The Thesis Is in Another Castle

I'm counting my victories, cutting my losses, and moving on with this project.

In summary, I wanted to create a new tool for procedural dialogue that rewrites lines of text automatically. The blueprints for such a tool still exist in the previous devlogs, and it's possible that I or someone else will try to pick up where I'm leaving off. The only blocker preventing me from continuing onward is time. I am 1/3 of the way through completion of the tool and 2/3 of the way through the time I had set aside.

I had hoped to make a tool that would help in the creation of a thesis project for my master's program in game development. However, I have been advised that my fascination with procedural narrative may be better fit for a dissertation when it comes to scope. So I have to tell myself that my thesis is in another castle.

This is the postmortem for "The Bitwise Collection" as a PROCJAM 2020 entry.

What Went Well

• In creating a piece of new technology, I started with the intended experience. This helped guide all technical design decisions toward the end I envisioned.
• I attempted to iterate on a previous success. I learned how to make a procedural dialogue system tick with You Teach! Romeo and Juliet before I tried generalizing such a system to a multi-purpose tool.
• When scope became a clear issue, I stripped as much as possible out of the project.
• I chose a good name for the project. The double entendre in Bitwise helped communicate the vision to others who were interested.

What Went Wrong

• It should have been obvious that my original plans would not fit the time allotted when I spent the first week or so journaling 40+ pages of tech design. The tech itself was originally a mammoth project, let alone the planned collection of demos to accompany it.
• I spent a long time trying to solve problems that other people had already solved. This happened both in the design phase and in the implementation phase.
• In reaching for the stars I promised more than I knew deep down that I could deliver.

What I Learned

• I had the opportunity to reflect on something I want to see in games: open conversational worlds.
• The (deceptively difficult) problem of programmatically diagramming sentences into their underlying structure has been eloquently solved by Temperly, Sleator and Lafferty in their program, Link Grammar, which is free for use.
• As a technical designer, time & energy resources are real concerns that can make or break a project.
• As a technical designer pioneering uncharted software territory, there is a trade-off between thinking outside the box, and conforming to existing paradigms. Flexibility vs. comfort.
• I learned when to call it quits and cancel my own project.