Last year I made the mistake of being over ambitious with my entry in to the GitHub GameOff competition.

Learning from my mistakes, and showing growth as a person and a developer, this year my submission to GameOff is over ambitious. Again.

Also last year I know I lost a lot of votes because of long loading times. So I’ve done that again too.

But there’s one clear difference, last year I made myself sick trying to hit the deadline, I did not do that this year. I was ill even before the jam started having come down with the seasonal plague.

The theme of the game was cliche, and I wanted to make a cyberpunk style game, so Clichepunk 2077 it was.

I had planned to make more quests, but my progress was too slow. I also ran into some framerate issues which stemmed from the rather poor job I did of importing the geometry into the project. Again, time was the enemy. I could solve it, but not in the alloted window.

Aside from these niggles there are some interesting aspects to what I’ve tried to do here. Some of it was new to me, and some I think is genuinely novel.

I thought it would be cool to make the player their characters own voice actor. This came together very quickly, but throughout development I found myself repeating phrases over and over again until the speech recognition could work out what I was saying. So pretty late in development some additional work on the heuristic after an aha moment in the bathroom led to what we have now - a system that actually works pretty well, at least technically.

I’m not sure what I think about it from a gameplay perspective, it’s interesting, but is it good? I’ll let the reviews be the judge of that.

Previously I wanted to write a comic book game, and I realised very early that I had no idea how to put together a quest system. I knew going into this that I would have to figure out how to do this, but that whatever I came up with would be a learning process and not something I was proud of… It was going to be my first stab at such a thing! And that’s exactly what it is.

Actually it works fairly well. But I ended up putting a few global hacks in for some quests. My key learning here was that a quest needs to be driven by its own code - even if it derives from some common interface with reusable event handlers. I know more about how to approach this for next time.

I also experimented with blend shapes which enabled lip sync and facial emotions for the first time. I had to learn about visemes (mouth shapes) and actually I think I did a fairly good job of it - it all came together pretty quickly (in about 5 hours work). My key difficulty was with timing the visemes to the audio. In the future I will look into making a timing editor for speech and having the avatars read the additional meta data that would necessitate, there are some existing standards for this so maybe a bit of reading up will shed light on existing tools that might be useful.

The city was an asset store grab, and it did not import well for me. I should have taken the time to atlas it. Technically I write this with enough time left to do so, but I want to get healthy. This years entry is now done for me.

I know it wont win as there’s not enough content and gameplay and the fps, loading times and browser support all work against me, but I hope it is entertaining enough to bring a few smiles and for people to enjoy what is there.