Lab 7 — The Debugging Clinic (Lecture 07 follow-up)

Machine lab · 2 hours · Pairs; rotating patients. Three "patient" programs, each with one planted defect, each failing a different way. You will practise the method from lecture 07 — reproduce, isolate, hypothesise, test, fix, re-run — under time pressure, then trade your method notes. Submit: the checkpoints at the end.

Setup

Python 3 as python; work in lab7/.
Your instructor hands out patient_a.py, patient_b.py, patient_c.py (prepared versions of the snippets below, wrapped so each runs and prints a wrong answer rather than crashing — silent wrong answers are the hardest kind and the point of this lab).

The patients

# Patient A — "average marks"  (prints a plausible but wrong average)
marks = [72, 85, 90, 60, 55, 91]
total = 0
for i in range(1, len(marks)):
    total = total + marks[i]
print("average:", total / len(marks))

# Patient B — "pass counter"   (counts one student twice / one never)
marks = [72, 85, 90, 60, 55, 91]
passes = 0
for m in marks:
    if m >= 60:
        passes = passes + 1
    if m > 60:
        passes = passes + 1
print("passed:", passes)

# Patient C — "last over first" (reports the wrong record)
marks = [72, 85, 90, 60, 55, 91]
names = ["ann", "bo", "cy", "di", "el", "fay"]
best = 0
best_name = ""
for k in range(len(marks)):
    if marks[k] >= best:
        best = marks[k]
        best_name = names[k]
print("top student:", best_name)

Part A — Reproduce and observe (15 min)

A1. Run each patient. For each: the observed output, the expected output (derive it by hand first — do not guess), and a one-line symptom description ("off by one mark", "counts 90 twice", …).

Part B — Isolate and fix, methodically (45 min)

B1. For each patient, before fixing: state the hypothesis ("the loop starts at the wrong index"), the observation that confirms it (a print of the loop variable), and then the minimal fix. B2. After each fix, re-run the patient and one input where it was already behaving (regression check). A fix that changes correct behaviour elsewhere is a new defect. B3. Log each patient as: symptom → hypothesis → confirming observation → fix → re-run result. The log is the deliverable, not the fix.

Part C — The method swap (20 min)

C1. Swap logs with another pair. For each of their patients, answer: did their hypothesis come before their fix? Was their confirming observation actually confirming (would it also fit a different hypothesis)? C2. One paragraph: which patient's defect was found fastest by your method and which would have been found faster by just reading the code? Say what that means for when the method pays.

Part D — Beyond the given (10 min)

D1. Patient C has a second, subtler issue tied to its tie behaviour. Find it, and state the input where it matters. D2. One sentence: why does a silent wrong answer (no crash) demand a different working style than an exception does?

Checkpoints (submit these as your report)

The A1 symptom table (observed vs expected, derived by hand).
The three B3 logs, in full.
Your C1 review of another pair's log.
Your C2 paragraph.
D1's tie-behaviour defect and its exposing input.
Your D2 sentence.

What completion looks like

Completion-graded as in Lab 1. Full credit requires the log order (hypothesis before fix) even for the patients you solved at first sight — the discipline is the deliverable; speed comes later.