Table of Contents for

Your Code as a Crime Scene

Discover Your Team’s Modus Operandi

Remember the geographical offender-profiling techniques you learned back in ​Learn Geographical Profiling of Crimes​? One of the challenges with profiling is linking a series of crimes to the same offender. Sometimes there’s DNA evidence or witnesses. When there’s not, the police have to rely on the offender’s modus operandi.

A modus operandi is like a criminal signature. For example, the gentleman bandit you read about in ​Meet the Innocent Robber​, was characterized by his polite manners and concern for his victims.

Software teams have their unique modus operandi, too. If you manage to uncover it, it will help you understand how the team works. It will not be perfect and precise information, but it can guide your discussions and decisions by opening new perspectives. Here’s one trick for that.

Use Commit Messages as a Discussion Basis

Some years ago, I worked on a project that was running late. On the surface, everything looked fine. We were four teams, and everyone was kept busy. Yet the project didn’t make any real progress in terms of completed features. Soon, the overtime bell began to ring.

Luckily, there was a skilled leader on one of teams. He decided to find out the root cause of what was holding the developers back. I opted in to provide some data as a basis for the discussions. Here’s the type of data we used:

Until now, we have focused our techniques around the code you’re changing. But a version-control log has more information. Every time you commit a change, you provide social information.

Have a look at the word cloud. It’s created from the commit messages in the Craft.Net repository by the following command:

The command extracts all commit messages. You have several simple alternatives to visualize them. The one was created by pasting the messages into Wordle.^[31]

If we look at the commit cloud, we see that certain terms dominate. What you’ll learn right now is by no means scientific, but it’s a useful heuristic: the words that stand out tell you where you spend your time. For the Craft.Net team, it seems that they get a lot of features in, as indicated by the word “Added,” but they also spend time on “Fixing” code.

On the project I told you about—the one that was running late and no one knew why—the word cloud had two prominent words. One of them highlighted a supporting feature of less importance where we surprisingly spent a lot of time. The second one pointed to the automated tests. It turned out the teams were spending a significant portion of their workdays maintaining and updating tests. This finding was verified by the techniques you learned in Chapter 9, ​Build a Safety Net for Your Architecture​. We could then focus improvements on dealing with the situation.

What story does your own version-control log tell?

Commit Messages Tell a Story

Commit clouds are a good basis for discussion around our process and daily work. The clouds present a distilled version of our team’s daily code-centered activities. What we get is a different perspective on our development that stimulates discussions.

What we want to see in a commit cloud is words from our domain. What we don’t want to see is words that indicate quality problems in code or in our process. When you find those indications, you want to drill deeper.

But commit messages have even more to offer; A new line of research proposes that commit messages tell something about the team itself. A team of researchers found this out by analyzing commit messages in different open-source projects with respect to their emotional content. The study compared the expressed emotions to factors such as the programming language used, the team location, and the day of the week. (See Sentiment analysis of commit comments in GitHub [GAL14].)

Among other findings, the results of the study point to Java programmers expressing the most negative feelings, and distributed teams the most positive.

The study is a fun read. But there’s a serious topic underpinning it. Emotions play a large role in our daily lives. They’re strong motivators that influence our behavior on a profound level, often without making us consciously aware of why we react the way we do. Our emotions mediate our creativity, teamwork, and productivity. As such, it’s surprising that we don’t pay more attention to them. Studies like this are a step in an important direction.

Data Doesn’t Replace Communication
	Given all fascinating analyses, it’s easy to drown in technical solutions to social problems. Just remember that no matter how many innovative data analyses we have, there’s no replacement for actually talking to the rest of the team and taking an active role in the daily work. The methods in this chapter just help you ask the right questions.