Every sailor remembers the first time they felt the weight of a boat moving toward a concrete pier. Was there a specific moment where you realized a textbook wasn't enough?
I remember it vividly. I had been sailing for two days with an instructor and was terrified because the exam for maneuvers was coming up. My brain couldn't map it out as much as I wanted to feel comfortable. I reverted to sketching on paper and spending hours processing maneuvers, but I had no interactive way to test my understanding. That first time doing docking, trying to remember the order of commands—it was the most stressful experience.
You chose a top-down 2D perspective rather than a hyper-realistic 3D environment. Why prioritize simplicity over realism?
In instructional design, there's a concept of 'friction.' Simulators that replicate mental friction can successfully help someone learn a specific process. Many 3D simulators are so rich in detail that they actually take attention away from the practice. My goal wasn't to create a beautiful 3D world—the 3D version of the experience is the actual sailing itself. This 2D simulator is a tool to isolate specific concepts and maneuvers, not to replace the experience.
How did you approach the physics of inertia and steering when the user is just moving a slider on a screen rather than a physical rudder?
It was about effective translation, not full replication. On a phone or computer, you aren't feeling the physical forces of a rudder. I had to ask: how can we replicate the feeling and the friction of it so it feels accurate to a sailor, even though the interface is completely different? I worked with AI agents to build multiple 'work-in-progress' simulators to determine if we were coding the right sensations. It’s not about being perfectly physically accurate; it’s about identifying the relevant forces that make the logic of the boat feel real.
Which parts of the interface proved the most difficult to 'tune' so they felt intuitive?
The steering wheel slider and the controls for the sails needed—and still need—the most work. We also found that some 'accurate' physics just didn't help the learning process. For example, physically accurate rope behavior was nowhere near being useful for explaining how to dock. You have to simplify the simulation to make the lesson clear.
