Post-Mortem: Final Version for GDEX Game Jam!

Hi! Welcome to my post-mortem for Herbal Tea Simulator for GDEX Game Jam 2020.

What went Right

Timeline

There was a lot of time for this jam, ~2 weeks, which meant that I had lots of time for procrastinating. Usually I start something in a 48-hour jam and get exactly far enough to realize I'm not in love with the concept before giving up. This time I could give up more than once and still have time to finish the project! The "engine" (subtitles queue and clickable choice gameobjects) was built in the first couple of days and I had built and animated an initial example of how I wanted the word art to work before running into writer's block for several days. Then once the first chapter of the story was in, life got in the way a bit and a bunch of time went by before I got back into it and started turning the text into interesting things. That still left me several days to sit on it before polishing and publishing. This game probably has 6 days of work and 10 days of thinking, and *would not* have been finished if it were just a week long.

Jam Friends

The GDEX Jam had a discord that had a lot of great people in it. They were all awesome, and it was a great outlet for random musings and questions for when I needed a distraction while I noodled a problem.

Scope

Related to the timeline, the game is made in Unity and uses DOTween, Simple Helvetica plugin, and Ink (which I'd only ever dabbled in before), but everything else needed to be built from scratch. I scope-limited hard from the get-go, but still "Chapter 1" had to be retrofitted to become the full game when I realized that taking the story further than what's already there was going to take longer than I had. I knew I didn't want to give myself too much work with beautifying choice text, and once I did that work (everything in the current game) it became clear that every chapter I added was going to be an extra 2-3 days of work. Luckily Ink is a very easy-to-iterate dialogue system, so adding a little "final boss" dialogue  and putting in some endings was quick and easy.

Tech

All the tech worked really smoothly, actually, which I think was mainly doable because Ink separates things so well into choices, tags, and responses. Choices got attached to a graphical prefab based on the text, and to an existing anchor gameobject in the scene via tag, which made it easy to prototype via a default text gameobject and then also to tweak that into the final product by just creating a prefab to get found and used automagically. 

There was a bit of "fun with delegates" to help give different choices different behaviors while retaining basic interactivity. So when you click something the default is to just respond to it, but you can make a new controller for that prefab and as long as it sets the delegate appropriately it'll just work. If I had a little more time to plan I probably would have turned that into an interface, but copy/pasting boilerplate code worked fine for the scope of this project.

Unity Plugins

One thing I'm really glad I found was the Simple Helvetica unity plugin, which as far as I can tell is just a set of 3d models of text glyhps and a script that puts them into the scene for you when you type a word into the editor. That made building the word art pretty much trivial, and added some constraints that prevented me from going overboard with any of it. The Television is probably my favorite "thing represented by text", the computer is my least favorite asset in the game, ugh, it's terrible, and the teacup with animated "tea" is what convinced me (on day 2) that I really was going to make a game called Herbal Tea Simulator.

Also, I can't say enough about DOTween, so I will say almost nothing: I am a huge fan. DOTween is the Frank's Red Hot Sauce of Unity plugins.

Music

I've worked with Javier before, and I always like the stuff he puts together. If I had gotten the story and music beats wrapped a little earlier I bet he would have been able to tailor things a little more closely to the narrative, but I think he did a great job for the poor direction I was able to give at the time "You're going crazy, and the music gets crazy and discordant when you do so".

Art

Not having any art was a great choice. I highly recommend it for all games.

What went wrong

Ink Tags

Apparently you're not supposed to want to tag your choices, because you can't do that in Ink for purely lexical reasons. It seems to me you could fix it by just adding a "##" char sequence to indicate the end of tags, but they've thought about it more than I have so maybe it's more complicated than that). I ended up writing a little parser and separated tags from choice text via "((". A little clunky, but it worked with the jam timeline.

2019 Macbook Keyboard

This keyboard is terrible. It is so bad. Writing code was awful, writing this post-mortem has been an exercise in industrialized rage manufacturing.

Sound

There's only one sound effect in the game, but I really wish I could have put more audio in. It's an often-overlooked thing that I usually take great care to not overlook, but bleah, there it is(n't).

Story

I'm a writer like I'm a poker player: I am maintaining my amateur status so I can do it in the olympics. No but really, I had grand plans for the story involving several relationships and a whole social media commentary thing, and that all seemed very reasonable right up until I started writing it. I can write just well enough to fool myself into thinking that the story will just fall into place and be amazing, and it is never, ever that easy. Anyway, there's a story, it just happens to be a very long-form joke with some tragedy bolted onto it. /shrug

Conclusion

Jams are good! Do them! Also do post-mortems, this has been very cathartic.

Thanks for reading!

-steve