▲ | lo_zamoyski 4 days ago | |
Indeed. All else remaining the same, a faster program is generally more desirable than a slower program, but we don't live in generalities where all else remains the same and we simply need to choose fast over slow. Fast often costs more to produce. Programming is a small piece of a larger context. What makes a program "good" is not a property of the program itself, but measured by external ends and constraints. This is true of all technology. Some of these constraints are resources, and one of these resources is time. In fact, the very same limitation on time that motivates the prioritization of development effort toward some features other than performance is the very same limitation that motivates the desire for performance in the first place. Performance must be understood globally. Let's say we need a result in three days, and it takes two days to write a program that takes one day to get the result, but a week to write a program that takes a second to produce a result, then obviously, it is better to write the program the first way. In a week's time, your fast program will no longer be needed! The value of the result will have expired. This is effectively a matter of opportunity cost. | ||
▲ | godelski 4 days ago | parent [-] | |
There's nothing more permanent than a temporary fix that works. |