Almost every modern computer science degree program requires its students to complete a Big Project. Sometimes it's the "Senior Project," and sometimes it's the "Capstone Project." Whatever it's called, the purpose of this Big Project is to expose students to "real" software engineering practices.

In other words, this typically translates to "coding with other people." Unfortunately, up until this point in a student's education, this has usually been discouraged as "cheating."

The problem is that these Big Projects actually tend to focus on extremely bounded problems. Most of the time, a small team of students works on a small project for a semester, and the result is, quite naturally, a small project. While good learning can take place in a small project, it actually does very little to prepare students to work on Really Big Projects.

To find Really Big Projects, one must venture out into the world, where there are Really Big Problems. The real world is full of gigantic applications that require build systems and revision control and defect tracking and prioritization of work. They are written in languages that one may or may not know, by people one may or may not ever meet. And in order to successfully navigate through these Really Big Projects, the novice developer must possess one skill above all others: the ability, in the words of our colleague David Humphrey, to be "productively lost."

The great advantage of open source, for the learner, is that the Really Big Projects of the open source world provide unparalleled opportunities to be productively lost. Complex codebases are immediately accessible by anyone who wants to participate. This accessibility is crucial to the learner, as participating in an activity is by far the most effective way to learn that activity.

Sooner or later, the coder aspirant must work at scale, with teammates. Open source provides that opportunity in a way that nothing else can.