Skip to main content

Git & Version Control

What is Version Control?​

A Version Control System (VCS) tracks changes to files over time, allowing you to recall specific versions, compare changes, and collaborate without overwriting each other's work. Think of it as an "unlimited undo" combined with a complete audit trail of who changed what and when.

What is Git?​

Git is the most widely used distributed VCS, created by Linus Torvalds in 2005 for Linux kernel development. Unlike older centralized systems (CVS, SVN), every Git user has a complete copy of the repository history, enabling offline work and robust backup.

Git excels at:

  • Tracking changes to code, scripts, manuscripts, and configuration files
  • Branching and merging for parallel development
  • Collaboration across teams and institutions
  • Reproducibility by linking analysis code to specific data versions

GitHub vs. GitLab​

Both are web platforms built around Git, but they serve different needs.

GitHub is the largest public code hosting platform, home to most open-source projects. It's ideal for public collaboration, sharing code with the community, and building a visible portfolio. However, free accounts have limited private repository features, and data is stored on US servers.

GitLab can be self-hosted, which is why many institutions run their own instances. LRZ GitLab (gitlab.lrz.de) is the recommended choice for TUM researchers working on non-public projects — your data stays within the European research infrastructure, and you get fine-grained access control for team collaboration.

In practice: Use GitHub for open-source and public projects; use LRZ GitLab for internal research, unpublished work, or anything requiring institutional data governance.

Key links:

  • GitHub: github.com
  • LRZ GitLab: gitlab.lrz.de — Use your LRZ/TUM credentials; recommended for non-public research projects

Additional Git functionalities: Git Pages & Wikis​

Git Pages (GitHub Pages / GitLab Pages)​

Host static websites directly from your repository — perfect for:

  • Project documentation
  • Lab websites
  • Research data landing pages
  • This Knowledge Hub itself!

Git Wikis​

Every repository can have an associated wiki for:

  • Extended documentation
  • Meeting notes
  • Onboarding guides for new team members

Both are version-controlled and free with your repository.

Should You Use Git?​

Git makes sense for:

  • Writing code, scripts, or analysis pipelines
  • Collaborative writing in text-based formats (LaTeX, Markdown, R Markdown)
  • Configuration files, workflow definitions, and documentation
  • Any project where you need to track changes over time or collaborate with others

Git requires extra consideration for:

  • Large binary files (images, videos, compressed archives) — Git stores every version, so repositories grow quickly. Consider Git LFS or external storage for large assets.
  • Sensitive or unpublished data — Never put research data on public GitHub. Use LRZ GitLab or institutional storage with appropriate access controls.

Git may not be the right tool for:

  • Purely experimental lab work with no computational component — a well-organized folder structure and regular backups may be sufficient.
  • Non-technical collaborators who work exclusively in Word/Excel and are unfamiliar with version control — the learning curve may outweigh the benefits for short-term collaborations. (Though for long-term projects, it's often worth the investment.)
  • Very large datasets — Git is designed for code, not data. Use dedicated research data repositories (Zenodo, institutional storage) and reference data via DOIs instead.
Rule of Thumb

If your work involves text-based files that change over time and you want to track who changed what and when — Git is likely worth learning.

Getting Started​

This is not a tutorial — excellent resources exist:

For research data management, the key insight is: Git tracks your analysis code; your data repository tracks your data. Link them via DOIs and version tags for full reproducibility.