AP Computer Science Principles · Unit 1 lecture · Part 1 of 11 · 6 min

1.1 CollaborationWhy the best programs have more than one author

By the end: Explain how people build programs together, describe a program by its purpose, inputs and outputs, follow the iterative and incremental development process, and find, name and fix syntax, logic, run-time and overflow errors with well-chosen tests.
Mr. Thompkins
Hi, I'm Mr. Thompkins, Mr. T if you like, and welcome to AP CSP. Let's start with a puzzle: a navigation app worked perfectly for the team that built it, and then failed thousands of its first users. Nothing in the code was broken. Today you'll see how that happens, and how a team prevents it.

Almost every program you use, from the game on your phone to the software inside a car, was built by a team. That isn't just because programs are big. It's because a team sees more than one person can. The College Board calls the programs and devices that change how we live computing innovations, and the first idea of this course is that people who work together, with different backgrounds and points of view, build better ones.

?Wait, what counts as a computing innovation?
A computing innovation is anything that has a computer program as an essential part of how it works. It can be physical, like a fitness band or a self-driving car; nonphysical software, like a photo editor; or a nonphysical concept, like online shopping or ride sharing, which only exists because programs run it.
Kind of computing innovationExampleWhere the program is
PhysicalA fitness band, a self-driving carInside the device, reading sensors
Nonphysical softwareA photo editor, a music appThe software is the innovation
Nonphysical conceptOnline shopping, ride sharingPrograms run the whole system
Three kinds of computing innovation. All of them depend on a program.

So why do different perspectives matter so much? Every developer has blind spots. If everyone on a team is a 17-year-old with perfect eyesight who speaks English, they will build an app that works beautifully for 17-year-olds with perfect eyesight who speak English. They won't notice that the text is too small for a grandparent, that a screen reader can't describe the map, or that the menus only come in one language. Nobody did anything wrong on purpose. They just couldn't see what they had never experienced.

That is what the course means when it says collaboration that includes diverse perspectives helps avoid bias in the development of computing innovations. Bias here means the program works better for some groups of people than for others, usually without anyone intending it. The cure is to bring more kinds of people into the process: on the team, among the testers, and in the conversations with users.

who tests itScreen reader userGrandparentSpanish speakerBus appwhat they catchMap has no labelsText far too smallEnglish onlymore perspectives, less bias
The same bus app, tested by three different people. Each one finds a problem the original team never noticed.
?Wait, does bias mean someone was being unfair on purpose?
Not usually. In computing, bias means a program's results or design favor some groups over others. Most of it comes from who was (and wasn't) involved in designing, building and testing the program, or from the data it learned from. You'll see the data side in Units 2 and 5.

Collaboration isn't only about avoiding harm. Teams also get more done and think of more ideas. When people with different talents share the work, one person might be great at layout, another at logic, another at talking to users. The innovation that comes out reflects all of those talents, which is something no single developer can match. Notice too that "the team" includes the users: consulting and communicating with the people who will use a program, and gathering information from them, is a form of collaboration.

Pause and think
A team of five builds a voice assistant. All five grew up in the same city and speak with the same accent. Before you read on: what problem is likely, and what is the simplest fix?
I thought about it. Show me
The assistant will probably understand their accent well and other accents poorly, because that's all it was tested on. The fix is to collaborate more widely: include testers and team members with many accents, ages and speech patterns, and gather feedback from them before release.
Explain how collaboration helps
A school builds an app for signing up for after-school clubs. The whole development team is made of seniors in the robotics club. Describe one way the app might be biased, and explain how collaboration could reduce that bias.
  1. Picture who is missing from the team: younger students, students who aren't in clubs yet, students who mostly use a phone instead of a laptop, families who read another language.
    Why?
    Bias comes from missing perspectives, so start by listing whose perspective is missing.
  2. Name a specific, likely bias: the team might list robotics and other tech clubs first, or design the sign-up for a laptop screen, so phone users struggle.
    Why?
    A specific example earns credit. "It might be biased" by itself does not.
  3. Explain the fix as collaboration: invite freshmen, students from different clubs and parents to test a prototype, or survey them, and change the design based on what they report.
    Why?
    The course wants the cause-and-effect: more perspectives, so problems are found, so the design changes.

Answer: For example, the app could be built for laptops because seniors use them, which makes sign-up hard for phone users. Testing with freshmen, other clubs and parents would bring in those perspectives, reveal the problem early, and lead to a design that works for everyone.

Stuck? Get a hint
  1. Who is on the team, and who will use the app?
  2. Think of one group that isn't on the team and one way the app could fail them.
  3. Then say how bringing that group into testing or design would reveal and fix the problem.
Check yourself
A company is designing software that suggests running routes. Which action would best help the company avoid bias in its design?
  1. AHiring more programmers who have the same experience as the current team
  2. BGathering feedback from runners of different ages, abilities and neighborhoods while the software is being built
  3. CReleasing the software as soon as possible and fixing problems later
  4. DKeeping the design secret until it is finished so ideas are not copied
Show the answer

B. Bias is reduced by bringing in diverse perspectives during development, so problems are found and fixed before release. The first choice is the trap: more people with the same background add hands, not perspectives.

  • A computing innovation has a program as an essential part. It can be physical, nonphysical software, or a nonphysical concept.
  • Collaboration that includes diverse perspectives helps avoid bias: people with different experiences notice different problems.
  • Teams produce innovations that reflect the diversity of their talents. Users are part of the collaboration too.