引導式測驗 · 中等 等級
程式設計課程 AI 自動評分系統測驗
測試自動評分器範圍、測試覆蓋率、回饋審查、再現性、公平性以及 GitHub 課堂的退役狀態。
問題 1 的 8
A student passes every configured test in a programming assignment. What does that establish most directly?
預覽每個問題8 問題
當您參加測驗時,就會出現答案和解釋。
- A student passes every configured test in a programming assignment. What does that establish most directly?
- The submission passed the behaviors exercised by that test suite
- The program has no bugs under any possible input
- The code meets every design and readability criterion
- The test suite covers all requirements even if some were not encoded
- Why should an instructor include edge cases in an autograder suite?
- To reveal the private expected outputs to students
- To make timeouts unnecessary for every program
- To catch failures outside the ordinary inputs used in examples
- To ensure all valid solutions use identical internal logic
- An AI model suggests a test that rejects a correct solution using a different algorithm. What should the instructor do?
- Hide the failure reason so students cannot dispute it
- Compare the test with the specification and reference solutions before grading
- Accept the test because AI-generated tests are objective
- Change the assignment specification to match the test after grading
- GitHub Classroom’s retirement took effect on August 28, 2026. Which description is now accurate?
- The Classroom service remains available to new organizations
- Only its autograding tests were retired, while the Classroom site remains active
- GitHub Actions was retired together with GitHub Classroom
- GitHub Classroom was retired on August 28, 2026; its old autograding docs describe a past service
- A grading suite compares output strings exactly. Which risk should be checked?
- Correct solutions with equivalent formatting may fail if the specification allows that variation
- Exact output comparisons always prove the internal algorithm is correct
- A formatting check makes edge-case tests unnecessary
- Every programming language produces identical output by default
- What helps reproduce a disputed autograder result?
- Only the student’s final score and course section
- A screenshot of the assignment title
- The model’s confidence in its grade
- The test version, dependencies, runtime, timeout, inputs, and recorded output
- How can an instructor reduce the chance that hidden tests distort grading?
- Let an AI model decide whether a failing test is fair
- Check known-correct and known-buggy solutions, include edge cases, and review score patterns
- Use hidden tests only and never inspect their failures
- Give every student the same score when a test is disputed
- Which feature is most appropriate for human review in addition to tests?
- Whether a program exits within its specified timeout
- Whether a dependency installs in the grading environment
- Code design, clarity, or reasoning that the test suite does not encode
- Whether a configured input produces its expected output