Explore sustainable scientific software development practices in this EuroPython 2017 conference talk. Discover the importance of applying scientific method principles to computational science, ensuring accurate, reproducible, and reliable results. Learn about the challenges of testing and documenting complex scientific codes, and gain insights into tools and techniques for building robust, reliable software. Delve into topics such as testing methodologies, continuous integration, and documentation strategies specifically tailored for scientific applications. Understand the unique aspects of scientific software development, including handling random data, simulations, and numerical errors. Gain practical knowledge on implementing version control, unit tests, regression tests, and code coverage in scientific projects. By the end of this talk, acquire valuable skills to enhance the sustainability and reliability of your scientific software development process.
Overview
Syllabus
Introduction
About neutron stars
About my PhD
Overview
What is scientific software
What makes scientific software different
Scientific method
Computational science
Experimental sciences
Developer workflow
Version control
Commits
Collaboration
Testing
Scientific Codes
Unit Tests
Unit Tests Example
Regression Tests
Science Specific Issues
Random Data
Simulations
Numerical Errors
Code Coverage
Documentation
Distribution
Conclusion
Taught by
EuroPython Conference
