C1 Master Catalog

Architecture: Open Assessment Standard (OAS v1beta1)
Layer: C1 (Exercise Types / Structural Scaffolds)

Objective: Define agnostic logical containers that multiply raw content (C2) and evaluate it against legal rubrics (C0) across ANY subject.

In the ColabEdu architecture, an “Exercise Type” (Layer C1) is an empty skeleton, blind to the subject matter. It contains no specific questions or laws. It contains the User Interface (UI) logic and the baseline evaluation directive for the AI.

When the Automated Ingestion Pipeline runs, it crosses raw texts, maps, or formulas with these templates to generate millions of transversal combinations.

1. Domain: Written Production & Document Analysis

Interfaces to process information and generate argumentative responses. Useful for Language, History, Philosophy, and Biology.

1. Standard Split-Pane Guided Essay:
   • Language Use: Commentary on a poem.
   • History Use: Read the Declaration of the Rights of Man (C2) and write an essay on its causes.
   • Science Use: Read a biotechnology paper and summarize the hypothesis.
2. Email Response / Specific Format: The student assumes a role and replies.
3. Strict Objective Summary: Severely limits length (e.g., max 100 words). Penalizes personal opinions or “copy-paste”.
4. Graph & Infographic Analysis (Multimodal): The C2 block contains an image/chart.
5. Creative Writing / Hypothesis Continuation:
   • Philosophy Use: Resolve an ethical dilemma arguing from Kant’s perspective.

2. Domain: Comprehension & Interactive Evaluation

Deterministic auto-grading exercises and guided analysis.

6. Advanced Multiple Choice Questionnaire (MCQ): AI generates semantic distractors based on the C2 text.
7. Short Answer Questions: The student replies in 1-2 sentences. AI evaluates semantic accuracy.
8. Cloze Text (Fill in the blanks): The system removes keywords from a C2 text.
9. Logical Ordering (Drag & Drop): Order chronological events or scientific method steps.
10. Fake News Hunt (Fact-Checking): A news article is presented (C2), and the student interacts with a bot to verify facts.

3. Domain: Oral & Multimodal Interaction

Utilizes Speech-to-Text (Whisper) models and image validation.

11. Interactive Socratic Tutor (Oral/Text Chatbot): The student debates with the AI. (e.g., Historical Roleplay with Abraham Lincoln).
12. Comparison Monologue / Oral Defense: The student records timed audio defending a thesis.
13. Image Description (Visual Stimulus): The student records a description of an image (e.g., a plant cell).
14. Listening Comprehension / Note-taking: Playback of a podcast or lecture. The student answers questions.

4. Domain: Mechanics, Identification & Tagging

Exercises where the student highlights, underlines, or associates elements within a canvas.

15. Inline Tagging (Syntax & Fact Tagger): Underline elements in the text (e.g., causes in red, consequences in blue).
16. Indirect Transformation / Editing: A flawed text is presented, and the student must correct errors (wrong dates, grammar flaws).
17. Lexical Expansion / Key Concepts: Define a highlighted technical term within the context of that subject.

5. The STEM “Delta”: Future Required Molds

To cover hard sciences (Math, Physics, Programming), the system can be expanded with these 4 molds, keeping the exact same RAG and C0/C3 engine:

18. Equation Editor (Math/LaTex Editor): To solve calculus and physics step-by-step.
19. Interactive Graphs (Cartesian Canvas): To draw functions, parabolas, or vectors.
20. Code Sandbox: Code editor with algorithmic AI evaluation.
21. Interactive Map (GIS Point-and-Click): Blank maps for Geography or Anatomy.

The “Recipe Schema” (The secret of agnostic evaluation)

If the UI is the same (e.g., Split-Pane Essay), how does the AI know it’s grading a History exam and not a Literature one? Through Rubric Injection (C0).

• Scenario A (Literature): C1 (Essay) + C2 (Poem) + C0 (Rubric: Meter, Literary Figures).
• Scenario B (History): C1 (Essay) + C2 (1812 Constitution) + C0 (Rubric: Causes, Anachronisms). The AI will ignore poetic meter.

Conclusion

The power of the Lego Architecture (OAS V2) lies in its absolute agnosticism. If ColabEdu launches a product for Medical or Law Schools, not a single new screen needs to be programmed. All effort goes into adding new texts (C2) and laws (C0).