The semester is coming to close, and I’ve completed the functional aspects of the system. All the editor, runtime, saving, and loading scripts are done (save for a few tweaks), and I just need to put together proper test scenes and documentation before I publish on the Unity Asset Store in January. The report is finished too.
In the spirit of reflection, it’s interesting to see how the project has both slowed down and sped up over its lifetime. Development took a long time, and meant that I wasn’t able to accomplish all my goals, but now I’m able to create Components extremely quickly, and demonstrate it live. The development is practically done, even though it took a while. I think that now my goal should be polish. Make it look nice, make it feel nice, and get to ready for public viewing.
As mentioned below, I plan to demonstrate Descant in TAG at some point in the new year. I was talking with Jonathan today and he recommended I showcase the project in an academic showcase/conference coming up in May, which is very exciting. I’ll definitely be applying. Good things are on the horizon, though I definitely deserve a break.
By this point in the project, I’m starting to think about the end of the independent study. I’ve been able to get a little bit of user testing through some students in Jongwoo’s class, and I plan to run a Descant workshop in TAG as well (either in December or in the new year). As for the project itself, it’m getting close to the end! I’ve successfully implemented pretty much every single Component! I just have to iron out a fair number of bugs, add comments to nearly everything, and develop some proper test scenes. I aim to get most of that done in the next week or so.
After that, the remaining tasks for the independent study are to write the report and get the ball rolling for actually publishing the package on the Unity Asset Store. I feel great about the current status of the project and hope to push forward just a bit more to see it to completion. The possibilities that it’s already affording are truly fantastic.
It’s been a little while since I last check in, and a lot of organizational changes, QOL tweaks, and bug fixes have come along. Most importantly, I finally settled on an appropriate package folder structure. I’d been messing around with a 2-folder and 2-folder systems with just as many Assembly Definitions, but I’ve settled now on what I like the best. It allows for Runtime scripts to reference Editor scripts, and allows for them to both reference Components scripts. When the Components will have to start referencing Runtime though, that’ll be a problem. We’ll figure it out as it comes. I also added a bunch of comments to new scripts, converted my Descant Graph files from .desc to .desc.json, and hid the Components scripts for the mean time, as I prepare to push a v0.1.1 build up to GitHub.
The reason for all this organization is the fact that I’ll be demoing Descant in its current state and talking about the future design goals with this semester’s CART 415 class. I had talked with both Jonathan and Jongwoo Kim regarding possibly doing a Descant workshop in their more practice-based classes. The students would benefit from using my system for their game jams (albeit in its simplified state), and I would get feedback regarding usability and my future goals. I was worried about getting it all working and buildable by tomorrow, when I’ll be doing the workshop, but everything is working perfectly now, and I couldn’t be happier!
The next goals after the workshop are to continue developing the Component system by allowing the Components to be saved to the DescantGraphData file. After that, it’s onto run-time implementation, which will be no small feat.
I’ve decided to focus on the node components for the moment. Once I get them down and working, then I can add the Actor and Conversation components if I want to / have time to. I’ve been able to get most of them half-written, I started work on saving them, and I’ve added probably 90% of their UI implementation in the Editor. Now I just need to read the data when saving, load it properly, and get it working in runtime. I’m quite pleased with my progress, though I know that I need to start interviewing / demoing the software soon, to get an idea regarding the UX and feasibility. I’ll want to meet with Jonathan soon too to discuss my progress.
As a sidenote, I also reorganized the project so that it only uses one Assembly Definition. This way I can avoid circular dependancy. This wasn’t an issue before, but now that I have the Components going, it’s come up. The Components use both the Editor and the Runtime, and those both use the Components. So I had to put them all in the same Assembly Definition.
I’ve added a new file to the documentation, detailing all of the ‘components’ that I will be writing for the system. I have node, actor, and full conversation components. Theoretically, these will be able to be added to their respective types in a totally free, totally modular fashion. Presently, I’m just getting a hold of how I’m going to program it. I’ve added a bunch of TODOs, and have started on about half of the node components. Some are invokable, meaning that they trigger some effect when the node they’re attached to is reached, but some are just for data storage, to be accessed by other scripts which might want to make use of their properties (e.g. a TimedChoice component won’t ever trigger anything itself, but its data members will be necessary for the DescantConversationController scripts to reference when updating countdowns and moving to the next node).
I’m feeling really good about what I’ve got done so far, but it’s definitely going to be a lot of work. Maybe more than I thought. I think it might be best to push my prototype date a week or so into the future, giving me time to get these all working. Instead of doing user testing right away, I’ll run a seminar where folks can comment openly, and I can discuss it in a casual manner. Until then, I’ll keep coding away at these components.
The base system has finally been settled! I have the Editor script all done (save for a few QOL or functionality tweaks), the saving works pretty well, and I’m now getting started on the Runtime scripts. While I do that, I’ll be working on the unique aspects of Descant. That’s where the real creativity will start to shine through, and I’ll be journalling a lot more (I hope). Presently I’m going to try writing one of these journals every 2-3 commits. To coincide with my general pattern of milestones. I’ll also be ramping up my explanations in commit messages, going beyond the standard dry descriptions of changes.
I’m nearing the proof-of-concept milestone that I played out for myself originally. I’m feeling good about my progress, though I know that the ‘components’ system that I had dreamed up will take some finagling to get working. I think that my next steps for the project are to make a ‘build’ of the current package, build a super simple runtime testing interface, then get going on the ‘components’ system. On Jonathan’s recommendation, I’ll also try to classify and narrow my sights down to just 2-3 for each of the node, graph, and actor scopes. I’ll be doing that soon too I hope.