12/11 First Deadline
Dec 11
Today was the first deadline for Trespasser. For programming, we were planning to show off the monster and other game-play elements in the dev-test world, separately from the main map that was not finished.
Failures:
Upgrading the engine version 20 minutes before demo time. This was the worst mistake ever and I will never try to do it again. The reason I wanted to update to UE 4.24 was because it had enhanced the Niagara particle system (which we were looking into using while developing the fire particle on the lighter) as well as better terrain editing and a new hair simulation system (Neo said that he was interested in doing hair simulation for one of the animal carcasses). I had thought that we had already upgraded the project version on Monday when the new engine came out, but instead we had all just downloaded it and hadn't actually converted the project. We realized this when there were new meshes imported in 4.24 and the 4.23 working copies weren't able to open them. This was 10 minutes before the deadline. In a panic I tried launching the editor in the new 4.24, not realizing that the engine would have to build shaders for the next 20 minutes, eating into demo time. Also in the engine upgrade, when it was finally able to play, the Vive pro wand controllers would not allow the player to teleport; an issue that has not yet been resolved and will hopefully be fixed tomorrow by the second attempt.
The other major issue was committing the level. David had made major progress in detailing the bunker, graveyard, and started to assemble some houses. When he tried to commit those changes to the SVN he ended up only committing a single stair asset, proceeding to revert the rest of the changes he had made that day. I really thought all had been lost and he would have to redo the parts he had done that day, but we came up with a solution. BACKUPS. The auto-saves single handedly saved a full day of work. Reverting the chunks of land from the backups folder brought back the graveyard. The only thing left to revert was the house asset. Once we did however, the thumbnail on the asset was correct but upon loading the asset, all transforms were reset to 0, 0, 0. We have yet to figure out why this is and it must be fixed by tomorrow.
If I have learned anything from these failures it is that demo builds should be made often and that nothing should be added on the day of a demo. These were our greatest downfalls as we had neither of them.
Successes:
We wanted to add a few last things and test before the deadline. Before we found the problems with engine versions not matching and the SVN reverting work, we were able to get a suprising number of features implemented. These included: adding new sounds, fixing old broken sounds, fixing old animations, fixing AI variable setting that caused constant detection and broke the blinding mechanic, and found a possible issue with hiding that will have to be addressed soon.
Everyone was able to come together within the last minutes to fix problems (maybe not in the most efficient ways but fixed nonetheless) and collaberate in a time of panic. I think this experience strengthened us as a team while exposing our weaknesses, and all I can hope is that we can continue tomorrow with the same confidence to get a working product.
Today was the first deadline for Trespasser. For programming, we were planning to show off the monster and other game-play elements in the dev-test world, separately from the main map that was not finished.
Failures:
Upgrading the engine version 20 minutes before demo time. This was the worst mistake ever and I will never try to do it again. The reason I wanted to update to UE 4.24 was because it had enhanced the Niagara particle system (which we were looking into using while developing the fire particle on the lighter) as well as better terrain editing and a new hair simulation system (Neo said that he was interested in doing hair simulation for one of the animal carcasses). I had thought that we had already upgraded the project version on Monday when the new engine came out, but instead we had all just downloaded it and hadn't actually converted the project. We realized this when there were new meshes imported in 4.24 and the 4.23 working copies weren't able to open them. This was 10 minutes before the deadline. In a panic I tried launching the editor in the new 4.24, not realizing that the engine would have to build shaders for the next 20 minutes, eating into demo time. Also in the engine upgrade, when it was finally able to play, the Vive pro wand controllers would not allow the player to teleport; an issue that has not yet been resolved and will hopefully be fixed tomorrow by the second attempt.
The other major issue was committing the level. David had made major progress in detailing the bunker, graveyard, and started to assemble some houses. When he tried to commit those changes to the SVN he ended up only committing a single stair asset, proceeding to revert the rest of the changes he had made that day. I really thought all had been lost and he would have to redo the parts he had done that day, but we came up with a solution. BACKUPS. The auto-saves single handedly saved a full day of work. Reverting the chunks of land from the backups folder brought back the graveyard. The only thing left to revert was the house asset. Once we did however, the thumbnail on the asset was correct but upon loading the asset, all transforms were reset to 0, 0, 0. We have yet to figure out why this is and it must be fixed by tomorrow.
If I have learned anything from these failures it is that demo builds should be made often and that nothing should be added on the day of a demo. These were our greatest downfalls as we had neither of them.
Successes:
We wanted to add a few last things and test before the deadline. Before we found the problems with engine versions not matching and the SVN reverting work, we were able to get a suprising number of features implemented. These included: adding new sounds, fixing old broken sounds, fixing old animations, fixing AI variable setting that caused constant detection and broke the blinding mechanic, and found a possible issue with hiding that will have to be addressed soon.
Everyone was able to come together within the last minutes to fix problems (maybe not in the most efficient ways but fixed nonetheless) and collaberate in a time of panic. I think this experience strengthened us as a team while exposing our weaknesses, and all I can hope is that we can continue tomorrow with the same confidence to get a working product.
Comments
Post a Comment