AI for Coding: Generate API Reference Docs That Match the Source

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-ai-coding-doc-gen-from-source-r8a1-creators

1. A development team notices their API documentation frequently shows incorrect parameter names in code examples. What is the root cause of this problem?

   1. The documentation team writes examples before the code is finalized
   2. The API uses too many parameters for humans to track
   3. Documentation lives separately from source code and drifts over time
   4. The documentation software has a bug in its rendering engine

2. Which task can an AI assistant perform most reliably when generating API documentation?

   1. Extracting function signatures and generating parameter tables from source
   2. Creating conceptual diagrams explaining API architecture
   3. Deciding which APIs are experimental versus stable
   4. Writing a getting-started tutorial for new developers

3. What does a 'doc-test' verify in automated documentation workflows?

   1. That documentation files are properly formatted in Markdown
   2. That API response times meet performance thresholds
   3. That code examples in docs actually execute without errors
   4. That all public functions have at least one docstring

4. An AI generates documentation for a public module but flags certain exports as 'undocumented.' What does this indicate?

   1. The module has no public exports
   2. The API is no longer supported
   3. The AI made an error in its description
   4. Those exports lack docstrings in the source code

5. Why is it risky to rely solely on AI-generated code examples in documentation without additional verification?

   1. AI cannot generate examples for async functions
   2. AI always produces incorrect syntax
   3. Code examples are copyrighted material
   4. Examples might look correct but use outdated parameter names

6. Which of the following is a task AI cannot perform without human guidance when producing API documentation?

   1. Extracting parameter types from function signatures
   2. Writing canonical descriptions of what a function does
   3. Generating return value descriptions
   4. Determining which APIs should be marked as experimental

7. A company wants to use AI to generate their entire API documentation suite. Which component would still require human writers?

   1. Exception documentation
   2. Return type descriptions
   3. Parameter tables for each function
   4. The getting-started guide for new users

8. What mechanism allows documentation to stay accurate after each code release?

   1. Generating reference docs from source on every release
   2. Hiring more technical writers
   3. Writing longer documentation
   4. Storing docs in a different repository

9. What does 'doc drift' refer to?

   1. The gradual movement of documentation files in version control
   2. A slow-loading documentation website
   3. Documentation becoming inaccurate because code changed but docs weren't updated
   4. Documentation teams moving to different offices

10. An AI documents a function with parameters (name: string, age: number). Later, the developer renames 'age' to 'yearsOld'. The AI regeneration finds a mismatch. What specifically caused the mismatch?

    1. The function was deleted from the codebase
    2. The documentation website crashed
    3. The generated example used the old parameter name in its code sample
    4. The AI forgot the original parameter name

11. Why is it insufficient to just generate documentation once and never regenerate it?

    1. Generated docs contain watermarks that expire
    2. Code changes over time, making generated docs stale
    3. Generation is too expensive to do once
    4. AI can only generate docs on weekends

12. What advantage does AI-generated reference documentation have over manually written reference docs?

    1. It always includes more examples
    2. It uses prettier formatting
    3. It can cross-check examples against actual function signatures
    4. It includes more jokes

13. A developer asks an AI to 'write documentation for my API'. The AI produces a conceptual guide explaining 'why' the API exists. Why might this output be unreliable?

    1. The developer didn't ask politely enough
    2. AI cannot write about concepts
    3. Conceptual guides are always wrong
    4. The AI doesn't know the product's purpose or user problems without explicit input

14. What should be included in a CI pipeline to prevent broken examples from reaching published documentation?

    1. A doc-test step that executes code examples
    2. A grammar checker
    3. A linter for comments
    4. A spell-checker

15. When generating reference documentation, what does AI extract directly from source code?

    1. The target audience profile
    2. Competitor analysis
    3. Function signatures and parameter information
    4. The company's marketing message