Sorry for any error in this post but English is not my first language. I'm trying to express and share my experience as best as I could. If you'd like to read the original version, please visit my blog (Chinese)

4/24/2017, 16:23.

I get out of the bed, with my neck and left shoulder hurting like hell. The vomiting has gone. It seems that the medicine I took is taking effect.

Five hours ago, I just won the battle against myself.

I knew the Ludum Dare last year. At the time seeing this competition, I was like striken by a thunderbolt, and decided to be a part of this next time when the event is held. So in Ludum Dare 38, I choose to participate in the Compo group. As a developer for about 10 years, I'm quite new to the game developing: I have been working in game companies for only a little over 2 years. I need this to know where my limit is and how capable I am. And, isn't it awesome to build a game in just 48 hours?

With these thoughts in mind, I started my FIRST Ludum Dare.

DAY 1, 4/22/2017

Working on the theme

The event started at 9:00 A.M. on Saturday in China. I didn't woke up until 9:40, because I needed enough energy to cope with the super tight schedule.

At the time I saw the theme, A Small World, I lost all the clues, thoughts and imagination immediately. The theme seemed not attractive at all: everyone today talks about how the internet makes the earth a small world; the earth is just a tiny tiny world in the whole universe; blah, blah, blah. Yeah, we have heard this all. How's this theme gonna be innovative.

I smoked for like 2 cigarettes, cooled my head down, and tried to interpret this phrase from another angle of view when my cat came and meowed at me. Looking at her pretty face, I had something in mind at once. Yes, my home is the whole world of my cat 'Doggie', and it is small! A Small World!

So I made this plan up as soon as I could: My cat, 'Doggie', is trying to be the ruler of the world. Not by violence but collecting all treasure in the world and hiding them in her treasury. The stupid Human, which probably is me, will try to prevent Doggie from doing this. It's a story about the cat queen defeats the stupid human.

Prototype Making

I designed the game to be an RTT game with 45° top-down view, like Commandos, Desperados and Shadow Tactics. But my game is kind of different: the map is quite small (just with the size of my apartment); there is only one enemy (me); the cat can move vertically frequently in the scene due to her extraordinary jumping ability.

I spent about two hours working out character control and NavMesh setup. I rarely used the Unity's NavMesh before. For the projects in the company, my team chose the A* Project, a super powerful 3rd-party navigation library. But I thought it was too much for this small game. Anyway, LD is about learning new things and challenging myself, so I decided to learn Unity's NavMesh.

After the navigation was done, I implemented the enemy's modules of sight sweeping, target searching, state switching, and the cat's module of collection treasure and hiding it in treasury.

By then the game had already with a rough shape. I planned to put the detail adjustments in the final few hours. The clock read the time for dinner, and I had worked for about 7 hours. The progress looked quite optimistic. I put a short video of the prototype on YouTube, and blogged for my progress.

After dinner, I started to put some bits and pieces into my game, such as the tooltips with mouse hovering on interactive objects, jump controlling on OffMesh points, etc. I tested and refactored my code as well.

Back to Reality

Everything looked perfectly until I met the first real trouble caused by myself.

In the very beginning of my design, the enemy's behavior was rather simple: patrolling, looking for the cat, staring, and chasing. It was a linear process, and several if ... else could get it done. However, when I tried to add a new behavior, "listen to the cat's meowing and go investigate", the old 'system' was corrupted and unable to develop at all, as the new behavior was too complicated for it. Even more, the old AI system was neither flexible nor extensible. I had to admit my design was like a pile of sh*t.

Struggling with all kinds of thought and fatigue, I decided to re-implement the whole AI system before going to sleep.

I decided to use FSM for enemy's AI. By separating the decision-making and action-control, the logics became much more clear and controllable. The improvement in state managing and back tracing was satisfying as well.

The implementation of the new AI system finished at dawn, around 5:00 A.M. The sun was rising. I wrote down a blog for my first day's progress, set my alarm, and lost my consciousness as soon as hitting the bed.

---

(To Be Continued) Part 2

And here is my game; I do hope you can try it out and leave me some comments and thoughts! https://ldjam.com/events/ludum-dare/38/this-world-of-mine