Here’s the long delayed recap of the work performed around IGF and the few weeks since.
Two Weeks before IGF Deadline
Aside from lots of bug fixing (gameplay, tools, stability, graphics), and iterative improvements, the following work was done:
- A first pass at the scripting system (triggers, moving platforms, etc).
- Checkpoint structure change, serialization improvements.
- Texture caching (cut load times to 10% of their original amounts).
- Improved fish behavior (both animation and stability).
- First pass of IGF levels (with Orioto’s level design really shining imo).
- Tool improvements.
- Continued work on the scripting, grouping objects, triggering different types of scripts, chaining events. We obtained our first working test at the end of the week.
One Week before IGF Deadline
I started this week by focusing on more improvements to the graphics system. I added support for VBOs, but I won’t be able to make full use of it until I’ve set up proper texture atlases. Still my work was not in vain: I was sorting the render list twice per frame, when I only really needed to sort it when the level changed. Nice little performance boost.
I spent a lot of time working on the menu system, gameplay logic (winning, losing, dying, etc), in-game GUI, and the tip system (both getting it working, and creating and placing the tips in the level).
One morning, I attempted to shoehorn ffmpeg support into the game to play the intro demo, but gave up when it became clear it was a job that would last days, not hours. We decided to simply send the intro video alongside the demo.
I made some basic shader improvements for the water shader. We ended up with something much inferior than what we had hoped to achieve, but again, this was simply through lack of time. This area which definitely will see significant improvement in the future.
The scripting also continued to improve, to the point that we had a semi-stable system in place. We made limited use of it in the demo, but the work done was very useful from an R&D point of view. We know how to make it more effective and easier to set up.
And of course we made tons of tweaks and changes to the levels, continuing to improve them all week long.
In the end, we finished the demo, but left a bunch of tasks on the cutting room floor.
November Work
I was mostly occupied with some contract work, so things moved at a slower pace, but we spent some time improving the IGF demo to address some of the more glaring issues (pacing, confusion, visual and audio feedback).
Some of the bigger changes came in the design of the first two levels. With the demo coming together so late, we didn’t get enough opportunities to get fresh faces in front of the demo and watch them progress through the whole set of levels. Unfortunately, a huge difficulty spike in the third level was left unnoticed. We made the mistake of introducing a new gameplay element without properly teaching the player how to make use of the tools he had so far. The end result was that most first time players got stuck and frustrated in the third level after breezing through the first two.
Aside from fixing those issues, we added many more visual and audio cues, little animations and audio effects to indicate important events to the player.
So we submitted an updated version of our demo last week, and hopefully some of the judges will get to experience that.
Work Planned This Week
I have some administrative work to handle, which is as fun as ever, but nonetheless has to be done.
We’ve been planning some of the work ahead. There’s still some ways to go, but the core elements of the game are in place, and the rest is beginning to take shape. We can actually evaluate all the major areas that require work before the game is complete. That’s a nice feeling.
Personally, I am working on improving our audio system to give us more flexibility on the assets used in every level. Up to the last demo, it’s been a single audio setup for the entire game, and that obviously won’t fit the requirements of the game. Music and environmental effects play an important role in setting the mode and creating an immersive environment. Audio effects are also important for providing cues to the player. So we’re designing a system to provide us with that flexibility.
And once more, I am continuing to improve the tools. Grouping of objects (of various types) has been a glaring omission for a long time. I hacked something together for the demo (in conjunction with the scripting/event system), and that worked surprisingly well, but now we need a robust system. We’ve identified how we want to build our levels, and at this point it’s all about reducing the number of steps it takes.
