How Long Does It Usually Take to Build a Simple Physics Engine?
Asked by
MellowCedar8
I've been working on a small physics engine in C++ using SFML for quite a while. So far I've implemented basic translation and I'm now refining the collision system. I'm not a computer science student—I'm studying mechanical engineering and learning a lot of the programming and math as I go. I've already spent more than 20–22 hours on the project, and I'm starting to wonder whether that's an unreasonable amount of time for something this simple. How long should building a basic physics engine realistically take, especially for someone learning along the way?
Answers
A useful alternative is to use an existing physics library when you want to focus on making a game or simulation. Building your own engine is a great learning project, but you don’t need to finish a complete general-purpose engine before you can create something interesting.
Reply to this answer
Twenty-two hours is practically nothing for a project like this, especially since you’re learning the underlying concepts instead of just copying an implementation. Collision detection alone can take weeks depending on how far you go. Keep at it—building a first engine is mainly about developing intuition for vectors, transforms, and simulation.
Reply to this answer
There isn’t one fixed answer because “physics engine” can mean anything from applying gravity to simulating complex three-dimensional worlds. A falling box might take an hour, while a basic billiards simulation could take several days. Add requirements like accuracy, performance, friction, air resistance, large numbers of objects, or real-time operation, and the project can grow dramatically.
Reply to this answer
The amount of time depends entirely on the result you want. A solo developer might create the physics needed for a small game in a day, while professional studios can have entire teams improving physics technology for years. If your goal is learning, the time spent is part of the value—not a sign that you’re doing poorly.
Reply to this answer
Even a two-dimensional physics system can become a substantial project. You have to deal with collision detection, collision response, numerical stability, and edge cases. It’s fine to start with a narrow goal, such as circles and rectangles with gravity, then add features one at a time.
Reply to this answer
Add your answer
Answers are reviewed before publication.
Related Discussions