acm - an acm publication

Articles

Workings of science
Is engineering applied science?

Ubiquity, Volume 2022 Issue May, May 2022 | BY Sharad Sinha


Full citation in the ACM Digital Library  | PDF


Ubiquity

Volume 2022, Number May (2022), Pages 1-6

Ubiquity Symposium: Workings of science: Is engineering applied science?
Sharad Sinha
DOI: 10.1145/3512336

There is often a debate to differentiate between science and engineering. There is also the view that engineering can be considered applied science. Drawing arguments based on historical developments, this article shows that both science and engineering complement each other. Hence, the question of one being the cause or the effect of the other does not arise.

Theodore von Karman (1881–1963), recipient of the U.S. National Medal of Science in 1962, once said, "Scientists study the world as it is, engineers create the world that never has been." It is apparent from the statement that there exists a view about two worlds: the "engineering world" and the "science world". They may be connected but neither is contained in the other. This article presents some examples of the intersection of science and engineering that clearly indicate that asking "Is engineering applied science?" is asking specious question.

First, how shall we distinguish engineers from scientists? A popular suggestion is engineers create and scientists study. This view is based on the etymology of science and engineering. "Science" comes from the Latin word "scientia" meaning "knowledge; and "engineering" comes from the Latin words "ingenuim and ingeniare" meaning "cleverness and to contrive, devise." Based on the origins of the words, scientist study in order to create knowledge, and engineers build machines or apparatuses.

But this basis for a distinction seems to unravel when one realizes that both scientists and engineers create and study. In his book The Art of Doing Science and Engineering, Richard Hamming quipped

  • In science, if you know what you are doing, you should not be doing it.
  • In engineering, if you do not know what you are doing, you should not be doing it.

He was suggesting that scientists are always searching for new knowledge whereas engineers are always trying to build safe and reliable artifacts. (Hamming even suggests that "applied science" is a field separate from science and engineering.) However, this quip does not make a clear distinction because engineers and scientists both spend a lot of time tinkering and experimenting to find what works.

Another attempt to make a distinction is the idea that engineers apply and exploit findings from science. But this does not work well either because there are many examples of engineers building things before there was any science to explain the laws that made them work. The Wright Brothers, for example, asked the Smithsonian Institution, the premier scientific institution of the day, for anything on the science of flight, but they had nothing. The brothers therefore conducted their own studies of birds' wings and combined with their engineering experience building bicycles. The success of their first aircraft led eventually to the creation of the science field of aeronautics. So in this case, the work of engineers opened the possibility of a new science. Yes, science history tells us that in the 15th century, 500 years before the Wright Brothers, it was speculated that air is a resistive medium. Meaning the resistance of air currents could provide the lift needed for flight. However, this insight would not have been useful for the Wright Brothers, even if the Smithsonian had provided it, because it had no bearing on how birds use their wings or how to steer a fixed wing aircraft.

It also works the other way around. Einstein invented his theory of relativity to resolve anomalies in the experiments to measure the speed of light in a vacuum. That led to the engineering of new kinds of instruments to observe and measure the heavens and opened the door for quantum physics and atomic physics. Newton's Laws of Motion conceived to explain and simplify a long series of inadequate scientific theories about motion of objects. Those laws, combined with calculus, enabled many engineering advances. Even in biology, which is generally understood to belong to the world of "science," laboratory work on gene editing is an engineering exercise to modify genes. The tools to automate this on a large scale have been developed by combining various types of electrical, electronic, computer, and mechanical engineering.

So which way does it go? Engineering preceding science? Science preceding engineering? There is no definitive answer. The best we can say is they dance together.

COMPUTING, ENGINEERING, AND SCIENCE

The field of computing illustrates this ongoing dance. Alan Turing's abstract machine (1936) has come to be seen as the foundation of computing by machine. However, his theory played no role in the development of the ENIAC, an engineering feat by any measure, though a few years later his ideas strongly influenced the development of the ACE computer. The earliest examples of cache memory hierarchy in computers dates to 1964 (the instruction stack in CDC 6600 is equivalent to the modern instruction cache). However, it was only much later that efficient ways were developed to write software code to increase the benefit from caches, which also benefitted from advancements in algorithms and cache design starting in the early 1960s. Clearly, however, the early cache designs were not influenced by some rigorous scientific principles. In fact, the 1964 paper by James E Thornton "Parallel Operation in the Control Data 6600" uses empirical observations as the basis for designing the instruction stack. Sir Maurice Wilkes is generally credited with introducing the idea of and analysis related to "memory caching" in his 1965 paper where he used the term "slave memory." It can be safely assumed that the work on CDC6600 happened at the same time but independent of Wilkes's paper. His paper does mention that the "slave principle has been applied to very small super-speed memories associated with the control of a computer."

Another interesting example in computing is related to floating point arithmetic. The floating-point number system is used in computers to represent real numbers. But computer registers and memory locations contain limited number of bits, so that only a tiny subset of real numbers can be represented exactly. For example, a 64-bit machine can represent only 264 numbers exactly. All others are rounded off to nearby exact numbers. This means iterative algorithms, such as differential equation solvers on a grid, can accumulate so many round-off errors that their results cannot be trusted. The field of numerical analysis emerged in early computer science to characterize rounding errors in computations and give ways to organize algorithms to minimize them. That field was enriched by contributions from both computer scientists and mathematicians. Recently, John L Gustafson proposed the universal number arithmetic framework and the posit number system that limit the rounding errors associated with the IEEE floating point representation.

Similarly, compilers began as engineered programs to translate source program files into machine code files. Computing theorists provided methods to assure that the translations preserved the original input-output function. Later they provided code-optimization methods that reduce machine code to a bare minimum.

These examples illustrate a common pattern: Engineers devise an initial solution without the benefit of a theory; the design is later improved by mathematicians and scientists who find ways to optimize its performance. A good design results from a dance over time between engineering, mathematics, and science.

COMPLEXITY REDUCTION BY THE HUMAN MIND

There are plenty of examples in scientific history and literature that demonstrate the two-way dance between science and engineering. Nonetheless we often find it attractive to say science "caused" the engineering in the sense that the final engineering product fits within a scientific theory. This oversimplified notion is a disservice that hides the co-evolution of an engineering design and a theory that explains it.

To some extent our modern idea that artifacts are the result of applying science stems from the thinking of Vannevar Bush, science advisor to President Roosevelt during World War II. In his 1945 essay "Science the Endless Frontier," Bush claimed basic research precedes applied research, which in turn precedes useful artifacts. This claim influenced U.S. science policy, in which government supports basic research because its time-to-use is too long for industry and leaves the responsibility for applied research to industry. This model dominates our thinking even though most modern research is actually a dance between looking for basic principles and instantiating them in prototypes. The Bush model turns out to be an oversimplification that was convenient for defining w research the government would pay for but does not describe the actual processes of research and the critical roles that engineering and science play.

CONCLUSION

The deeper we dig into the history of scientific development, the more we learn the "science world" and the "engineering world" are in a constant dance of interactions. Thus, while it is facile to suggest a distinction between science and engineering, and even flattering to the scientists or the engineers, this distinction is artificial. There is not cause-and-effect relationship between engineering and science.

It is often the case that science is inspired to seek to understand something that engineers have discovered to be useful either through chance, experience, or inspiration. It is also often the case that engineers have adopted models invented by scientists and mathematicians. The best answer to the original question of this essay—is engineering applied science—is that neither science nor engineering causes the other. Look how silly the reverse question sounds: Is science applied engineering? In truth, science and engineering are simply different aspects of the same thing. Trying to separate them into distinct entities is bound to be a fruitless enterprise.

Author

Sharad Sinha is an assistant professor of computer science and engineering at the Indian Institute of Technology (IIT) Goa, India. He is fascinated with the developments in a variety of science and technology related topics in general and computing in particular. His primary research is in computer architecture, embedded systems and high performance computing. He enjoys discussions with people from different disciplines including arts and social sciences and is always looking forward to do something new, interesting and possibly impactful. He is also the editor-in-chief of IEEE Potentials for 2020.

2022 Copyright held by the Owner/Author.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2022 ACM, Inc.

COMMENTS

POST A COMMENT
Leave this field empty