Code Hosting

Every program consists of text that a human has written: line by line, instructions for the computer. This text is called source code. Code hosting means that this source code doesn’t just sit on your own laptop, but additionally on a computer on the internet. There, several people can work on it at the same time without sending each other files by email. The service also remembers every change: who made it, when, and why. Major providers of such services are GitHub, GitLab, and Bitbucket.

Why software is rarely created alone today

A modern program often has millions of lines of code. Dozens or hundreds of people work on it, spread across different countries and time zones. Without a shared place for the code, this would end in chaos. Two people would change the same file differently, and no one would know which version is the correct one.

Code hosting solves exactly this problem. There is one binding main version, and every change is documented in a traceable way. If a program crashes after an update, the team can review the recent changes and find the bug. In an emergency, the code can be reset to last week’s state.

Economically, this has meanwhile become its own industry. Source code is a software company’s most valuable asset of all. Whoever hosts it sits at a very central point. This explains why Microsoft paid billions for GitHub, even though the platform itself barely made a profit.

Repositories, commits, and pull requests

A project lives on such a platform in a so-called repository, or repo for short. That’s a folder with all the files plus the complete history of changes. When someone saves a change, a commit is created: a snapshot with a date, name, and short note. The series of all commits is the project’s history, and you can jump back to any point at any time.

So that people don’t overwrite each other, work is done in branches. A branch is a copy of the project in which you can freely experiment. Once the new feature is finished, you open a pull request: the request to merge the branch into the main version. Others read through the code beforehand and comment on it. This mutual review is called code review.

The technology behind this is usually Git, a program from 2005 for version control. It’s important to distinguish: Git runs on your own computer, GitHub is just the service on the network. So you can use Git without a platform. The platform contributes additional things, such as an issue list, discussions, and automatic tests for every change.

From open source to AI assistants

Almost all free software resides publicly on such platforms. The browser Firefox, the operating system Linux, and the tools behind nearly every AI application can be viewed there. Anyone is allowed to look inside and suggest improvements. For job applications in IT, a public profile with your own projects is often worth more than a certificate.

Code hosting platforms show up in the news for two reasons. First, in security incidents: if a company accidentally leaves passwords in public code, attackers find them within minutes. Second, in connection with artificial intelligence. Programming aids like GitHub Copilot were trained on huge amounts of public code, which sparked disputes over copyright.

A common misconception is that public code is automatically free to use. Visible does not mean rights-free. Every project has a license that precisely specifies what is allowed. Some licenses require that your own extensions also be made public.