Profiling performance
The first step in creating a great program is getting the program fully working. The second step is to perform automated testing to prove that the program works as intended. The final step ought to be tweaking the code so that the program runs as fast and as efficiently as possible. Poorly performing programs run the risk of having a low adoption rate at best, and may simply be unusable at worst. In England, the NHS has an algorithm that was designed to match organ transplant recipients to recently harvested organs. The algorithm is complicated but extremely time-sensitive. Harvested organs must be transplanted quickly; otherwise, their tissues will die and become useless. In short, the algorithm must be extremely accurate; otherwise, the transplanted organ may be rejected, resulting in the patient’s death. It must also be fast since the organ will lose viability, which may also result in patient death. Suddenly, I’m very glad for my job dealing...