* They can return things in tuples
* They can be perf tested & faster
* They can make it more obvious how the process the input slice (instead of modifying it, they return a new slice that is the 'next' part of the data.