Getting you, from here to done.

The Current

Our Summary of “The State of Code Review 2017” by SmartBear

Dynamo /
Collaborative Code Review

SmartBear—a company that creates monitoring tools—released another 35 page, detailed study of code review and collaboration. You can read the whole study here, but who has the time for that? No need to worry about missing out, we read it for you and pulled out what we feel are the highlights.

The findings are a compilation of responses from 550 developers representing about 30 different industries. Some of these industries are computer software, mobile development, systems integration, aerospace, insurance, industrial manufacturing, and game development. All sizes of organizations are represented from smaller teams with just a few developers to large companies with over 50 software team members.

US DEV TEAMS ARE CENTRALIZED; DECENTRALIZED TEAMS ARE INTERNATIONAL

The survey found that just over 50% of development teams are distributed across multiple locations, and 38% of those are international. Only 15% of development teams are distributed across locations in the US. Surprisingly, with the rise of remote work, many companies are still centralizing their development teams. They may distribute developers, but they keep teams in the same physical location. Team members that aren’t in the same location, in this case, live internationally.

THE BEST WAY TO GET BETTER CODE IS TO DO STRUCTURED CODE REVIEW

Regarding code reviewing in general, the survey found that performing some sort of code review is the number one way for collaborative development teams to improve code quality. The respondents use a number of different types of review including ad-hoc reviews, over-the-shoulder reviews, meeting-based reviews, and some form of a code review tool. Based on survey results, teams using a code review tool are much more likely to review code on a regular basis compared to other methods. Teams using ad-hoc or meeting-based review generally only review code weekly or monthly. In contrast, teams that use some sort of code review tool are much more likely to review code on a daily basis, increasing the odds for improved quality.

WHAT SIZE OF TEAM IS BETTER? IT’S UNCLEAR.

The data is interesting around a team’s perception of their code quality. Larger teams report lower satisfaction with their code quality and less ability to meet deadlines. Smaller teams report a higher satisfaction with code quality and greater ability to hit deadlines. Yet, small teams do far less code review and review code much less often. Larger teams, on the other hand, are much more likely to use a code review tool and collaborate on code daily. We could assume that smaller teams are more flexible and able to move more quickly than larger teams. But it also could be that larger teams, who do more code review with a specific tool and process, are simply more aware of their code quality. Further, it could be that larger teams have more difficulty hitting deadlines because they are creating cleaner and more accurate code. Unfortunately, this data isn’t deep enough to conclude the reason for these intriguing findings.

In conclusion, large and small teams alike in any location or team environment should be reviewing code often and, if possible, consider using a code review tool.

 

We’d love to hear about your experience with code review. What method does your team use to review code? How do you feel about the quality?

Share This Post