S4 ICT: debugging and version control (challenge)
- 2026-11-20
S4 ICT: debugging and version control (challenge)
Third paper: defect removal efficiency, the lines behind a coverage rise, bisect steps over a million commits, mutation scores, static analysis and branch coverage.
Challenge: debugging economics and test strength The challenge paper turns debugging into comparable measures. Defect removal efficiency of 190 out of 200 is 95 percent, coverage rising from sixty to eighty percent on 12,000 lines adds 2,400 lines, a million commits need 20 bisect steps and 180 of 200 mutants killed is a 90 percent mutation score. A 2,000 production fix on a 100-fold cost ladder means 20 in development, and twelve percent of 500 static warnings are real, that is 60 defects. Eighty percent line coverage against fifty percent branch coverage shows one side of each decision untested.
01 Paper details (check answers afterwards)
| Item | Detail |
|---|---|
| Questions | 10 |
| Suggested time | 25 |
| Difficulty mix | Foundation 4 · Intermediate 2 · Challenge 2 · Error redo 2 |
| Chapter | Chapter: S4 ICT: Debugging and version control |
| Level | Challenge (10 items) |
02 Questions (10 items)
- FoundationA team finds 190 of 200 known defects before release. What is the defect removal efficiency?
- FoundationA 12,000-line system raises coverage from sixty to eighty percent. How many more lines are covered?
- FoundationUsing bisect across 1,000,000 commits, how many tests are needed at most?
- FoundationMutation testing kills 180 of 200 mutants. What is the mutation score?
- IntermediateA defect costs 2,000 to fix in production and production is 100 times the development cost. What does the same fix cost in development?
- IntermediateOne defect pattern appears 5 times and each trace takes 3 hours, with two developers splitting the work. What is the minimum time?
- ChallengeA static analysis tool raises 500 warnings and only twelve percent are real defects. How many real defects are there?
- ChallengeOne module shows eighty percent line coverage but only fifty percent branch coverage. What does that mean?
- Error redoIs it true that bisect tests every commit one by one?
- Error redoIs it true that revert deletes version history?
03 Answers and one-line notes
| # | Answer | One-line note |
|---|---|---|
| 1 | Ninety-five percent, since 190 over 200 is 0.95. | DRE counts only what was caught before release; defects reported by users later sit outside that denominator. |
| 2 | 2,400 more lines: sixty percent is 7,200 and eighty percent is 9,600. | Each ten-point rise on a ten-thousand-line system means over a thousand lines, so the cost must be estimated up front. |
| 3 | Twenty, because 2 to the power 20 is about 1,048,576 and already covers a million. | Logarithmic growth means a million versions cost only ten extra steps, which is the point of bisect. |
| 4 | Ninety percent, since 180 over 200 is 0.9. | A mutation score measures whether tests catch deliberately broken code, which is stricter than plain coverage. |
| 5 | 20, since 2,000 divided by 100 is 20. | Read the other way, catching it one stage earlier saves 1,980, which is what justifies automated tests. |
| 6 | Eight hours: 15 hours of work split two ways is 7.5, rounded up to 8. | Parallel work shortens calendar time but not total effort, and rounding up stays conservative. |
| 7 | Sixty, since 500 times 0.12 is 60. | Too many warnings stop people reading them, so thresholds and regular cleanup are needed or real defects stay buried. |
| 8 | It means most lines ran, but one side of many decisions never executed, for example an if whose false branch was never tested. | Comparing the two numbers exposes the gaps in testing, which a line figure alone hides. |
| 9 | No. It tests the middle version and discards half each time, so 1024 versions need 10 tests while checking each one could take 1024. | Treating a logarithmic search as a linear one is the commonest efficiency mistake in version control. |
| 10 | No. Revert adds an inverse commit that cancels the old change and keeps the record; history rewriting belongs to reset or rebase. | Keeping history is what lets you see who broke what and when, so pushed versions should be reverted. |
04 What to do after marking
- After marking, redo only the wrong items and write down which step went wrong (formula, unit, or misread number).
- Repeat the same item two days later; move to the next paper only when both attempts are correct.
- Parents only need the answer and note columns; there is no need to explain every item.
Want syntax, runtime and logic errors clearly separated, plus bisect debugging, commits and version numbers understood? Book a TutorZone tutor and settle the whole debugging and version control workflow in twenty-five minutes of ten items.Tutor Registration →
Questions and answers are written originally by the TutorZone team and are not taken from any school, publisher or examination authority paper. Every value is checked by computer.



