GIT also falls down when the project gets large enough, which is why it's not used by Google/Microsoft/Amazon for their larger projects. For that you'll want something like perforce (which is admittedly a bit painful). Though somehow I doubt they're making nearly enough documents for that to matter.
For some areas in the minimum spec, that is certainly true (EX: RAM), but for compute resources, that is not necessarily the case. You could easily raise the hardware requirements by requiring more functionality out of the hardware instead increasing utilization. Some good examples of this would be DirectX 11 support on your GPU and SSE4 for x86 and NEON for ARM. Requiring DirectX 11 could conceivably reduce the memory required by the window manager by using tessellation for some details instead of storing static geometry. This in turn could allow for a prettier interface while keeping utilization numbers the same. Requiring SSE4/NEON to be available could make the OS smaller and faster. It could be smaller because optimized portions wouldn't need to support multiple code paths for legacy hardware and faster because of compiler optimizations that can be done to existing code using the newer instructions. Obviously it's not all that much, but we're all paying a price for the low hardware functionality requirements.
Slashdot Mirror
GIT also falls down when the project gets large enough, which is why it's not used by Google/Microsoft/Amazon for their larger projects. For that you'll want something like perforce (which is admittedly a bit painful). Though somehow I doubt they're making nearly enough documents for that to matter.
For some areas in the minimum spec, that is certainly true (EX: RAM), but for compute resources, that is not necessarily the case. You could easily raise the hardware requirements by requiring more functionality out of the hardware instead increasing utilization. Some good examples of this would be DirectX 11 support on your GPU and SSE4 for x86 and NEON for ARM. Requiring DirectX 11 could conceivably reduce the memory required by the window manager by using tessellation for some details instead of storing static geometry. This in turn could allow for a prettier interface while keeping utilization numbers the same. Requiring SSE4/NEON to be available could make the OS smaller and faster. It could be smaller because optimized portions wouldn't need to support multiple code paths for legacy hardware and faster because of compiler optimizations that can be done to existing code using the newer instructions. Obviously it's not all that much, but we're all paying a price for the low hardware functionality requirements.