Terraria taught us building is fighting. Then 3D top-down said: that's expensive.
Hello everyone, I’m one of the developers of Emberhaven (a top-down survival-building sandbox game). We're a small indie team of four. I'd like to share some of the experiences we've had during development with you all. This is our first post in this community.—English isn’t our native language, so please forgive any grammatical errors.
Since we're developing a survival-building sandbox game, it only makes sense to talk about building in this first post.
As one of our favorite sandbox games, Terraria has an unspoken design philosophy that every player can feel: building isn’t just decoration, nor is it merely preparation for combat—building is an integral part of combat itself. Through building, players can create their own battlefields anytime, anywhere.
We're absolutely obsessed with this concept and wanted to bring it to our game. This idea took us through three full iterations...
Since we're developing a 3D top-down sandbox game, we didn’t initially realize that the change in perspective actually represented a shift in the entire experience. We still followed the building approach we were familiar with from top-down games, using 3D models to represent various components (floors, walls, roofs, and so on), which we then placed manually. You had to choose the location, rotate the angle, adjust the height (though we did implement various programmatic adjustments), and then piece them together one by one.
The implementation went smoothly, and we quickly finished implementing the feature. It looked almost identical to the mature building systems found in most top-down 3D games, and we were soon able to build a decent-looking house in our own test environment (though, of course, the variety of building materials was limited). However, we soon discovered a problem: First, our terrain was uneven (due to another gameplay design requirement), and we allowed players to modify the terrain themselves. From a top-down perspective, aligning two sections of wall at the same height was an absolute nightmare. Furthermore, stopping to open the building menu and fine-tune a model's rotation while being chased by monsters was practically a death sentence—which was very un-Terraria-like.
We held a team meeting, beat up a certain programmer, and then completely overhauled the approach. This time, we went fully automatic—just click anywhere, and the system handles the walls, floors, roofs, and multi-layer stacking. On the day the build was released, we all praised the programmer and started testing it. At first, it felt great—we could quickly build walls and tall structures just by running around and clicking. It looked very much like a top-down Terraria. But after just half an hour, we all noticed a problem: Since we’re using a top-down perspective, building two or three stories not only reduces maneuvering space but also introduces unnecessary construction interference. For example, when you want to build quickly, how does the system determine whether you’re building upward or horizontally? This inevitably forces you to take an extra step. Furthermore, Terraria is suited for building in the four directions—up, down, left, and right—because it’s a 2D side-scroller. Even if we were to support only automatic upward building, it would undoubtedly increase the complexity from a single dimension to multiple layers (with each additional floor adding four new directions).
We held another meeting... and then we decided to stop automatic upward generation. Now, horizontal expansion is instantaneous and highly automated. On flat ground, a single click places a wall. No rotation, no height calculations—just a "bang," and it's in place. By rapidly clicking your mouse and moving your character, you can place a continuous string of walls.
Of course, you can still build tall structures, but that’s something you’d do when you’re safe—for base planning, not during combat—just like planning your home in Terraria.
Furthermore, we've refined our monster AI and other related designs, so horizontal expansion in city-building is a highly cost-effective choice for both combat and resource gathering, while building upward is often driven solely by aesthetic preferences.
We've now largely achieved what we set out to do, and we’re pleased that we’ve stayed true to our design philosophy: the functionality of building comes first—it serves survival, exploration, and combat, rather than merely creating a beautiful home. I believe this is one of the core features we want to preserve.
If you have any ideas, please let us know. We’ll respond thoughtfully and compile them for discussion at our team meetings—this is crucial for an indie game team.