Month 10-Week 2: Capstone build sprint 1¶
Week summary¶
- Goal: Heads-down build week. 3-5 substantive features added. Eval after each. Read source from one respected library and steal patterns.
- Time: ~9–10 h over 3 sessions.
- Output: Capstone with 3-5 new features, eval results updated daily, source-reading notes.
Why this week matters¶
Velocity weeks are how capstones happen. The discipline of eval after each feature is what makes the project real (vs claimed).
Prerequisites¶
- M10-W01 complete.
Recommended cadence¶
- Session A-Tue/Wed evening (~3 h): 1-2 features
- Session B-Sat morning (~4 h): 1-2 more features + source-reading
- Session C-Sun afternoon (~3 h): 1 more feature + eval week roundup
Session A-1-2 features¶
Goal: Add and test 1-2 substantive features.
Part 1-Pick (15 min)¶
From your DESIGN's roadmap. Cut anything that takes more than ~75 min in this session.
Part 2-Build (150 min)¶
Heads down. Tests where applicable.
Part 3-Eval (15 min)¶
Re-run eval. Note what changed.
Output of Session A¶
- 1-2 features shipped + eval delta noted.
Session B-Source-reading + features¶
Goal: Read source from a respected OSS library. Add features informed by it.
Part 1-Source-reading (75 min)¶
Pick a library you respect in your specialty. Read 200-500 lines.
What's elegant? What's pragmatic? What pattern would steal-with-attribution?
Part 2-Features (150 min)¶
Apply lessons. Add 1-2 features, ideally informed by what you read.
Part 3-Eval (15 min)¶
Re-run.
Output of Session B¶
- Source-reading notes.
- 1-2 more features shipped.
Session C-1 feature + week roundup¶
Goal: One last feature. Recap the week's eval delta.
Part 1-Build (90 min)¶
Part 2-Aggregate eval delta (60 min)¶
For each feature added this week, what did the eval do? Build a simple table.
Part 3-Forward look (30 min)¶
Re-read DESIGN. Are you on track for v0.1 by end of M10?
Output of Session C¶
- Final feature.
- Week-over-week eval delta documented.
End-of-week artifact¶
- 3-5 substantive features added
- Daily eval results
- Source-reading notes
End-of-week self-assessment¶
- I can articulate what each feature added quantitatively.
- My capstone is materially better than at start of week.
Anti-patterns this week¶
- Refactoring before features work. Make it work, then make it nice.
- Skipping eval to "get more done." False economy.
- No source-reading. Borrowed patterns are how libraries grow up fast.
What's next (preview of M10-W03)¶
Find a first user. Watch them use the tool. Fix the top confusions. The single highest-leverage 15 minutes of M10.