I tend to hang out in pathfinding-space. it's where I navigate. I don't really navigate path-space, and I don't really navigate space. pathfinding-space is where I move around, and the rest just kind of unfolds based on where in that space I'm standing at the moment.
* name what you want, completely detach from the idea of going after it manually, and do reps of taking care of yourself-in-this-now. you'll evolve into something that *has* what you want - i.e. you'll gain what you want *as a side-effect* instead of *as a landing area*. which is good, because you-as-process proceeds, always proceeds, always moves. the landing area strategy works *temporarily*, but you-as-process will have leave the area at some point - and if what you want is projected as an *area* you'll have to leave it behind when you go.
* precisely define what we don't have and why we don't have it, negative constraints. take the resulting shape, hold it in hand, and start over at the *top* of the deductive stack, and walk through each deduction, holding that shape you made up to the local skyline, looking for patterns. if the shape changes, send another thread back to the top of the deductive stack with the new shape. this strategy tends to either run the proof *through itself* as a solution, or to reveal a path that already has what it needs. 'Key insight: While holding the shape, you become a different observer (you-plus-shape) and cannot re-enter the stuck space because it was contoured to you-without-shape.'
* a fork in the road: might not be able to prove anything about which path to take. you might be able to prove that it doesn't matter, though. if you know that your choice doesn't matter, what you observe next is of a different shape than "ah I know for sure that there's alternate visibility I sacrificed to get here". 'The obstruction was never "do or die" - it was "include both branches and both branches are mechanical."' 'When well-definedness seems blocked, the collinear case COLLAPSES both sides to the same direction.'
* two points are identical, and you know it, but you can't prove it. so: prove the existence of a third point from which the two points are interchangable without loss.
* negative geometry can compound safely. if you get stuck with complex negative constraints, you just zoom in and you're fine, you can find everything else in there with you. "Correct shape cannot be expected. Reach for positive mechanisms → find negative constraints instead."