in this stage of our growth, as a broader range of developers begin using Rust. |
There was a problem hiding this comment.
I would make clear that the absolute number of daily users is growing, even though the percentage is falling
|  | ||
|
|
||
| Developers have largely opted to move to nightly or a recent stable (with some on beta), showing that developers are eager to upgrade and do so quickly. This simplifies the support structure a bit from last year, where developers were on a wider range of versions. | ||
| Unfortunately, despite our efforts with procedural macros and helping move crates like Serde to stable, we still have work to do to promote people moving away from the nightly Rust compiler. This year shows an increase in nightly users, now at 1,852 votes it represents 51.6% of respondents using nightly, up from 48.8% of last year. |
There was a problem hiding this comment.
This seems not necessarily negative? It's only bad if they don't want to be on nightly, it seems possible that people's trade off is moving towards nightly because it is more stable than last year? Or there are more enticing features.
There was a problem hiding this comment.
In general, requiring people to use nightly is seen as a negative by many folks and a sign of Rust not having the level of maturity that some corporate/conservative users look for. A lot of the push this year for getting things like serde on stable was to help address this.
|
|
||
|  | ||
|
|
||
| Linux still features prominently as one of the main platforms Rust developers choose. Of note, we also saw a rise in the use of Windows as a developer platform at 1,130 of the 3,588 total respondents, putting it at **31.5% of respondents**, up from 27.6% of last year. |
There was a problem hiding this comment.
A 3% rise in the number of Windows users doesn't seem very note-worthy to me - how does it compare to movement on the other platforms? Is there a reason to call out Windows in particular?
There was a problem hiding this comment.
It's 4%, and it's the only change since last year, so I called it out :)
|
|
||
|  | ||
|
|
||
| Likewise, **full-time** Rust commercial users saw medium- and large-scale projects grow to taking a larger part of the pie, with projects over 100,000 lines of code making up almost 18% of the all full-time commercial respondents, and a large shift in the 10,000-100,000 lines range from 39.7% up to **52.6%**. |
There was a problem hiding this comment.
Hmm, I assume the 100,000+ people are all Mozilla employees (unless someone is doing a giant Rust project we don't know about)? Not sure how much attention we should bring to that.
There was a problem hiding this comment.
From a quick look, the vast majority aren't Mozilla employees (as best as we can tell). We know of some pretty large external Rust projects, but it does make me wonder if we shouldn't reach out to those people who say it's okay to and see what they're doing...
There was a problem hiding this comment.
Huh, interesting, we should def. reach out if we can. Perhaps the total project size is > 100,000 lines rather than > 100,000 lines of Rust?
|  | ||
| *(hover for more info)* | ||
|
|
||
| An important piece of the Rust community is to be one that is welcoming to new users, whether they are current users or potential users in the future. We're pleased to see that over 3/4th of all respondents said they feel welcome in the Rust community, with 23.6% not sure. |
|  | ||
| *(hover for more info)* | ||
|
|
||
| The first surprise we had here was how long people gave Rust a try before they stopped. Our initial hunch was that people would give us using Rust in the first day, or possibly the first week, if it didn't suit them or their project. Instead, what we see is that people tried Rust for a much longer time on average than that. |
|
|
||
| Finally, at the end of the survey we we provided a free-form area to talk about where Rust could improve. Before we get to the themes we saw, we wanted to give a big "thank you!" to everyone who posted thoughtful comments. There are many, many good ideas, which we will be making available to the respective sub-teams for future planning. With that, let's look at the themes that were important this year: | ||
|
|
||
| * 17% of responses underscored the need for **better ergonomics** in the language. People had many suggestions about how to improve Rust for day-to-day use, to allow for easier prototyping, to work with async programming more easily, and to be more flexible with more data structure types. Just as before, the need for a much easier and smoother experience with borrow checker and how to work with lifetimes was a popular request. |
|
|
||
| # Conclusion | ||
|
|
||
| We're blown away by the response this year. Not only is a much larger number of responses than we had last year, but we're also seeing a growing diversity in what people are using Rust for. Thank you so much for your thoughtful replies. We look forward to using your feedback, your suggestions, and your experience to help us plan for next year. |
There was a problem hiding this comment.
grammar - "Not only is a much larger number of responses"
|
Spell check stuff is mostly false positives or jargon, we should add that stuff but it shouldn't block this! Thanks so much @jonathandturner ; everyone has been waiting on this post, and you've done a great job! ❤️ |
r? @rust-lang/core
Here's the 2017 survey results post