Why invest in Open Source programming skills?

WoCCode Summer 2023 Bootcamp

Good Research is enabled by Good Software Engineering

https://hdsr.mitpress.mit.edu/pub/f0f7h5cu/release/2

3Rs: readability, Resilience, and Reuse

• Readable code is an investment in future you
• Informative variable name choices
• Commented code with descriptive docstrings
• Intuitive structure and order of operation
• Resilient code that “breaks gracefully”
• Provide checks with useful error messages
• Modular design, minimal repetition
• Testing infrastructure to build confidence and reliability
• Reusable code
• Less time spent reinventing / repeating tasks
• Make it general enough to use in many scenarios

Different Scenarios have Different Practice

Personal Use

• Data visualization code (e.g., matplotlib style files, color maps, corner plot)
• Commonly repeated tasks (e.g., github.com/eblur/gastronomy)
• Can still be made public as a service to others
• Can still benefit greatly from version control (e.g., Github)
• Can operate privately or publicly

Different Scenarios have Different Practice

Lab Use (2-20 users)

• Comments and docstrings need to be clear enough for multiple people to read and learn from
• Version control and testing ensures reliability for all, allows for multiple contributors and updates
• Can operate privately or publicly

Different Scenarios have Different Practice

Open Source Community

• Astropy, numpy, scipy
• Provides toolkit for other projects
• Requires full software engineering toolkit (testing, documentation, continuous integration, deployment, community input and governance)
• To participate: discuss needs, submit an issue, and/or submit a pull request

Not just good for you: Good practice is the Gold Standard

There is a world-wide push towards FAIR data principles

https://www.nature.com/articles/sdata201618

https://www.nature.com/articles/s41597-022-01710-x

Not just good for you: Good practice is the Gold Standard

https://science.nasa.gov/open-science-overview

How we will get there: Readability and Reuse

• Python in practice
• Object Oriented Programming
• Docstrings
• Code design

How we will get there: REsilience

• Version control
• Code design
• Code testing
• Optimization

Your Homework: Decide on a project (aka “Hack”)

Some ideas: https://docs.google.com/document/d/1htz32t6Lfe2wiX4lmEhcrfPnpCUkDcLq0pBztzoyPtc/edit?usp=sharing

Research task that you are already working on? (e.g., visualizing project results, performing a computation)

Project you have always wanted to try? (e.g., design a game, explore a dataset)

Something not too big! (i.e., not a PhD thesis) However, you might set up the code infrastructure to design and test a lengthy problem.

Something that can use thought on design, testing, and optimization.