In the early days of Aleph One development, there were few free-to-use engines, and even fewer with high-quality development tools, so the desire to build on that foundation was strong. Unfortunately, new features couldn't be added as easily to Forge and Anvil, whose source was never released, and the move to OS X made them increasingly more awkward to use. In the meantime, the game engine market exploded: Source and Unreal offer a vast array of features and power, while Game Maker and Flixel make it easy to quickly build retro 2D games.
Bungie's engine comes from a transition period in gaming history: it's too limited for modern 3D standards, but too complex for the build-a-game-in-48-hours crowd. If you're looking to build something that's not recognizably Marathon, now there are better options than Aleph One to use as your starting point, and that would still be true even if we had Carlos' missing source code.
The focus at this point, then, is not to evolve the engine but to keep the core Marathon experience alive, and add modern conveniences and enhancements where they don't interfere with backward compatibility. That might include dynamic lighting; Treellama's even experimented with that a bit.
Given our lives outside Aleph One, and non-engine projects like documentation, tools, and Lua scripts, it could take years before the next major engine improvement arrives. Support for Marathon 1 data files might be next out of the gate. My current experiment is movie export, to share your carnage skills on YouTube. I have a host of other ideas (as does Treellama), but most of them will never come to pass; life is short and there's so much to do. You'll just have to be surprised when the time comes. :)
