Is Gcstd_eval.from_nodes asymptotically optimal for our use case?

[nlevenson]

We build a hashmap from Gcntree nodes so we can look up node numbers in O(1), which makes the whole algorithm O(m + n) instead of O(m * n) had we just done linear search against the list of node numbers for each node in the Gcntree...

Are the setup costs and space complexity tradeoffs worth it? How big do we imagine that standards might eventually become? Can we imagine a world in which Ground Control is used for some purpose other than we originally intended, and our inputs get really big?