Software Design for Scientists

Chapter List:

What is Software?

• Something one tells a computer to do, no matter how small
• Should you write something?
• Improvements to science
• Finding and implementing what is already there
• Making it work
• Making it answer the questions you want it to

User-Centered Design

• Who are your users?
• Think about software as always having a user who's not you
• General software design principles
• You're not going to necessarily become a great developer, but can be a pretty good developer by thinking of your code as having a life cycle
• Think beyond the script to do one experiment

Steal, Share, Repeat

• Participate in open source
• Two way street
• Everyone benefits
• Steal like a developer - learn & build on the code don't just slot in a pretty good solution because you don't know how to do it

Scientific Reproducibility

• Requires codesharing as well as data sharing
• ... And there are so many good options!
• Requires a community to reuse and/or update the code.

The Scientific Method of Software: Agile

• Why agile will feel more natural to scientists (than waterfall)
• Agile isn't one thing and you don't have to follow a particular process
• Agile is a mindset not a process

Data Artifacts Part 1: Bugs

• Not letting your software work against you

Data Artifacts Part 2: Communicating Parameters

• Making sure that everyone gets the same answer, every time

Software Solutions Part 1: QA

• Pyunit
• Flask and coverage.py for code coverage
• But what about R?

Software Solutions Part 2: Relatively Painless Documentation

• Documentation for science part
• Documentation for software part
• Keep em separate as practical
• Do both
• What software artifacts are most useful
• Making it part of experimental design

Database Design: Setting Yourself up to Make the Data Work for You and Others -- Not the Other Way Around.

• Solr faceting of flat files
• Organizing your data - think beyond relational databases

Scoping Data Science Projects, why it is the most important difficult thing you will ever do.

Appendices:

Appendix 1: A glossary of useful terms, circa 2016.

Appendix 2: Some tools that might be useful to you.