got good enough?

May 12, 2017

People think engineering is difficult and engineers pride themselves on their technical prowess and mastery over complex topics. The truth is engineering is all about making life simpler and better. I used to think engineering was about finding the absolute best solution to a problem after having considered everything necessary. However, after graduate school and internships, I realize that the real question in engineering boils down to “is this good enough?”

  • Is this level of accuracy when measuring leak pressure good enoughfor the purposes of testing a tube’s leakage?
  • Is this linearization of a non-linear relationship good enoughfor the range we care about? For example, is this physiologically relevant range of temperatures good enough to linearize a non-linear temperature transducer relationship?
  • How should a surgeon place this neural electrode so we get a good enoughreading of the neurons we’re trying to stimulate?
  • Will testing it under these three conditions be good enoughto estimate how the device will work under other conditions?
  • Is the clarity and simplicity of my code good enoughto give me an accurate output in a good enough processing time that I’m happy with?
  • When will the motion of the robotic hand be good enoughor smooth enough that it won’t jerk when grasping an object?
  • Is the trajectory of this catheter good enoughto prevent arterial damage as it is inserted into the body?

Nothing is perfect in research or in industry. There’s never enough time to run all the experiments, try all the possibilities, and test all the results. So, you make informed decisions based on what you know and what problem you’re trying to solve. Often, engineers must estimate, run statistical analyses, and use basic knowledge of the system with which you’re working to create your own good enough criteria. For one problem, you may need 1% accuracy; for another, you may need 10%. As students, we think this is clearly defined in the real-world. But as a graduate student and intern, I have realized it never is. It is critical to define specifications and determine what defines good enough success on one component so you can move on to figuring out all the other components in the complex project jigsaw.

What this means is that every engineering problem is an optimization problem where your constraints determine what is good enough for success and conversely, what is irrelevant and can be discarded from your analysis. You can’t solve everything about everything; there are always trade-offs. Nevertheless, define your constraints correctly and you can always find a good enough solution to any problem using engineering.

Published on May 12th, 2017

Last updated on August 16th, 2018