The AI is perfect. Onto the next task on the whiteboard.

Design that never was

My original design included finding more objects such as batteries and oxygen tanks to make a truly crammed room, but were cut due to time restraints. Also cut were plans of having the player put batteries and oxygen in generators that depleted over time. I fell in love with my little humans after I made them out of 100% pure panic, so I decided to just keep the focus on them. It also meant I could summarize the game-play in a single sentence: "Save the stupid humans" and not be forced to have all kinds of UI elements telling you what to collect.

Audio

All my audio was super rushed last second things like recording my desk drawer for the doors and pitch-shifting it until it sounded closer to what I wanted.

For the BGM I had a bassline/lead track that weren't adding much so I ended up stripping it down to a short drumbeat.

Graphics

Quick and simple botch jobs. Simple textures and rapid fire unwrapping just to make things not look stretched or untextured. I started working on a texture for the player character, but this was quickly cut as I wasn't sure the time properly unwrapping the model and painting something would be worth it as by then the lighting was in place and it looked "fine". All the objects were then given primitive colliders in Unity as to not glitch out the physics engine.

Lighting was incredibly lazy and consisted of point lights and a simple gobo/cookie for the spinning alarm light. I fell for the classic trap of making it potentially too dark in the attempt to create mood, and I probably should have made it a little brighter than it was even though it was displaying fine on my main monitor.

I created foreground and background versions of the doorways and walls so that the player wouldn't lose sight of the gameplay while still being able to keep the camera at a reasonably low angle. I'm quite happy with this subtle effect that saved me from having to write a dynamic system while still being able to keep some tall walls for that claustrophobic feel.

20/20 Hindsight

Getting 8 hours of sleep both nights was a solid decision as I was working at very high efficiency almost the entire time with very few mistakes due to fatigue. I did notice my habits were getting worse the closer to the time limit I got though. If you look at the code you'll notice that the later parts have less and less comments and other bad practices in place..

I got very used to the controls during the hours of making the game so I may have under-estimated the difficulty picking up the controls for a new player. I'd keep the tank controls just as they are as I think they work as intended, and there wouldn't be much "game" left if the controls were too easy. What I would do is modify the player collider and reduce wall friction to keep the player from getting "hooked" on surfaces and rotated out of sorts, removing a significant amount of frustration. I'd perhaps also increase the mass of the player to ease movement slightly.

As polish I'd add a condition to end the game early if there were no survivors outside of the safe-room as I noticed this was a common issue, but never got around to implementing a fix for it since it wasn't game-breaking. Graphics and Audio would get another pass to bring them up to scratch.

There are a couple of minor glitches in the web version of the game like intersecting physics objects and imperfect audio looping, but having never exported to web before I'm very impressed with how well it worked for something like this.

Conclusion

At the end of the day I'm happy with finishing on time and the process as a whole given the time limit. Though I do feel the end result was a little too safe/vanilla and I wish I'd been a little more experimental in the design stage.

I'm left feeling that I have no reason to do another compo entry in the future as too much was sacrificed for the sake of the strict 48 hour time limit which I don't feel adds anything of value to the end result. Maybe if I aimed smaller and more experimental instead...

However I can absolutely see myself doing a Jam entry with other people with different skill-sets and a more relaxed time limit.

If there are any questions I'd be happy to do my best to answer them in the comments.

https://ldjam.com/events/ludum-dare/42/herding-humans. (web version available)