Article
April 4, 2024 · 7 min read timeNitor conducts an annual developer survey to determine which languages, frameworks, databases, tools and other technologies our software developers and architects favor. In this post, I present an overview of the main results and highlight interesting changes from last year. I also nominate my three picks for technologies of the year.
This was the fourth year we arranged the survey using the same basic methodology. We asked Nitoreans which technologies they had used during last year for at least ten days. The respondents could answer whether they like, dislike or are neutral towards the technologies they used.
Additionally, the respondents were asked if they are interested in using a given technology in the future. There was a pre-curated list of technologies divided into 15 categories, but Nitoreans were also able to write in missing options. The developer survey got 80 responses, which means about half of Nitor developers and architects responded.
Languages
The programming and query language category is relatively stable this year, with no changes in the top ten. Java saw a 7 percentage point (pp) drop in usage, although alternative JVM-based languages like Kotlin, Clojure and Scala are still far behind. The largest growth of the category goes to Scala, with a 10 pp increase in use from 2023. TypeScript usage is now very close to JavaScript, as almost all projects benefit from static type checking.
Rust and WebAssembly are the two most interesting technologies in this year's survey. Both saw a very high interest already last year, but this has yet to translate to significant use. It might be hard to find projects where the performance benefit of these languages outweighs other factors. Interest in GraphQL has, however, has dropped sharply since two years ago, when GraphQL was the most interesting technology. Kotlin is the most liked language among users, with ClojureScript close behind.
Chart of the most used languages. The solid part of the bar displays the share of respondents who have used the language. It is divided into three parts for users who like, dislike, or are neutral regarding the technology. On top of the solid part, the share of interested users is shown with diagonal stripes. Finally, the label above each bar is the percentage point change in usage since last year. New items in this year's survey are marked with ✨.
Backend
This category includes backend frameworks and JavaScript runtimes. The most notable change is the 13 pp increase in usage for the Express framework for Node.js. We also have a high interest in Deno and Bun, which are alternative JavaScript runtimes, although their usage is still almost non-existent. It will be interesting to see if this will eat into the popularity of Node.js in the coming years.
Frontend
In the frontend category, we see a significant 20 pp increase in React usage, while Angular usage shrunk by 7 pp. The other main alternatives to React, Vue.js and Svelte, remain stable at much lower usage, although Svelte is highly interesting. It looks like React will remain the dominant frontend framework for years. Next.js, the most popular React meta-framework, saw usage grow by 12 pp.
Vite, a build tool, also saw a considerable 19 pp growth, bypassing the now deprecated Create React App. AngularJS, the old version of Angular, remains the most disliked technology in the survey by a large margin, although usage is not shrinking as quickly as developers would hope. Reagent, which provides React support in ClojureScript, is the most liked in the category, with Vite not far behind.
Testing
Jest is still the most used unit testing framework for JavaScript and saw usage grow by 11 pp, although the Vite-powered Vitest is catching on with a large 19 pp growth from almost no usage the previous year. Among end-to-end testing frameworks, Playwright has passed Cypress in usage. Playwright also has the highest interest in the category. Selenium, Robot Framework, and Puppeteer, three other E2E testing frameworks, are highly disliked.
Data
The data category is also relatively stable, although there are a few larger shifts. The most notable is the 12 pp growth of Amazon DynamoDB usage. This makes DynamoDB the most popular NoSQL database by far, although PostgreSQL is still the most used database and has also seen a significant 10 pp growth. Apache Kafka has the highest interest in the category.
Cloud platforms
Although the usage of almost all cloud platforms grew in this year’s survey, AWS grew the most by 15 pp, and it continues to be in the top spot. Azure has the highest interest in the category, although it is not as liked as other options. Vercel, a frontend-focused platform by the creators of Next.js, continues to grow and is highly liked. Cloudflare is notable as the most liked technology in the survey, with 100% of users liking it.
Continuous integration
The main result in the CI category is that GitHub Actions usage continues to grow, this year up by 12 pp, putting it further ahead of the relatively disliked Jenkins as the most popular technology in the category. The use of Azure Pipelines is also growing.
Deployment and infrastructure
This is one of the most stable categories since last year, so there is little to say. Terraform and AWS CDK both remain relatively popular infrastructure-as-code tools. The largest changes are drops in usage of Kubernetes and Ansible.
AI tools
AI tools are a new category this year. The survey specified that only tools used for work should be included in the responses.
Although 2023 will likely be remembered as a year of generative AI breakthroughs, it is still surprising that ChatGPT is now used by 64% of respondents. Last year, it wasn't even included in the survey. GitHub Copilot has also seen a big 26 pp increase in usage since last year, when it was included in the other category. Large language models have now become established tools for increasing developer productivity, although the field is changing rapidly.
Collaboration
Collaboration is another new category this year. The usage rates might be less interesting as several of the tools are used internally at Nitor. Interestingly, Slack is highly liked, while Teams is one of the most disliked technologies in the entire survey. Confluence and Jira are both also quite disliked. The usage of Miro grew by 22 pp, becoming the whiteboard of choice in a world where physical whiteboards are less practical than before.
Technologies of the year
Last year, my three picks for technologies of the year were Go, Next.js and GitHub. All three have continued strong this year. Based on the survey data, here are my three subjective picks for this year.
My first pick is ChatGPT, the generative AI chatbot by OpenAI. Unless you live under a rock, you may have already heard quite a lot about ChatGPT. Still, the meteoric rise of the tool from zero to 64% use within a year makes it an obvious pick. There are risks with using AI to generate code and answer questions. Still, it can be a considerable productivity boost in the hands of an experienced developer, who is able to notice hallucinations and refine responses. There is a lot of controversy and competition in the generative AI space, so the situation might look different in a year, but 2023 was definitely the year of ChatGPT and the GPT-4 large language model.
My second pick is Vite, the frontend build tool created by Evan You, who also created Vue.js. Vite utilizes native ES modules in the browser, which makes it very fast as a development server, even on large codebases. It supports multiple frameworks, including React, Vue.js and Svelte. It provides a plugin system that has allowed, e.g., the Remix framework to be reimplemented as a Vite plugin. Vite saw a large 20 pp increase in use this year, and the Vitest unit testing framework, built on top of Vite, also saw a 19 pp growth.
Finally, my third pick is the Slack communication platform. Used by most Nitoreans every day, it is easy to take the tool for granted, but a good communication tool is critical for hybrid and remote work. Slack is very well-liked, especially compared to alternatives like Teams. Slack can be extended through numerous apps and integrations to enhance the development process. Slack has also launched features to improve focus and reduce distractions, which is appreciated by developers looking for productivity and peace of mind.
Do you agree or disagree? We’d love to see your input as a fellow Nitorean next year!