Hello, the question I have is a little open-ended and a lot of factors could affect the answer, but I've been struggling with this question all through the preliminary design of the game I'm trying to get off the ground. My question is: Is it a realistic goal to develop my game without using an existing game engine? I have just about every aspect of the game's architecture mapped out in a massive hierarchical directed graph, and things such as gameplay elements and lore are extensively documented. So with the game's design stage coming to a close, I need to start considering my options for developing the game.
The reason I'm considering developing the game without an existing proprietary engine is that my game has very specific graphical goals; which I think could be tailored to optimize performance. For example, many environments in the game will be completely statically illuminated, which would allow me to skip a huge amount of calculations for anything within a total-black location. Most proprietary engines such as Unreal Engine don't allow users to modify their core source files, so I would lose many opportunities for improving performance. There are also a huge amount of features that I simply don't need for the game concept. If I were to build the game closer to the metal; only implementing features that are necessary to the game, I feel that in the long-run I would save time and end up with a more streamlined game. Also, from a business standpoint, not having to license an engine and then paying royalties would definitely be beneficial.
What are your opinions on this? Is the added workload of building the game from the ground up worth the potential advantages? And if the answer is yes; other things to consider: Java or C++?, also; what hardware abstraction graphics API to use for rendering system?