What are Git and version control?

Git and version control: definition

Version control stores the history of a project: who changed what, when and why? Instead of copying files such as “quote_final_new2,” the system records every change as a traceable step. For software projects this is indispensable, because many people work on many files at the same time.

Git is the most widely used system for this. It was created in 2005 by Linus Torvalds for the development of the Linux kernel and is distributed: every developer has a complete copy including the history. Repositories are often hosted on platforms such as GitHub, GitLab, Bitbucket or Azure DevOps.

How Git works

The project lives in a repository. Changes are saved as a commit—a snapshot with a short description. For new features or bug fixes, you create a branch, a separate line of development that does not disturb the main branch. When the work is done, the branch is merged back into the main branch.

On platforms such as GitHub and GitLab, this usually happens via pull requests or merge requests: the change is proposed, reviewed (code review), tested automatically and then accepted. Releases are marked with tags, so that every published state can be reproduced exactly later.

Real-world example: a faulty update in an online store

An online retailer has its store extension developed further on a regular basis. After an update, one shipping method no longer works and orders get stuck. Because every change is documented as a commit with author, time and description, the team sees within minutes which change triggered the problem and who made it.

It returns to the last working state (rollback), fixes the error in a separate branch and applies the correction again after a test. Without version control, you would have to search through file copies. In addition, the source code is not only on one developer’s computer but also on the platform—another gain.

Benefits, limits and common mistakes

Git provides traceability, parallel work and quick rollbacks, and forms the basis for automated testing and delivery. The learning curve is noticeable, Git is poorly suited to very large binary files, and without clear rules for branches, a repository quickly becomes unruly. Typical mistakes:

• Storing passwords, keys or credentials in the repository
• Working directly on the main branch, without branches and review
• Meaningless commit messages such as “fix” or “changes”
• Keeping the repository only on one computer or solely in a service provider’s account
• Checking in very large files such as videos or database dumps

What businesses should look out for

For every software project, clarify where the source code is stored and who has access: it should be in a repository belonging to your company, or with permanent access for you. Nano Concepts works with Git throughout, uses branches, code reviews and automated checks, and on request gives you access to the repository.