Simon van Luxemburg
The Final Print
"You're a journalist investigating a suspicious case during the Harlem riots. Take pictures of evidence, gather information by interrogating people, and put it all together in an article to show people the truth."
For this project's concepting phase, I was initially brought on-board for helping the team find fun gameplay concepts to further develop. However, since we were leaning towards a game with a lot of dialogue, and because one of my prototypes featured a fairly robust custom conversation manager, the team asked me to further develop this feature for them.
Prototypes
For the initial prototypes, we had the constraints of an adventure game in a suspenseful noir setting. In the first few weeks, I collaborated to do research into suspense in games and other media such as writing and film, to see what can evoke this feeling in the audience.
We found that this could be achieved by giving or withholding information, so I made a small prototype where the audience can hear a bomb before the characters do, and then need to find the person who set the bomb off.
The first prototype I made, aimed at creating a conversation manager that can change dialogue and game events based to create suspense
This prototype was made in Unity for the sake of speed, which paid off, since we had this whole prototype including dialogue ready in just a few hours.
For our next prototype, we needed to gamify this a bit more, so we combined it with another strike team's findings of having information be available for a limited time in a limited space.
My second prototype had a day-might cycle, took place in the scope of a single street, and had a multitude of NPCs who would move to different locations based on the current hour.
The second prototype I made, testing out a day-night cycle where NPC's would move around say different things depending on the daytime
This prototype took another afternoon, and the speed helped quickly answer questions about the feasibility of a day-night cycle, and how large the scope of our level could be. But the team continued work in Unreal Engine 5. And since we knew dialogue would be a prominent feature, I was asked to focus my efforts on translating my custom conversation manager to Unreal.
Conversation Manager
So, for the second half of the short concepting phase, I looked into how I could translate my custom conversation manager to Unreal Engine. Specifically which of Unreal's assets were comparable to Unity's.
My prototype translating the Unity system to Unreal after learning its components
I learned about Unreal's Interfaces, Structures, Data Assets, Data Tables, Enumerations, and got a lot more thorough practice with blueprinting, Components, Widgets, and manager objects the same way I did in Unity. For example, the dialogue was managed by a ConversationManager that got inputs from the PlayerCharacter and could then talk to a HUDManager and MainManager.
The result came very close to my Unity prototype, as well as two Unreal plugins that were researched, so the team was very content with having a custom dialogue system to work with.

Afterwards, I would leave the project. So, I documented how my code worked, what could be improved, and what was yet missing.
I also documented my findings on the researched plugins, so that the team could quickly pick up my system once they needed to work with it, wanted a comparison, or needed to switch to using a game-ready store asset. And they were very thankful for this.
Documentation I made on the custom Unreal conversation manager, to explain to future developers how it works and what should or could be reworked
Once the project was publicly released a year later, I reached out to the development team to congratulate them, and asked out of curiosity how much of my initial dialogue system had survived development. And surprisingly, they said it was almost exactly the same system!
Of course, UI designers and programmers touched up the finer details, and a level designer eventually designed the conversation tables themselves, but overall, it served as a good enough system that didn't need to be replaced with something like a store asset.

