[carterschonwald]

Indeed. As one of those performance obsessed haskellers, a golden rule is "choose your data structures based upon your work load". This means not just the time space complexity of the operations , but also whether you want persistence vs shared state. Etc etc. finger trees are nice for workloads that are heavy on adding / removing values on the ends, with some random access, sequential scans are ok too.

I may wind up using this or a similar data structure in a number of current commercial and open source projects