Building STAR-EX: Part 2 - Fullstack Development

Hey everyone! Following up on my previous post about STAR-EX (link), which introduced the vision and key gameplay modes of STAR-EX, this article delves into the Expedition gameplay system. This mode revolves around resource management, risk assessment, and strategic decision-making, all built on robust architectural principles.

With the support of Amazon-Q, the development process has been streamlined, allowing us to design high-quality systems while iterating quickly. Let’s explore the expedition concept, supported by diagrams and insights into how these systems will evolve.

Quick Video Overview:

// 2-min explanation video expedition
https://www.youtube.com/watch?v=fed4YgzA-_M

// 1-hour workflow with Amazon Q video

https://www.youtube.com/watch?v=AMTbhh10xFA

## What We're Building: Expedition Mode

Now let me explain what we're trying to achieve with Expedition mode. Think of it as your strategic command center in the STAR-EX universe.

The core loop is beautifully simple:

1. Players start with basic resources (ships, food, fuel, ammo) and access to zones

2. Deploy missions to these zones to gather resources

3. Successfully return fleets to claim rewards and use materials to craft more resources, and advance more civilisation, as well as get more rare items to sell into the marketplace (as NFTs) or craft them to advance your fleets. Ideally you need to explore more zones, and explore more mysterious/hidden feature later.

4. Rinse and repeat - but with increasing complexity and strategy! the more you own, you are richer in resources, and once the game scale. you are prepared facing the next challenge and can survive because we will implement some balance to keep the cycle to be interesting. whether you can lose the risk of your fleets, encounter hidden feature. let's see the game later

## The Technical Challenge: Building a War System

One of the biggest challenges was designing the war system that would drive zone control and fleet management. After several iterations and great conversations with Amazon Q, we landed on this architecture:

The system integrates DynamoDB for persistent storage, Redis for caching, and a War Manager service to handle the complex state management. We're using scheduled jobs to trigger wars every 8 hours, making zones contestable and keeping the gameplay dynamic.

## Database Design: Planning for Scale

Speaking of data, here's how we're structuring our database draft relationships (ofc when we scale it next month, it will be more massive and some stuff will change too):

While this is our initial design, it's built to be scalable. We'll likely rework some aspects as we learn more about player behavior and performance requirements.

Currently player will receive basic items -> zones, spaceship, foods, fuels, ammos.
and then player can select zone to begin the mission.

Image not found

Image not found

Deploying the fleets are simple too as long as you have the resources.
Once you ask them to return, they will bring you items for you to craft to the next fuels, next foods, etc.

## My Experience with Amazon Q

This is where it gets interesting! Building complex game systems is challenging, but Amazon Q has been an incredible partner in this journey.

1. Intelligent Code Assistance

// image: tooltip/tips and generate next code

Image not found

Image not found

It's quick feature when you try to do live code and having a senior developer looking over your shoulder, suggesting improvements and catching potential issues before they become problems.
It is indeed like cursor-editor, but Q is free, and can be implement in many editor and perhaps vscode-web.

2. Comprehensive Understanding

Image not found

What sets Q apart is how it ensures mutual understanding. It doesn't just throw solutions at you - it confirms understanding and asks relevant follow-up questions. It is similar like Claude AI-agent i think but it has no the next feature (reference)

3. Well-Referenced Solutions

Image not found

Every suggestion comes with relevant documentation and resources. This has been invaluable for learning AWS best practices while building. Although other AI-agent perplexity.ai also offer reference, but im surprised with Q for focusing to improve my knowledge here.

4. Continuous Conversations

Image not found

Even in complex discussions, Q maintains context and can continue detailed explanations across multiple responses. Seems the LLM provide more tokens? perhaps competing against Gemini Ai?

## What's Next?

I'm currently focused on implementing this system completely, and I'll be updating again the game next month once my current urgent project wraps up. The current goal has been accomplish to have a fully functional Expedition mode where players can:

- Deploy fleets strategically

- Participate in zone wars

- Manage resources across multiple missions

- Interact with the dynamic event system

but on next month, you will see the game integrated with the smart-contract and proper auth. I done without so people can try it seamlessly the sneakpeak gameplay.

So Stay tuned for Part 3, where I'll dive deeper into making the Ai-agent with AWS-bedrock, because STAR-EX will have many recipe of items, and ton of items. So we need to onboard and help players seamlessly to understand quickly their reward, or the game flow or even making the ai-agent to copilot the IDLE game.

Also I'd love to hear your thoughts:

- What other features would you like to see in Expedition mode?

- Suggestions for balancing the 8-hour war cycle?

Drop your comments below