Hm I think you are right. The Guava code is apparently the simpler version of the algorithm that takes linear time (but still no memory). (Apologies, it appea

Hm I think you are right. The Guava code is apparently the simpler version of the algorithm that takes linear time (but still no memory). (Apologies, it appeared to be different at first and I downvoted you; wish I could undo it). It seems likely that the new part is the optimizations. Still, it is pretty interesting as I didn't think Guava overlapped with this area and these authors much within Google (I work there). It's usually inaccurate to speak of Google as one entity with specific reasons for doing things a certain way; there are many people working on many different and overlapping things :)