Skip to main content

Semester 0: Orientation: Answer Key and Solution Outlines

World-Class Evidence Integration

Calibrate the expected shape of World-Class Strand evidence without prescribing one tool, architecture, or implementation.

These outlines calibrate self-assessment for the semester-level checkpoint, exam, and project. They are intentionally concise: a correct submission may use different examples or wording, but it must show the reasoning and evidence described here.

Checkpoint Gate Solution Outline​

A passing checkpoint response demonstrates readiness to continue toward Semester 1 by showing independent command of CS orientation, algorithm exploration, computing maps, and disciplined learning workflow.

Expected evidence:

  • answers are written from memory first, then corrected with notes only after the first attempt
  • definitions are connected to concrete examples from the semester rather than repeated as isolated terms
  • problem-solving steps expose assumptions, invariants, tradeoffs, and failure cases
  • explanations include at least one connection to an earlier semester or prerequisite skill
  • remediation notes identify the exact concept, practice page, or project artifact that would close any gap

A strong checkpoint solution usually includes a short self-audit table: prompt, answer quality, evidence, uncertainty, and next action. The answer does not need to be polished prose, but it must be specific enough that another learner can tell whether the standard was met.

Semester Exam Answer Key / Solution Outline​

Use the exam prompts as a synthesis assessment, not a memory quiz. A complete solution should include:

  1. Concept accuracy: key terms are defined precisely enough to avoid misleading simplifications.
  2. Applied reasoning: each answer shows how the concept changes an engineering or mathematical decision in the semester domain.
  3. Worked method: calculations, proofs, traces, diagrams, or design alternatives are shown step by step where the prompt requires them.
  4. Tradeoff analysis: when multiple choices are plausible, the answer names the chosen constraint and explains what is sacrificed.
  5. Failure analysis: at least one answer identifies what would go wrong if the method were applied mechanically.

For Semester 0: Orientation, high-quality exam answers should repeatedly tie back to CS orientation, algorithm exploration, computing maps, and disciplined learning workflow. Award full credit for logically equivalent approaches that are correct, justified, and inspectable; do not require the exact vocabulary of the guide when the underlying reasoning is sound.

Semester Project Solution Outline​

The project is successful when it produces an inspectable artifact and a defensible explanation of its construction. The reference solution shape is:

  • Problem framing: the submission states the user, purpose, constraints, and non-goals.
  • Core artifact: the implementation, analysis, notebook, design packet, or capstone evidence directly addresses the required scope.
  • Verification evidence: tests, proofs, traces, benchmark results, diagrams, review notes, or operational logs support the most important claims.
  • Decision record: the learner records meaningful alternatives and explains why the final approach fits this semester's constraints.
  • Reflection: the final write-up names what changed in the learner's understanding and what remains risky or incomplete.

A valid solution may be smaller than an ambitious portfolio piece if the vertical slice is honest, reproducible, and aligned to the semester objectives.

Self-Calibration Procedure​

  1. Attempt the checkpoint, exam, or project review without looking at this page.
  2. Mark every answer or deliverable as meets, partial, or missing against the outline above.
  3. Repair only the highest-risk gaps first: conceptual misunderstanding, unverifiable claims, or missing core deliverables.
  4. Re-grade with the rubric and record the final decision: pass, conditional pass with remediation, or repeat.

Workload and advancement contract​

Use Workload and Progression for the complete hour allocation, prerequisite order, evidence reuse, score thresholds and repair rules. Its estimates do not establish learner-tested duration. All required gates apply together; conditional pass means repair remains, not permission to skip a gate.

Learning-support calibration​

For current learning-calibration work use the fresh scenarios in Module 4, Launchpad Module 4, Launchpad Module 5 and Launchpad Module 6. The Launchpad records are inherited entry evidence, along with its digital recovery and AI learning gates; do not repeat completed modules. A good answer distinguishes independent from assisted attempts, data from inference, feedback from a personal judgment and a feasible plan from an aspiration. Do not substitute these written examples for a learner's actual observations.

Technical practice calibration​

Use data calibration, HTTP calibration and AI-evaluation calibration. Preserve the distinctions between bytes/meaning, valid JSON/valid response and authored policy scores/live model performance. Each changed-requirement task requires an independent implementation or explanation under its rubric.

Concrete code-review and project calibration​

The exam uses Python and the current Algorithm Explorer project. No additional study-tracker feature or JavaScript knowledge is required.

Streak review: the loop never updates best after a final run of true values, and a later shorter run overwrites an earlier longer best. For [True, True], expected 2 but the defective function returns 0. For [True, True, False, True, False], expected 2 but it returns 1. An empty list expects 0. One repair updates best = max(best, streak) on every iteration after updating/resetting the current streak. Assume a list of booleans; if broadening the input domain, explicitly define validation rather than inventing a third mandatory defect.

Search review: the code mixes half-open initialization with closed-interval updates. In [1, 3] searching for 1, it can discard the remaining candidate and return -1. In [1] searching for 2, low = mid makes no progress and the loop repeats. Trace these with a step bound. A consistent closed interval uses high = len(values) - 1, loops while low <= high, advances low = mid + 1 and otherwise sets high = mid - 1. Empty input returns -1. For duplicates, any matching index is acceptable unless a first/last-match contract was explicitly added.

Project data shape: a saved scenario can be a dictionary containing an algorithm identifier, input values and optional target; trace steps can be a list of plain records with boundaries, current operation and visible state. Generate/validate the trace outside display code, and serialize a declared JSON shape for this small project. A plain terminal display is valid. Explain assumptions and invalid-input handling; a database or web framework is unnecessary.

Scoring: parts A–D each contribute 25%. Within A, score its five questions equally; within B, its three challenges equally; within C, its two reviews equally. Score each response on four equally weighted criteria: correct result, explicit reasoning, relevant example/test and stated assumptions or limits. Use 0 (missing/incorrect), 1 (partial), 2 (complete) per criterion, divided by 8 and multiplied by 100. For D, use its four displayed 1–4 criteria and convert with 100 × (sum - 4) / 12; a minimum-description response should not receive an automatic 25% baseline. Preserve criterion scores. The 70% overall and 60% per-part floors and all separate gates remain unchanged.