In the wild

In reality, there are a lot of factors that may influence the choice of space and runtime complexity. Typically, these factors are forms of resource constraints, such as power consumption on embedded devices, clock cycles in a cloud-hosted environment, and so on.

Since it is difficult to find out the complexities of a particular algorithm, it is helpful to know a few, so the choice comes intuitively. Often, the runtime complexity is not the only important aspect, but the absolute execution time counts. Under these conditions, a higher runtime complexity can be preferable if n is sufficiently small.

This is best demonstrated when Vec<T> contains only a few elements, where a linear search is a lot faster than sorting and then running a binary search. The overhead of sorting might just be too much compared to searching right away.

Getting this trade-off and the overall implementation right is hugely beneficial for the entire program and will outweigh any other optimizations....