Quo Vadis, Code Review? Exploring the Future of Code Review Journal: IEEE Software
Michael Dorner, Andreas Bauer, Darja Šmite, Lukas Thode, Daniel Mendez, Ricardo Britto, Stephan Lukasczyk, Ehsan Zabardast, Michael Kormann · 2025
Read Quo Vadis, Code Review? Exploring the Future of Code Review Journal: IEEE Software by Michael Dorner, Andreas Bauer, Darja Šmite, Lukas Thode, Daniel Mendez, Ricardo Britto, Stephan Lukasczyk, Ehsan Zabardast, Michael Kormann (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-06879.txt.
Introduction
Michael Dorner and co-authors ask software developers where code review is heading, in a 2025 study. Code review — colleagues reading each change before it merges — is a bedrock habit of team programming. Automation now writes, checks, and even reviews code, which puts the human ritual in flux.
The team surveyed one hundred developers at five software-driven firms. Questions covered how much time review takes today, what gets reviewed, and how both may shift over five years. Open answers invited hopes and worries in the developers' own words. Tensions surfaced between speed, trust, and learning on the job. The results give tool builders and team leads a grounded view of what practitioners expect.
Catalog record
| CREATOR | Michael Dorner, Andreas Bauer, Darja Šmite, Lukas Thode, Daniel Mendez, Ricardo Britto, Stephan Lukasczyk, Ehsan Zabardast, Michael Kormann |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 23 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-06879
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Quo Vadis, Code Review? Exploring the Future of Code Review Journal: IEEE Software Michael Dorner, Andreas Bauer, Darja Šmite, Lukas Thode, Daniel Mendez, Ricardo Britto, Stephan Lukasczyk, Ehsan Zabardast, Michael Kormann Abstract Code review has long been a core practice in collaborative software engineering. In this research, we explore how practitioners reflect on code review today and what changes they anticipate in the near future. We then discuss the potential long-term risks of these anticipated changes for the evolution of code review and its role in collaborative software engineering. \jtitle IEEE Software \sptitle Quo Vadis, Code Review? “Nothing is constant except change.” —Heraclitus \chapteri During the early days of software engineering, code was often designed and implemented by individual developers working in isolation. As the software systems became more and more complex, software engineering became a collaborative effort requiring many developers with diverse skill sets and specializations, organized in different teams, and later even distributed around the globe. To maintain and enhance a collective understanding of the code base and the changes to it, developers need to discuss a code change before it gets merged into the code base. Over the years, these discussions have taken various forms ranging from the formal and heavy-weight code inspections in the 1980s [ 9 ] to synchronous and steady exchanges between two developers in pair programming [ 13 ] . Nowadays, the asynchronous and less formal discussions around code changes are known as code reviews , a practice that has seen widespread adoption in collaborative software engineering [ 3 , 8 ] . However, change is constant in software engineering—and code review is no exception. Advances in large language models (LLMs) and new regulatory frameworks, such as the EU Cyber Resilience Act, are beginning to influence how code reviews are conducted. In this article, we take a glimpse into the future of code review. To that end, we begin by analyzing how practitioners reflect on code review today and what they expect to change, based on a survey of 92 professional software developers from four large software-driven companies, each representing a distinct organizational and software engineering context. Building on this analysis, we discuss potential long-term risks for the future role of code review as a core practice in collaborative software engineering. Table 1: Organizational and Software Engineering Contexts and Participant Numbers by Company Attribute SAP Ericsson A Bank JetBrains Industry Domain Enterprise Software Telecommunications Banking and Finance Developer Tools Business Model Software products (B2B SaaS, ERP) Infrastructure provider (network equipment, telecom software) Financial services (retail and investment banking) Commercial software vendor (IDEs and tooling) Primary Software Focus Customer-facing enterprise platforms Embedded and network control software Internal financial transaction and risk management systems Developer productivity and language tooling Number of Employees >100,000 >100,000 >25,000 >2,200 Number of participants 32 24 25 11 Sidebar: Study Design Sampling To gain a diverse perspective on the future of code review grounded in real-world practice, we employed quota sampling , a non-random, two-stage strategy. First, we purposively selected four relevant companies representing varied organizational and software engineering contexts. Then, we aimed to recruit approximately 25 professional developers from each company, closing recruitment after eight weeks or once the quota was reached. Participants Our sample includes 92 developers from (1) one of the world’s largest enterprise software providers (SAP), (2) a globally leading provider of telecommunications infrastructure (Ericsson), (3) one of Europe’s largest banks, which has requested anonymity, and (4) a mid-sized vendor of professional developer tools, including widely used IDEs (JetBrains). The number of participants per company varied slightly. See Table 1 for an overview of the participating organizations, their contexts, and participant numbers. Though not statistically representative and with varying participation per company, the sample captures diverse industry perspectives on code review. Questions To capture developers’ current code review practices and their expectations for how the practice may evolve over the next five years, we asked six questions: How many hours on average do you currently review code per week? (numerical response) Do you expect to spend more, less, or the same amount of time on code review in five years (compared to today)? (Multiple choice) What software artifacts do you review today? (Multiple selection) What software artifacts do you anticipate reviewing in five years? (Multiple selection) What major changes do you anticipate in code review in five years? (Open-ended) Based on those changes, what implications for code review do you see? (Open-ended) The survey covered six types of software artifacts covered in code review: Production code, Test code, Parameter and configuration files, Documentation, and GUI-based test code (end-to-end testing). Respondents could also add further artifact types or choose to select none. Data collection The survey was conducted end of 2024 and beginning of 2025 via an online questionnaire, shared through internal communication channels at the respective companies. Participation was both voluntary and anonymous, allowing respondents to share their perspectives freely. To further protect confidentiality, we deliberately excluded demographic questions such as age, gender, or company affiliation, and report the results on sample level rather than by individual company. Participants were informed about the study’s purpose and assured that their participation was voluntary, and their responses would remain anonymous. Data Analysis Quantitative and multiple-choice responses were analyzed using descriptive statistics. Open-ended responses were examined using thematic analysis [ 5 ] to uncover common patterns and emerging trends. Data availability All anonymized data and analysis scripts are publicly available at github.com/michaeldorner/quo-vadis-code-review Code Review Five Years From Now What do developers actually do in code review today and how do they think that will change? In this section, we report the results of our survey, capturing how developers currently experience code review today and how they expect it to change in the coming years. 1 1 1 See sidebar for more details on study design. The analysis is structured around three central questions from our survey: the time they currently spend on code review and how they expect that to change; the types of artifacts they review today versus those they anticipate reviewing in the future; and who they expect to be involved in the code review process going forward. Not Less Time Currently, respondents report spending a median of approximately three hours per week on code reviews. To contextualize this finding, we compare it against two external data sources: first, a 2013 survey by [ 4 ] involving 416 Microsoft developers; and second, the 2019 Stack Overflow Developer Survey, which includes responses from 49\,790 developers 2 2 2 https://survey.stackoverflow.co/2019 . 0 1 2 3 4 5 6 7 8 9 10 0 0.25 0.5 0.75 1 Hours per week reviewing code Our data Stackoverflow (2019) Microsoft (2013) Figure 1: Cumulative distribution of hours spent on code review today ( \leq 10 hours). In the absence of raw data, we approximated the cumulative distribution from the reported ordinal-scale results at Microsoft [ 4 ] . Figure 1 shows the cumulative distributions of weekly review hours across all three data sources. The distribution in our sample is consistently lower compared to both the 2013 Microsoft study by [ 4 ] and the 2019 Stack Overflow Developer Survey, which show strikingly similar distributions with a median of around four hours spent