Explore the concept of anchored metadata in software development through this 55-minute conference talk. Learn about an alternative approach to associating metadata with code that stores it separately using anchoring techniques. Discover how this method overcomes drawbacks of embedding metadata directly in code, such as language specificity and code clutter. Examine the 'spor' tool, which implements anchoring and automatic anchor updating using interesting text alignment algorithms. Gain insights into the tool's Python implementation and its partial Rust version. Understand how anchored metadata is applied in Cosmic Ray, a mutation testing tool for Python. Delve into topics like Smith-Waterman alignment, context snapshots, and the challenges of updating anchors as code changes. Consider future developments, including IDE integration and alternative tokenization methods.
Overview
Syllabus
Introduction
The problem
Cosmic Ray
Mutation Testing
Exceptions
Embedded if directives
Embedded if problems
Externalize
The Challenge
Research
SmithWaterman Alignment
Computer Science II
Context
Anchor
Contexts
SmithWaterman
The Fundamental Algorithm
Updating an Anchor
Python Implementation
Commandline Interface
Demo
Work to do
IDE integration
Anchoring directories
How we deal with this
Alternative tokenization
Scoring functions
Python to Rust
Speed
Learning Curve
Positive Experience
Cake vs Rust
Quick demo
Mirror display
Creating an Anchor
Taught by
NDC Conferences
