464: Modelling the stars with Rémy Hannequin

What It Covers

Rémy Hannequin discusses building Astronomy, a Ruby gem for celestial body positioning calculations, covering domain-driven design patterns, value objects, and translating complex astronomical mathematics into developer-friendly Ruby code for telescope positioning.

Key Questions Answered

• •Value Objects for Units: Wrap raw numbers in objects to prevent unit conversion errors (meters vs miles) and enable natural operations like distance addition while automatically handling conversions, similar to Rails' ActiveSupport Duration with five.days syntax.
• •Domain-Driven Design Translation: Use terminology from scientific literature when naming code objects and methods to enable direct collaboration between domain experts and developers, eliminating translation overhead and making contributions more accessible to astronomers reviewing code.
• •Light Time Correction: Astronomical positioning requires correcting for light travel time since observed celestial bodies appear where they were in the past—moon one second ago, sun eight minutes ago, planets thirty to forty minutes ago—creating difference between geometric and apparent positions.
• •Performance Trade-offs: Value objects improve safety and readability but hurt performance in high-iteration calculations. Switch to raw numbers (like radians) when instantiating and converting objects repeatedly becomes a bottleneck, prioritizing practical results over architectural purity.