Devlog #02: Blocking Out an Island When You've Never Made a Game

This is Devlog #02 of building Shrines of Mythora. In the last entry I covered where the game came from and how personalisation works. This one is about the first wall I ran straight into: building the actual island.

I'd never blocked out a level in my life. I opened an empty scene, looked at it, and just… stared. For longer than I'll admit.

What the island has to do

Before I could build anything, I had to be honest about what the space needs to do. So I wrote it down as a brief rather than a vibe:

• The player arrives on the shore — a beach, washing up at the edge of the world. I'm not married to it, but that's the image in my head.
• The first thing they see is the courtyard. (Probably a bad name — it's really the home base.) It's where the player always returns, and once all three shrines are complete, it activates and reveals a floating island holding the final puzzle.
• I want the line of sight from the shore to land on the courtyard and nothing else. One clear focal point to pull the player forward.
• From there, they're free to explore out to the three shrines: the cave, the forest, and the mountains.
• For performance, once the player is inside a region I want everything out of line of sight unloaded.
• And the courtyard should overlook the sea — because that's where the floating island rises from once everything's activated.

That's a lot of intent for someone who didn't know how to place a single cube.

The art kit

On the art side I'm working with Synty low-poly packs, which let me focus on layout without drowning in modelling:

• Meadow Forest as the island base
• Tropical Jungle for the beach and the edges
• Enchanted Forest for the forest shrine
• Alpine Mountain for the mountain shrine
• A handful of dungeon packs for the cave

It's a fantasy island, so I'm not too worried about biomes clashing — but keeping each shrine close enough for a no-combat exploration game means I'll have to blend them in a small footprint. More on that fight in a later devlog.

Why I was actually stuck

Here's the real reason I froze, and it took me a while to see it.

I'm a developer. So I treated the island as a pure art problem — something to be made beautiful, first try. And I've always struggled to look at something ugly and half-finished and trust that it'll get there. I want it to look right immediately, which is exactly the wrong instinct for a work-in-progress.

The frustrating part is I know better. When I build websites, you don't start with brand and imagery. You start with content and wireframes, you test the journeys and the navigation, and only then do you layer on the design. I knew the same logic applied here. I just couldn't feel my way past the blank scene.

The perspective shift

So I asked Claude for help, fully expecting it to be useless for something this visual.

It wasn't. It reframed the whole thing: this isn't an art problem, it's a spatial one. It started from a completely different place — roughly how long do you want the player walking for? Turn that into meters. Convert the meters into Unity units. Now you know how big things should actually be.

That was the unlock. Suddenly the island wasn't a painting I had to get perfect. It was a set of distances and sightlines I could lay out with grey boxes. A navigation problem. That I know how to do.

ProBuilder, and it clicked

I installed ProBuilder and started blocking out the island against Claude's sizing.

It came together fast. Within about an hour I had a grey-box island I could actually walk around and feel — the right scale, the right rough distances. After weeks of avoiding it, an hour of real progress felt enormous.

← →

/ 03

Once I had a workflow, I could finally see the vision for where this could go.

Layering the big moments

From there I started adding the larger focal points — some mountains, a couple of waterfalls. Nothing detailed, just the big shapes the player's eye should catch.

I had my wife walk the sections to get early feedback on how the paths felt. We noted the mountain path ran a little long, and immediately started brainstorming things to add to fill it.

Then I stopped myself. "No." Flesh out the rest first. We were about to design a fix for a problem we might not even have once the full layout exists. Solve the real thing, not the imagined one.

Documenting the build

The last thing I did was set up a screenshot system: cameras placed at key spots around the island, and a small script that cycles through every camera and captures a shot. The plan is to document the same angles all the way through — blockout, then dressing, then polish — so you can watch each part of the island evolve from grey cubes into (hopefully) a beautiful low-poly scene.

You'll see those evolution shots build up across the next few devlogs.

What this taught me

A few honest takeaways:

• Sometimes a single perspective change on a problem reveals a dozen solutions that were hidden a second ago.
• My procrastination wasn't laziness — it was "I don't know how," dressed up as avoidance.
• The moment I found the right framing, I couldn't stop. Layout, space, navigation, player sightlines — it went from the thing I dreaded to the thing I couldn't put down. Genuinely addictive.

In Devlog #03, I start turning those grey boxes into an actual place. See you there.