Team SquareWars Postmortem

It didn’t end up looking like that at all.

Check out the entry here.

What went right:
– We came up with an original gameplay mechanic that is simple, yet fun to play but also to translates well to pc and mobile interfaces.
– We honed on the core of the game almost immediately and stuck with it to the end.
– A quick sketch for the battle field gave us a good starting point of how the game would look.
– The initial was for to use a 2D top-down view for an old-school look. Halfway though, it became apparent that the look would be too simplistic, and was changed over to full 3D. This improved the look of the game and allowed us to add 3D specific features like the camera following the units which gives a better feel of the “action” in the battlefield.
– Unity proved a good tool for rapid game prototyping and production.
– Using a Google docs spreadsheet for managing tasks. Although we tried a bug tracker, a simple task list with multi-user editing is perfect for managing tasks as well as a brain dump for all good (and not so good) ideas and features for the game.
– We had lots of ideas about what the unit AI should do, and spend some time implementing infrastructure for supporting it. Luckily time constrains forced us to simplify, so most of the behaviours of units rely on a randomness, which produced surprisingly good results. In part it was also an experiment to see if players would be able to sense any intelligence behind the unit AI, something like a cheesy Turing test.
– The fact that we always planned to keep going on making this game a success even after LD was over. This helped us push back features that would have taken way too long for the 72 hour limit, otherwise we would have ran out of time. “Will the time spent implementing this be good enough for the increase in fun in the game compared to other features?”

What went wrong:
– The Visuals. Even with a favorable theme like “minimalism” the visual quality of the entries was higher expected. In the end graphics proved to be very important and we spent a fair amount of time improving the look of our game, but programmer art has its limits.
– Although on average turnaround in Unity is fast, there were many cases that things need to be done in a certain way.
This invariably means scouring the internet for a Unity-specific solution and generally wasting time on trivial things better spent working on the game itself.
– Source control. We used TortoiseHG and BitBucket for managing and sharing our assets and code. Although Mercurial is an excellent tool for source control, Unity’s reliance in part to binary scene assets which cannot be merged, file commits that generate unwanted branches, added with the complexity of the TortoiseHG UI become an occasional obstacle to our work-flow.
– Not enough time to actually  implement all that we wanted.
– The UI is not good, it’s functional and does what it’s supposed to do, but I wish we had had more time to actually make it nicer.
– In the end, one of us had made much better explosions (the current ones aren’t that good) but after trying them on the web player it would break the game, units wouldn’t die. A bug that made us go back one revision because just eliminating nice things from the explosions wouldn’t fix the problem. Same thing happened with reflected lasers (they don’t reflect in the final LD version).
– Unity, although great in some aspects, holds too many nice features for the pro version, which is pretty expensive actually. It also means that what you do is tied to their engine and not really portable to anywhere else.

What went weird:
– Shouting “Square Wars” repeatedly at the laptop to record a sample for the title screen. In the end sample was not used. (Best for all concerned!)

Conclusion: 
Overall it was surprising with how much of a game can get done in just a few days. Focusing on the core mechanics was very important, as well
as resisting the urge to implement things the proper way. Prioritizing everything by the how much they will improve the game as it is at that point in time is also critical, as this will prevent you from wasting time developing stuff which may or may not be used in end.

http://www.youtube.com/RpkhuqspuYA

Tags: postportem, unity, unity3d