Let’s review the steps we went through to get to the final MinuteHourCounter design. The process is typical of how other pieces of code evolve.

First, we started by coding a naive solution. This helped us realize two design challenges: speed and memory use.

Next, we tried a “conveyor belt” design. This design improved the speed and memory use but still wasn’t good enough for high-performance applications. Also, this design was very inflexible: adapting the code to handle other time intervals would be a lot of work.

Our final design solved the previous problems by breaking things down into subproblems. Here are the three classes we created, in bottom-up order, and the subproblem each one solved: