1. 16
  1. 8

    What astonishes me most is that Rust again is the most loved language and the percentage of people who do love it goes up every year*…

    • 2021: 86.98% (First place!)
    • 2020: 86.1% (First place!)
    • 2019: 83.5% (First place!)
    • 2018: 78.9% (First place!)
    • 2017: 73.1% (First place!)
    • 2016: 79.1% (First place!)
    • 2015: 73.8% (Third place, losing to Swift and C++)

    *from 2017 on


    1. 19

      The loved/dreaded ratio is from actual users of each language. My theories why it’s high:

      • Rust is still niche enough that people who use Rust have to seek it, so Rust users use it because they want to. Compare this with Java or PHP that dominate the job market, and many programmers may use them only because they can’t find another job.

      • Rust is young enough and evolving, so it probably hasn’t grown many dreaded “Enterprise” legacy codebases.

      • Rust isn’t really competing with all programming languages. It focuses on a niche owned by C and C++. For programmers whose only choices were either a language that peaked in 1989 or a language known as “an octopus made by nailing extra legs onto a dog”, Rust just showing up is already a godsend.

      As for why it’s increasing:

      • Users who don’t find Rust useful, leave. Rust isn’t unreasonable when you need a high level of safety and maximize performance, but if you’re fine with a bit less control or strictness, there are plenty of other languages to choose from.

      • As Rust gets more popular and known for what it actually does, it’s driven less by hype, and more by actual needs, so it’s picked by users who need Rust, rather than users who are just curious and want to play with something new.

      • Rust has released many quality-of-life improvements over the years, which were very well received by its users (better borrow checker, simplified modules, async/await, proper arrays). There has been a steady improvement in quality of error messages, tooling, compilation speed, and overall maturity of the ecosystem.

      1. 7

        The “no legacy codebases” point is huge. If you dig into why people hate Java and PHP, a lot of it really translates to, “I have had to work with some absolutely awful Java and PHP code from the stone ages.” Neither language is perfect, of course, but people aren’t viewing either one of them from the kind of unexplored clean slate perspective of a person who has just chosen to use Rust to build a brand-new application.

        Put an army of poorly-trained junior engineers with no domain expertise and an impossible deadline on a big Rust project, and the people who come after them will probably have a lot less love for Rust.

      2. 10

        I have never used most languages out there, including various that seem really neat. How can I really judge these languages? It takes me at least a few months of actual use before I feel comfterable really having an opinion. Only 7% of people report they actually use Rust. Although a number of people have probably played around with it for a bit, the gap is still huge.

        There’s been loads of “zomg, this is amazing, we should apply it to everything!” hypes in the past, such as threading, OOP, dynamic languages, micro{kernels,services}, The Cloud (on-going), The Blockchain (on-going), etc. None of these are bad things – most of them are actually really neat things – but they also didn’t really live up to the hype and it turned out things were a tad more complex. I have no opinion on Rust one way or the other – and clearly it’s a good programming language on account of many useful things being written with it – but is it so good that 86% of people think it’s just brilliant? Colour me skeptical; when was the last time 86% of programmers agreed on anything?

        I always found these kind of figures rather meaningless and not all that informative, and I dislike how some people use them as a cudgel to “prove” something. I stopped doing the survey years ago.