A Maiden Voyage into Open Source, Datatypes, and Scala
summary
Implements an immutable B-list where each node holds a fixed-size array instead of one element, benchmarks block sizes up to 200 and chooses block size 36; concat and random access cut pointer dereferences by roughly the block factor and random access runs about 4× faster than Scala’s List. Prepend uses a share-and-fork strategy with an atomic boolean shared flag to avoid copying; prepend ends up about half as fast as List but concatenation and indexed ops show clear, measured wins.
AI-written summary · about
