NASA’s new dark energy space telescope can also detect killer asteroids
technologyreview.com·2h ago
AI Research
arXiv CS.AI·Jun 9·5 min read
Jas: AI-Paired Engineering as a Revival of N-Version Programming
TL;DR
Software engineering often struggles with ensuring code reliability across different programming languages. An AI-assisted approach was used to create five ports of a vector illustration application in Rust, Swift, OCaml, Python, and browser-based platforms, guided by a YAML specification and differential testing.
✦ Why It Matters
Engineers can leverage AI and structured specifications to improve the efficiency and reliability of multi-language software development.
Key Takeaways
- Utilize AI tools to assist in software development tasks.
- Implement a detailed specification to guide development.
- Adopt parallel implementations for effective testing.
- Explore N-version programming principles to enhance reliability.
How It Works
The methodology combines AI assistance with a precise YAML specification, which serves as a comprehensive guide for development. This specification ensures consistency across different implementations.
Additionally, parallel implementations allow for differential testing, where the outputs of various versions can be compared to identify discrepancies and ensure correctness.