Knowledge Engineering

“Science is about understanding the origins, nature, and behavior of the universe and all it contains; engineering is about solving problems by rearranging the stuff of the world to make new things.” -Henry Petroski

Science is engineering.

Last December, Henry Petroski wrote an article in IEEE Spectrum titled “Engineering is Not Science” where he described problems that can occur when conflating science with engineering.  The premiss was that science can get in the way of engineering progress if people are too conservative when it comes to using tools that they don’t understand.  He uses the example of the steam engine which revolutionized commerce long before a theory of thermodynamics was fully developed.

This has also been my experience, that with many scientists, there is a hard distinction made between science and engineering in a similar vein.  I’ve been told that science is not engineering, and that combining the two can lead to biased science because the plan for action will invariably influence the results.  Or more succinctly: “You can’t put science in a box.”

As a compliment to that, I’ve often heard engineering described as a mechanistic process of filling in the blanks between a need and a solution… totally algorithmic.  The “creative part” being described as the science and the mechanical execution part the engineering.  I am of the opinion that these attitudes are problematic.  They create a culture that lacks appreciation for the products of engineering and an academic community that rejects a structured approach to problem solving, relying instead upon an ad-hoc process to knowledge discovery without structured planning, execution, and a quality control process.

I prefer the following characterization:

Science is the acquisition of knowledge;  Engineering, the application of knowledge.

In this case, we do not have a competing distinction.  This places science as a form of engineering just like electrical or civil engineering.  Science is the application of knowledge to acquire knowledge.  It is “Knowledge Engineering.”  Instead of an iPhone, the scientist designs, constructs, and validates a model of reality.  As an electrical engineer would test the software and antenna to make sure a device meets the market derived specifications, a scientist will test the constructed model to verify that it creates predictions in line with known observation.  The scientist then provides the new tool to the community for usage in their work.

But why do we find a considerable lack of process and process metrics in academic science?  Why must discovery remain a magical ad-hoc process?  Why can’t creativity be the result of a structured plan of attack into a new knowledge space from multiple angles?  The fact is that the process of science can be characterized, understood, and improved.  You don’t have to live in the cold dark world of ad-hoc project execution.  You don’t have to stab blindly in the dark.  Stop doing science.  Start doing knowledge engineering!