For these (and many other!) reasons, we strongly encourage people to use some form of version control as a part of best practices for research data management, reproducible research, and documenting her/his scripts and code changes over the course of a research project (see our Best Practices page). Version control can be used to keep track of versions of a piece of work that either a single person is working on, or a shared document.

Version control systems start with a base version of the document and then save just the changes you made at each step of the way by taking a so-called "snapshot". A snapshot records information about when it was taken, and all the changes between the current document and the previous version. The user (you) decides when these snapshots are collected, and this allows one to ‘rewind’ your file to an older version. Multiple users can make independent sets of changes based on the same document and have separate snapshots documenting the changes. You can then "merge" various sets of changes onto the same base document. Version control systems are used to track changes in text documents (HTML pages, text files, Markdown documents, etc) and binary files (MS Office files, images, etc), though the changes in these types of files are tracked less efficiently (and is beyond the scope of this document).

There are many different version control systems available, but one of the most widely used is Git, which is the program that will track changes of your scripts (source code) over time. And Git should not be confused with GitHub, which is a hosting service for Git repositories, the bundle of files and folders that represent your work and the typically-invisible data files that track your changes & history. Other hosting services include GitLab, SourceForge, and BitBucket.

If you'd like general information on how to use Git, please see the excellent, self-paced materials for Git at the command line from Software Carpentry. For a GUI version of Git, we recommend GitKraken, and our own GitKraken usage tutorial.

Git and GitKraken are installed on the HBSGrid. GitLFS, the extensions for handling files larger than 50 MB, is also installed.

Two of the most commonly used hosting platforms within the Harvard community are described below.

Members of the Harvard community can take advantage of Harvard’s Git hosting platform at https://code.harvard.edu by signing in with their Harvard Key (an account will automatically be created). Features of this platform include:

• Public repositories are only visible to people within Harvard
• Private repositories are only visible to the user and collaborators
• Maximum total file size for a repository is 1GB
• Users can create an unlimited number of repositories

Free accounts can also be created on github.com, a web-based code repository. While similar to the repositories at code.harvard.edu, the platform has slightly different features for the free accounts:

• Public repositories are open to everyone on the platform
• Private repositories are only visible to the user and collaborators, but the free account limits the features and only allows you to share with up to three collaborators
• Maximum total file size for a repository is 100GB, but if your repository exceeds 1 GB GitHub support may contact you to reduce the size
• Users can create an unlimited number of repositories