Previous Post:
First, before we start, I’ll give a quick update on the state of the tool. It has officially reached 1.0 and is released on the Unity Asset Store! I have the trailer for the tool on the actual project page, so I’ll share a more in depth preview of the tool here.
Branching Dialogue 1.0
Postmortem:
What Went Well:
I’m extremely happy with how this all turned out. When I started on this project, I wanted to take my previous branching dialogue tool and make it something even better. As I worked more, I realized that I kept surpassing my previous expectations. What I planned to make, and what I ended up making, were to different levels of polish of the same idea. And this brings me to the first thing that I thought went well, my flexibility.
To address the elephant in the room, Coronavirus messed up everything. I had to pack up and leave my college, go home, and work in an area that I had only relaxed in for the past four years. I make a big effort to keep my work at work, so moving into a mixed working space was a little challenging. But, I got a good system working, and after a week or two my motivation had returned to normal. In terms of the project, I was flexible as well. As I worked on this tool, I realized that my original plans weren’t complete or they weren’t possible given the architecture I had set up. I had to make quick decisions on what to keep, cut, and add when that happened. But since I acted decisively, ghosts of unfinished features never came back to bite me.
The second thing that went well is that I released the tool! After a two week process it was approved, and now that the tool is registered, pushing changes becomes a 24 hour task. It’s sitting at $4.99 on the Unity Asset Store, and it’s my first tool, well first anything, that I’m releasing for money. Because I went into this with the idea of releasing, I had a high standard of polish for myself. That helped me push through long nights to get certain aspects “just right” to where I was happy with it.
The third, but certainly not last thing that went well in this project was my node creation pipeline. Right after making my first few nodes, I realized that I needed to find a better system to make them. Previously, they were a mess of dependencies and data all over the place, so I decided to spend some time organizing it. Within the week, I had a system that was tagged in code, so any time I wanted to make a new node, I just searched for the ADD NODE HERE comment and filled out the new node info there. Now, it only takes me 5 minutes to make a node from start to seeing it working in a tree.
There’s a ton more that went well, including my debugging system, the variables, and blog posts, but I’ll leave it at three.
What Went Poorly:
Since I wanted to release this tool, I ended up getting bogged down in the details. I spent a lot of time on stuff like font colors, spacing of words, and QoL features for edge cases when some major features weren’t in yet. Personally, it worked out well because it fit my style of working, but it could have gotten out of control if I wasn’t aware of it.
Also, it was nice that I was flexible with my work, but it did mean I had to change around my milestone definitions a lot. As I said above, I realized that my original plans weren’t complete or they weren’t possible given the architecture I had set up. So my schedule was messed up as I tried to compensate for my lack of detail planning.
For the last thing that went poorly, I’ll bring up the lack of unique visual direction in this tool. It’s very similar to other assets visually, and it doesn’t have anything unique to stand out. I’m even using the built in skins for my nodes. While it’s certainly not a cookie cutter tool, I doubt many people would see a screenshot and know it’s my tool.
