Chapter 10 - Key Coding Standards for Automotive Software Development

82% of those surveyed are using at least one coding standard, a slight decrease of 4% compared to last year. Even a 4% decrease, though, is alarming. Respondents who cited their top overall concern as Testing are using a coding standard the most (87%), followed by Security (86%) and Quality (84%). Interestingly, those with Safety as their top concern had the lowest usage of a coding standard (79%), even below Productivity (80%).

As the overall safety concerns this year revolved around difficulty fulfilling every ISO 26262 requirement, those who do use a coding standard are better positioned to fulfill essential functional safety standards.

The use of a coding standard is important for code quality, which helps to ensure that the software is safe, secure, and compliant.

Many of those we surveyed are using multiple coding standards. MISRA use increased by 8%, returning to 2024 numbers, and continues to lead as the coding standard of choice for the majority of automotive professionals. With the recent release of MISRA C:2025, this increase was expected. It is important to use a static analysis tool that enforces the most recent MISRA guidelines.

Following MISRA, AUTOSAR C++14 (36%) replaces C++ Core Guidelines as the second-most popular coding standard this year. CERT usage increased by 5%. This reflects the greater focus on safety and security in 2026.

Of those surveyed, the majority of respondents are required to track code quality metrics (despite a decrease of 18% since last year). 

Most respondents who need to track code quality metrics are required to track test coverage (59%), which was added as an answer this year, over the previously leading cyclomatic complexity (54%). The metrics most likely to be tracked are part of the HIS metric set and are particularly applicable to the C programming language. 

Cyclomatic complexity is useful, as it gives an indication of the number of test cases required—but it should be considered as a comparative measure and not to be below a definitive value. The number of lines of code is particularly difficult to evaluate for more modern languages and requires a definition of “line of code” to be useful.