State of Clojure 2020 Results

Clojure developers will be particularly interested in the version and tooling related questions in the survey.

One new question we asked this year was about the primary developer operating system to give us better guidance about tool-related work. As expected, MacOS was the leader (55%), followed by Linux (35%) and Windows (9%):

For primary development tool, there were only minor updates this year. Emacs dropped slightly to 43%, IntelliJ/Cursive rose slightly to 32%, and VS Code with Calva had the biggest increase to 10%.

Clojure users have a wealth of fine development environments, each suited to different communities and tastes, and we’re glad to see them all getting better day by day.

For many years, the survey has included a question about "build tools", but this idea of a single monolithic tool has become increasingly less reflective of how people are managing Clojure projects, where they may use multiple tools for different parts of their process, particularly in mixed Clojure/ClojureScript projects. In response to this, we tailored this question more tightly to dependency management and made it multi-select this year:

Looking at previous years, we continue to see strong (but slightly reduced) use of Leiningen, and a steady increase in use of clj/deps.edn. For ClojureScript work, shadow-cljs has made big strides over the last couple years, with big support from Clojurists Together.

In the greater world of Java and the JVM, Java has migrated to a new release strategy where releases come out every spring and fall, and every 3rd release is a "long term support" (LTS) release - 8, 11, and 14. Java 14 is just in the final ramp down stages now and is available only in early access builds right now. Java 9 introduced a major change with the module system and in all JVM communities this has caused a significant user base to remain on Java 8. Clojure reflects this as well (although probably shows more shift to Java 11 than other language communities):

Releases like Java 9, 10, 12, and soon 13 are effectively dead when the next release comes out and we would recommend sticking primarily to the LTS releases and maybe the latest release, if it’s not an LTS release.

One aspect of Java 11 that is underappreciated is significant work to make Java work better in containers like Docker. If you are deploying in containerized environments with Java 8, you should really be looking closely at the changes in Java 11 and considering an upgrade.

Clojure itself has been using Java 8 as the baseline JVM for a couple years and will continue to do so (while also supporting newer versions of Java). When running Clojure, we recommend Java 8 or 11 right now.

Since last year, we’ve seen strong uptake of Clojure 1.10.0 and 1.10.1. The latter was a maintenance release this year with error handling improvements building on the changes in Clojure 1.10.0 and mitigations for some Java performance regressions in their service releases after Java 8u201. Use of Clojure 1.8 and earlier continues to dwindle:

In addition to the prior dependency management question, we also added a new question on how respondents are starting their apps in production. Based on feedback, it’s likely the wording and answer choices will need some fine-tuning next year, but there is some interesting feedback in the results:

The majority of users are using launchers like Leiningen or clj to start their production apps, more so than by building jars or uberjars and launching them directly with Java. We do see a small group also experimenting with Graal native images (particularly common with smaller scripting apps).

from Hacker News https://ift.tt/39L62wG