
make one data-based revision to my team’s code, test it in a new situation, and compare the new version with the old one.
Phase 4: Build
One change at a time. Ava’s team changed three things at once and their code got better. Which change helped? Think about it.
Why should a team change only ONE thing before testing again?
Revise with evidence. You have three kinds of evidence now: your test run data, your checkpoint, and your feedback notes. A good revision uses all three. Pick the change that best helps your criteria.
What a design needs. Look back at what you did: you did research, you built, you tested, and you talked with other teams. Each step made your design better. Engineers do all of these, not just the building.
Worked example 1: Malik’s revision. Malik’s team had 2 of 4 correct from far away. Feedback said “3 and 4 flashes look alike.” They changed “line up” from 4 flashes to 1 long flash. They kept everything else the same.
| Version | Situation | Correct (out of 4) | Time (seconds) |
|---|---|---|---|
| Version 1 | Far, noisy | 2 | 35 |
| Version 2 | Far, noisy | 4 | 30 |
Version 2 was more accurate AND faster. The data shows the revision helped.
Worked example 2: comparing patterns. Priya compares Version 1 and Version 2. What is the same? Both use light flashes. What is different? Version 2 mixes long and short flashes, so each message looks more different. Different patterns are easier to tell apart.
Rehearse. Practice the full field test with your team: sender, receiver, crowd, and timer. Test in a NEW situation, like lights off or near a window.
Make your revision, then rehearse. Record both versions.
| Version | Situation | Correct (out of 4) | Time (seconds) | What we changed |
|---|---|---|---|---|
| Version 1 | (none) | |||
| Version 2 | ||||
| Version 2 | New situation: |
Reflect on the whole build.
Describe what your team needed to design its solution: what research you used, what you learned from talking with others, and how testing made it better.
How confident are you that you can I can use data to show whether my revision made my design better.
Your code is revised and rehearsed. The Build phase is done! Next: your 3DX field trip, from the telegraph to fiber optics. Then the real field test.
Exit ticket: compare Version 1 and Version 2 of your code. Name one thing that stayed the same and one thing that changed.
Materials: prototypes, code keys, stopwatches, feedback notes from 2.20, checkpoint results from 2.19.
Setup: teams that need re-teaching from the checkpoint meet with you for the first 8 minutes while others plan revisions.
Listen for: teams changing many things at once; teams who revise without evidence; “faster is always better” (accuracy matters most for a help signal).
Support / stretch: give a revision planning frame (“We noticed… so we will change…”). Stretch: test a third version that combines two teams’ best ideas.