Most of the stuff you mentioned is going to be looked at, absolutely, and has already been suggested. It's also insane to think that the FPS death won't be fixed between now and 1.0. But by then we'll all have supercomputers the size of a fingernail, and they'll be able to run Dwarf Fortress easy!
Lol. But probably yes.
I've been programming a voxel engine as a hobby for two years now (which is capable of rendering about a 2km radius of terrain at good speed
), and I can tell you that by far the hardest thing out of all the topics the OP listed will be optimization. It's pretty much the most abstract, time-intensive (with slow results), clever-thought-intensive, analytical, prone-to-bend-your-mind-over-itself thing a game designer has to deal with. Not at all insurmountable, but very tricky. Biggest issues in FPS seem to be single-threading and RAM-related with Dwarf Fortress, which makes it peculiar and a poor match for most hardware setups, which are geared towards multi-threaded, GPU-intensive games. So Toady is going to have to either find very clever, huge speed improvements on many things*, or rewrite some parts of the game (potentially a lot of it) to be multi-threaded - which still doesn't solve the RAM issue, but that should be getting better if the typical hardware progresses to DDR4 (has it already? Mine's not.)
* Pathfinding, active history, temperature, fluids, and probably plants/creatures (incl. sentient) - the most-discussed solutions being in pathfinding (like jump-point search, or similar improvements to A* - idk what Toady has already implemented), but all are important, and the others are harder for us observers to figure out without seeing code (we just know that it has an impact, whereas pathfinding is widely studied outside of DF)
Probably a lot of Toady's motivation for making this game is based on the development of new features, and not necessarily making the game playable. I can totally understand that. It will get better once he gets the "fun" developments out of the way and is left with trickier new features vs. equally tricky optimizations. Then he might choose optimizations for a while.
But, of course, see the above quote - he might never even have to optimize, given that DF is on such a long-term development cycle.
(Note: I've been trying to get DF to run well on my computers for a long time now, mostly unsuccessfully. So I feel your pain for sure.)
I think others have spoken to the rest of the things the OP mentioned. I will echo that siege frequency is a problem from a gameplay perspective, but sieges might be more realistic now (minus the micro scale of populations), since there are motivations behind them (although hard to diagnose.) Also, he said he thought he fixed a glitch in the next release that should increase sieges in some situations. Hoping for the best!