Revisiting old ideas with new tech

#programming #gamedevelopment #ai

In which we breathe new life into old ideas, with some friendly help.

I had an idea.

You can read about it here and here and here.

The tech stack I was working with had some limitations. I started with RubyMotion (because I :heart: Ruby, and it compiles cross-platform) with a Firebase realtime backend (because I wanted pubsub stuff in a database). The problems were that:

• any system-based stuff (location, etc.) had to be plugged into RubyMotion, which had mixed success
• the data had to be no-sql, and I was getting the heebies from that
• any server-side logic (for gameplay, and especially for Voronoi calculations) had to be in Javascript, which was a major let-down

So, while I got a working prototype from it, I was reaching the limit of my patience working with that stack.

The first weak point I wanted to eliminate was Firebase. While it's still a pretty cool tool, the second I poked around in Supabase I was sold. We're talking:

• normalized SQL, ahhhh relief
• really solid pubsub
• and the most best discovery, PostGIS with built-in Voronoi.

As a service, I couldn't really ask for more for this project.

However, recently at work we've been doing a lot of research into AI tools. Claude proved itself quite useful, and in my efforts to understand it's limitations better, I thought I would put it through a personal project in the holiday off-time.

Well, color me impressed.

After an initial stumbling period (the XCode project still isn't set up properly, but heck if I'm touching it now), things came together quite quickly. It knows how to make iOS apps. It knows Supabase (at least the 2024 version of Supabase, which is close enough). I was able to recreate the months of work using the original stack in about 20 hours.

So now I've had some opportunities to playtest, but really need more. But that's an issue for another post.

Screenshots