Ludum Dare 23
Archive for the ‘LD #23’ Category
Finally, relaxation! Our two-man team (Ryan and Jon) buckled down over the 72 hours to churn out this Unity powered scandalous adventure game. However, it didn’t all go smoothly. A relaxed Monday turned into a stressful race to submit the game as the deadline inched closer.
Lorc’s Icon set (used for a portion of our action icons)
Our team reduced the use of outside materials as much as possible. The icon set did not cover all possibilities for actions and required us to either manipulate the icons in Photoshop or create a new icon from scratch (there weren’t many lotion bottle icons in the RPG set *cough*).
A terrible phone camera
Preheat oven to 425 degrees. Insert brain. Leave for 72 hours (or until well done).
The code creation was frankly a pretty slapdash sort of thing. After we fleshed out our idea, Ryan dove head first into the gameplay. The first thing that was done was the general gui for interacting with the environment (clicking on stuff, popping up the actions interface, and performing actions). After the gui was done, next came hooking up all the gameplay, and smoothing out the action interface out a bit. The system soon got very specific. There quickly became things that had to be done a certain way, or named a certain way for things to work right. This was one difficulty with the super-rapid prototyping. There wasn’t much time to go back and change things if they weren’t structured well enough to start.
One thing that was very helpful was keeping a rolling document throughout the process that contained things like the specific steps needed to hook up actions, or naming conventions. In the end this became our big todo list, and it worked out pretty well for us. It got a bit messy at times, but overall it was good to have some documents to record thoughts.
As for sound, Ryan made most of the sounds (with a few done by Erica, Jon’s girlfriend) and pretty rapidly. All told, about 45 minutes was spent on music and sounds throughout the course of the competition. There was discussion about making a few tracks in garageband, but it was far behind bug fixing and assembly, and so it never came to fruition.
The world and character asset creation were both handled entirely in Blender. The level was almost all contained in one file with many objects being split into separate objects within the same scene. This process got a little messier as the room was expanded upon and filled with content (as scene navigation and object placement became especially cumbersome), but it was very easy to combine and unwrap objects throughout the scene that would be using the same texture. The reference for the house is our apartment which helped speed up the house creation.
The character creation was a blast in Blender. It took about 4-5 hours to model and animate the main character and an additional 3 and a half hours for the human that walks around the house (shitty though they both are). In a solo compo, it would be extremely difficult to have completed 3D characters and have a decent game to surround them with.
Our team split art and code down the middle (with one box having the “primary” scene); we worked in our own scenes until Sunday morning. Early Sunday morning we started hooking up all of the actions in the primary scene, assembling the majority of the game.
Starting early! We spent about an hour talking about design before diving in. Coming up with a reasonable idea quickly and plowing forward worked out well us.
Splitting up the art and programming allowed us to both launch headlong into our respective areas without getting bogged down doing the hookups for each individual item immediately.
The humor, though somewhat crude, is part of who we both are and made it fun to implement all of the ridiculous actions. The actions gradually became more twisted the further we got along– probably due to partially fried brains. Hopefully nobody gets any ideas about spreading pink eye…
What Didn’t Work
Unity and Dropbox combined. We experienced multiple difficulties even when working in different scenes (and with the Library folder of the Unity project omitted from syncing). Partway through Friday, the frustration of losing work led to us both turning off our Dropbox syncing.
Some of the actions that the character performs were originally intended to be more complex (going from two step to three steps for completion), but the time required to add the functionality (along with the risk of introducing bugs) prevented us from extending the actions. Additionally, several textures are unlit in the version that has been submitted for judging. This was a result of waiting much too late to submit the final version– it would have been better to start making the final builds even 4-5 hours before the deadline to test the builds and catch anything that has slipped through the cracks during development.
Also, I think it would have been good to have a few different camera options to use coming into this thing. We spent a fair bit of time trying to tweak one of unity’s default controllers to do what we wanted, and in the end found ourselves spending too much time on it. If we were to choose a small bit of code to bring into the next jam, it would be some custom character controllers we are bound to write by then!
The final submission is something that our team is proud to have pushed out in just a couple days. Next time we will probably have a rapid prototyping toolkit full of modular, reusable code that can be used in a wide variety of projects. For now, though, we can take some time to relax and enjoy the creations of our fellow developers– better get back to that. Thanks for reading!