Team Project Submission (week 16)

What week 16 is for

Week 16 has a lecture only. There is no lab session, and no defence. The team project is finished when you submit the archive described below.

---

Run it yourself before you submit

Do not submit something you have not run. Open a terminal and actually run it:

a = Array([[1, 2, 3], [4, 5, 6]])
print(a)
print(a.transpose())
print(a.reshape(3, 2))
print(a + a)
print(a.sum(axis=0))

Then run the tests:

python3 -m pytest -v

A test suite going green is more persuasive than any sentence in your report. Save that full output — it is one of the things you hand in.

While you are at it, ask yourselves

These are the places a TA looks first. Go through them yourselves; anything you cannot answer is something that is not finished.

• Why did you choose that internal representation? What is wrong with the alternative?
• How is reshape implemented? Why does it not have to move the data?
• What happens on a shape mismatch? Is there a test for it?
• What is dot’s complexity? How long does 100×100 take?
• Which test was hardest to write? Why?
• Which parts did an AI write, and how did you verify them?

---

Marking criteria

25% of the course grade, allocated as follows.

Required functionality (40%)

Tests (15%)

Code quality and design (10%)

• complete type annotations (3)
• clear naming and structure, no duplicated code (4)
• consistent error handling (3)

Documentation (5%)

The README must contain:

• project conventions (internal representation, naming, error handling)
• how to install and run it
• usage examples (at least three)
• division of work
• known limitations

Technical report (10%)

• design decisions explained, with the trade-offs (5)
• test strategy and the problems you hit written concretely, not in platitudes (3)
• division of labour accurate and consistent with the code (2)

Optional extensions (up to +30%)

Capped at 100% overall.

---

Final submission (Sunday, 23:59)

One archive containing:

1. all code (implementation and tests)
2. README.md (see the documentation requirements above)
3. report.md, at most two pages, answering:
   • What is your internal representation? Why that one? What did you consider and reject?
   • Which part was hardest, and how did you solve it?
   • What was your testing strategy? Which bugs did cross-checking find?
   • Division of work: who did what, in what proportion
   • If you did it again, what would you change?
4. the complete output of pytest

---

Wrapping up

The final exam

• Time and place per the university notice
• Closed book, 50% of the course grade
• Scope and question types in L16, section 3
• Focus on: the frequent-traps table

How to revise

1. Work through L16’s frequent-traps table until you can explain each entry
2. Revise along the five threads (reference model, cost awareness, divide and conquer, verification, the dirty outside world), not lecture by lecture
3. Write code by hand — no completion in the exam hall, and no AI
4. Modify the runnable blocks in the notes and re-run them

Questions

• Ask in the course group any time
• There will be a revision session the week before the exam; the time will be announced

---

Finally

That is the end of the course.

Sixteen weeks ago many of you had never written a line of code. You have now finished a project of several hundred lines, with tests, with documentation, built by two people working together.

That is a larger distance than it currently feels.

Good luck in the exam. And I hope that after this course, the code you write is code people are willing to read.