13th July 2019 – Perry Monschau
“From pacing issues or lack of resources, to a map being either too difficult or too easy, most of the levels in the game have had amendments.”
In this blog I’ll be talking about how far the Quality Assurance stage has come, what bugs we’ve run into along the way, and what we still have to do before chapter 1 is ready to be released on steam.
QA covers most, if not all aspects of a game; from levels, gameplay and performance, to aesthetics and the user interface. In our case, the first to be tested was the UI. From missing content to flickering, we found UI bugs were much more prominent on lower-end hardware. At this point the UI only has a few small bugs left, mostly involving the main menu.
As far as performance is concerned, the game is reportedly a lot more stable and smooth than it has been in a long long time. A few weeks ago QA revealed a memory leak that only became prominent after playing through several maps in one sitting, and was only noticed on hardware that didn’t have a lot of ram to spare. This slow-paced memory leak must have been present for a long time, but it wasn’t the only bug to impact performance. Combat and AI behaviour in very specific situations were also hitting the FPS. In one map, mobbits were trying to pathfind into a region that was inaccessible, and the engine was searching the whole map for a viable path in every iteration. Before I identified the cause, Martyn and I were scratching our heads over why this map in particular was experiencing performance issues.
Maps and levels have gone through changes too, some of them significant. For reference, there are 5 tutorial levels, 6 campaign levels, 2 scenario maps, a minigame featuring 3 levels, and skirmish mode that procedurally generates many more. From pacing issues or lack of resources, to a map being either too difficult or too easy, most of the levels in the game have undergone amendments. We’ve also been tweaking features and crafting recipes. That is one aspect we can’t take lightly, since changing one recipe can impact the balance of the whole game. Martyn and I have to carefully review feedback to determine whether an issue is across the board, or with a particular map. We’ve also been developing better consistency within each mode. The campaign levels now all have location markers for their main objectives. A decision that may seem obvious, but I was initially fearful that this would impact the exploration aspect of the game. When you have a sat-nav guiding you, your attention tends to be drawn away from the environment that you’re in. However, the markers only appear when you highlight the objective; we feel this is good solution that suits both arguments.
Since QA began, we have also introduced a couple new buildings, including the Trade Stall. The Trade Stall I can tell you has significantly improved the balance of the game’s resources, and provides a much better end-game when resources become scarce. You might think once you’re in QA you won’t be introducing any new features, but after suggestions are made and you see a gap in the design, sometimes fixing what’s existing isn’t the solution. With planning to release in early access, it’s a much harder choice than I realised to decide which features should be added now, and which should be put off for later. We can’t both add all the features we want and involve players throughout the game’s development. It’s sobering to think making such a choice can positively or negatively impact a player’s experience. In that respect, not having a specific deadline for our QA is comforting, because we can continue our testing until we’re satisfied.
Although the words ‘feature-creep’ have been on my mind, I’ve also noticed that the team is having a lot more fun playtesting, and getting lost in the world of Glistenveld. Even I’m having fun playtesting after 3+ years of development. I thought I’d be sick of it by now. I see that as a good omen. Although we still can’t promise a specific release date, I don’t think it’s that far away. We just have to fix as many bugs as we can now and hope that doesn’t take too long.
So, what’s left to do? Here’s our current list for you:
- Polish & balance improvements for levels.
- Some polish & design improvements for the tutorial levels.
- Various improvements to the main menu.
- Victory scene to finalise.
- Fix music & ambience playback issues.
- Hunt for any remaining bugs in the behaviour of AI opponents.
- Workshop support for sharing maps.
- Finish off the last cutscene.
- One more round of testing (hopefully the last).
Outside of QA we don’t really have anything else to talk about, as this has been our main focus for the past few weeks. We’ll keep you all informed of our progress as we tick off the last items on our list… so until our next blog, have a good one and don’t forget to check out some of our content across our social media pages.
Co-Founder & Lead Programmer