Collect Knowledge Graph Use Cases

Background

The X-Ray Graph project proposes using a knowledge graph approach to enhance our AI-powered developer tools. Before proceeding with full implementation, we need to validate that this approach will solve real problems and provide value that our current solutions cannot deliver.

Objective

Collect and validate specific use cases where a knowledge graph-based approach would provide capabilities beyond our current feature set. These use cases will help us:

  1. Validate the business value of the X-Ray Graph project
  2. Guide the implementation priorities for our proof of concept
  3. Identify potential early wins for the project

Some examples of the types of questions we may want to ask of a knowledge graph:

  1. If I modify this function, what other files/functions are impacted?
  2. Given this issue description, where should I start? - Repository navigation for new developers
  3. Can you tell me where this code is used for additional context with my code suggestion?

Investigation Areas

Current Limitations

Document specific examples where our current tools fall short:

  • Questions our AI tools cannot currently answer
  • Features we cannot implement with current architecture
  • Performance or accuracy issues in existing features

Target Teams for Input

TBD - Could be most Duo Features and beyond

Expected Outcomes

  1. Use Case Collection Document
  • Specific questions developers want to ask about their code
  • Current workarounds for missing features
  • Priority ranking of needs
  • Estimated impact of solving each use case
  1. Technical Requirements
  • Required relationships in the knowledge graph
  • Data freshness requirements
  • Performance expectations
  • Scale requirements (repository size, number of users)
Edited by Matt Nohr