Git, GitHub, and Version Control

One of the first topics encountered by individuals trying to break into programming is the idea of version control. After all, being a professional developer requires using a version control system, so it makes logical sense to introduce this early on. However, it is a tough topic for a budding programmer to cut their teeth on. The concepts necessary to understand version control (and in particular git) are tough enough on their own, but it becomes doubly as confusing and intimidating when wrapped with another entirely new topic - i.e., programming.

Version control, however, is nothing more than a systematic way of storing and retrieving different versions of a project. The project can be a Word document, PowerPoint file, or - in its most common usage - a code base. If you have ever had to save version 1, version 2, version 3, etc., of a file, then you have used a form of version control. At it core, what professional version control software does is automate that process, make it less error-prone, and - as we'll see with git / GitHub - help us collaborate with others on the same file(s).

In fact, if you have ever worked in an office environment, the workflow described above, and its associated pain, is likely very familiar to you. Let's use that familiarity as an analogy to understand some key concepts of one of the more popular version control systems - git. Git is very customizable, but there are certain conventions that are used by developers to make it easy for folks to orient themselves. In the examples below, I will be using these conventions.


In version control land, think of a repository as a folder. This folder is going to store all the files and folders related to a single project.


Think of a branch as another folder within your main project folder - i.e., your repository. In every repository, there is a special branch called "master". This branch holds the project in a state that is always production ready. If your VP came along and asked you for the latest PowerPoint presentation and all the related files - i.e., all the Excel files used to build the charts in the PowerPoint - the master files are what you would give him. All the numbers / charts will match, and there will be no spelling errors.

So what good are branches? Well, you likely have the need to update all your PowerPoint / Excel presentation at some point. Maybe new numbers are out, or maybe you want to add a new chart. Instead of polluting your master branch (you never know when the VP will be back demanding the materials again), you create a new branch / folder and make your changes on that branch. If someone needs the documents, but you haven't finished updating your files, you just send them the master branch.


So you now have two branches. One - the master branch - is there to make sure that if anyone asks you for the documents, you have a version that will make you look good. The other branch, where you are adding a new chart, is your feature branch. The new chart is going to take several days to build while you grab the information, cleanse the data, and figure out the best way to display it. At each stage of that process (say, when the data is clean), you save your progress by committing your changes. Basically, you are hitting CTRL + S, but on your entire branch as opposed to just one file.

Unlike simply saving a PowerPoint file, a version control system keeps track of the files at each commit. If you made a mistake, you can go back to an earlier commit, when you know the files were in a desirable state, and continue working from there.This is the feature of a version control system that does away with the need to keep saving version 1, version 2, etc. Not only that, but by using a commit message, you can write down what were the changes made to the files compared to the last commit.


After a few days, the new chart is ready, you add a new slide in your PowerPoint - the new feature is ready. At this point, you are ready to merge your new feature into your master branch. Instead of figuring out what needs to be added where, the version control system takes care of this for you. Just use its command to merge the feature branch into the master branch.

Version control systems are much more powerful than this. We didn't even get into collaborating with others, and explaining concepts / commands such as pull,fetch, fork, or clone. Hopefully, however, this analogy is helpful in understanding version control at a conceptual level, and makes learning about more advanced topics a bit easier.