katrinleinweber / Git-for-bio-scientists

Slides of this presentation with recorded audio (archived notes)

Tree of Life

Tree of Life

Tree of Projects

Git tracks digital evolution of files

• highlights exactly what changed, line by line
• makes you comment file changes

Git is your file watchdog.

Good against: too many files

Good for: tidy project folders, but...

Good for: ...version history still accessible

Version control with Git

Technical

• works best for line- or paragraph-based files
• built for source code, but also useful for texts and simple images
• less useful for complex or large file types
• not a substitute for full backups

Cultural

• helps build self-explanatory & verifiable project documentation (for your future self, colleagues, students, reviewers...)
• enables less painful collaboration on same set of files
• conflicts only if same lines in same file get changed
• for text files: better than syncing tools

Basic Git vocabulary: repository/repo

• project folder watched by Git
• is still a normal folder: add, edit & delete files normally
• database of file versions in hidden subfolder

Basic Git vocabulary: committing

• saving a logical set of file changes in Git, together with meaningful comment
• changes can be within single file, or across different ones
• commit history: self-explanatory & verifiable documentation

Basic Git vocabulary: .gitignore

• file that contains a repo's rules for ignoring files
• No preview of highlighted changes? Large file? Auto-generated? Rather ignore those!

Basic Git vocabulary: .gitignore in GitHub for Windows

Vocabulary summary

repository/repo

• project folder watched by Git, plus hidden database of file versions

committing

• composing logical set of file changes and meaningful commit message

.gitignore

• a repo's ignore rules for non-essential files & file types

How to start using Git?

1. download client from windows.github.com (any other from git-scm.com/downloads/guis) & install
2. start with small sub-project (protocol optimisation, diagram in R, report, etc.)
3. besides working on the files themselves: commit logically connected changes often
4. recommended: start writing texts in Markdown¹ format
5. optional (for collaboration etc.): create account with Git hosting service like GitLab or GitHub

Thanks to & Further Reading

Writing with Git

• "Markdown Basics" and "Mastering Markdown" by GitHub
• "Scientific Markdown" by Jens Erat
• "Paper Now" by PeerJ

Git in Science

• "Git/GitHub: a Primer for Researchers" by Carly Strasser
• "We Need a Github of Science" by Marcio von Muhlen
• "Git for Scientists" by Molly Gibson

Git concepts & tech

• "Learn Version Control with Git" by Fournova
• "10 Years of Git: An Interview with Git Creator Linus Torvalds" by Jennifer Cloer

Thanks for your attention! Questions?