I read many systems papers and appreciate how research makes my career more interesting. Some interesting papers use optimal when they mean better in describing performance and I wonder if I am an outlier in thinking that optimal is occasionally misused. Optimal is formal and requires one of:
- A proof
- An objective function and math to show where it is minimized or maximized
- Exhaustive search
Objective functions and math are wonderful but they work on performance models and it is a challenge to accurately model a complex system. Proofs share that challenge.