The last three days before the submission date were a really hard time. The original plan was to do some fine tuning, improving the game mechanics, provide a tutorial about the goals of Mage's Delight, and instructions how to play.
But things turned out differently. Monday morning after doing the final merge of all branches, the Lollipop deplyoment bug showed up again. Since then we spent most of the time to find a workaround for this. Without the possibility to test our software on the Gear itself, it was hard to do any fine tuning. We tried so many things but did not find the simple solution to get things working.
So we ended up in a binary search try and error session to find the corrrupting assets in our main game scene. About 1350 game objects in our main game scene were to blame.
Two days after the sumission ended I successfully found 5 faulty tree assets. The most surprising thing was that they contained different materials and meshes. On the other hand, the same materials and material combinations were fine when used in conjunction with other meshes.
Thursday afternoon the user cyberreality from Team Oculus finally posted a workaround for this annoying bug: Simply set the Install Location_ to Force Internal and the previously corrupt version can be deployed without problems. Well, the signing tool does not work any longer, so no further official updates. But I am happy that at least I have a playable version to show my friends :)
What we take home:
Developing for 3D VR is different. It needs very short deployment iterations to always ensure that you are on the right track. Although it takes an annoying amount of time, you should always test your app on the device at the end of the day. The user experience in the Unity editor play mode definitely cannot replace the actual feeling on the device.
Using brand new technology and SDKs with low version numbers is always a risk which has to be considered when planning your project. We felt safe at the beginning to get this straight during the project, and underestimated the difficulties. At the end we had no buffers for the bad luck case we were running into.
Developing for the Gear was exciting. None of us regrets the huge time invest, because participating in the VR Jam was an amazing experience. We will definitely continue with VR projects and make use of what we have learnt in the past four weeks.
Log in or sign up for Devpost to join the conversation.