(more) performant memory storage #32
Description
Look into using one of the newer pieces of swift collections, specifically the pending-1.1-release HashTreeCollections, which provides an optimized append-only structure that sounds like it would be lovely for the backing store of operations. (right now, it's just using Set<> which, while functional, might not provide the best operations for scaled up sets of data.
To verify the performance aspects, I think I'd want to bolster the benchmarks applied here, and seriously consider some of the classic CRDT benchmark structures, such as the string-sequence one that Martin Kleppmann (encoded at https://github.com/automerge/automerge-perf)