Build Better Software

Under development

Build Better Software provides foundational principles of software development tailored for scientists and researchers. It offers practical guidance, best practices, and illustrative examples to support effective, maintainable, and reproducible research software.

The accompanying repository, learn-dev, is available at: https://github.com/ExPlanetology/learn-dev

Dr. Dan J. Bower

Senior Scientist

Institute of Geochemistry and Petrology

Department of Earth and Planetary Sciences

ETH Zurich, Switzerland

Content

• 1. Why we are here
  • 1.1. Lessons learned the hard way
  • 1.2. Course structure
• 2. Top tips
  • 2.1. Foundational tips
  • 2.2. Intermediate tips
  • 2.3. Advanced tips
  • 2.4. Enjoy the ride
• 3. Object-oriented programming
  • 3.1. Class definition
  • 3.2. Using classes
  • 3.3. Class vs instance attributes
  • 3.4. Defining your own classes
  • 3.5. Exercises
• 4. Interfaces
  • 4.1. Interface definition
  • 4.2. Implicit or explicit?
  • 4.3. Exercises
• 5. Functional programming
  • 5.1. Benefits
  • 5.2. FP and OOP
  • 5.3. Pure function in a class
  • 5.4. Pure function with a class argument
  • 5.5. Exercises
• 6. Refactoring
  • 6.1. Anti-patterns
  • 6.2. Technical debt
  • 6.3. How to refactor
  • 6.4. Refactoring vs code creep
  • 6.5. Exercises
• 7. Writing new code
  • 7.1. Scenarios
  • 7.2. From science to code
  • 7.3. A common design tension
  • 7.4. Guiding principles
  • 7.5. API and library layer
  • 7.6. Exercises
• 8. Jupyter notebooks
  • 8.1. Introduction
  • 8.2. Strengths and limitations
  • 8.3. Practical engineering approach
  • 8.4. Notebook hygiene
• 9. Testing
  • 9.1. Types of tests
  • 9.2. When tests get written
  • 9.3. Test-driven development
  • 9.4. Python example
  • 9.5. Striking a balance
  • 9.6. Continuous integration
• 10. The journey continues

Bonus

• JAX
  • Overview
  • Mental model
  • Design considerations
  • Getting started

Appendices

• Glossary of Terms
• IDE setup
  • Why an IDE?
  • Visual Studio Code (VS Code)
  • AI agents in VS Code
  • Python
  • Mac specific
  • Windows specific
• Typing
  • Static vs. dynamic typing
  • Type hints in Python
  • Why use type hints?
• Documentation