What engineers do: rules of thumb, the method and the kinds of engineering knowledge: Engineering is not applied science
All courses / Understanding Engineering / Module 1: What engineers do: rules of thumb, the method and the kinds of engineering knowledge Part 1 of 4Engineering is not applied science
A common picture has scientists discovering and engineers applying. History does not support it well. Working devices often come first and the explanation later, engineers act on rough rules that have earned trust, and their profession keeps knowledge that no physics course teaches. This part sets out how the engineer's question differs from the scientist's, how decisions get made with partial facts, and why a margin of safety is the honest price of not knowing.
1.1 Science asks why something happens; engineering asks what will work
Picture a clay water pot, the matka, standing in the shade of a verandah in May. The water inside is several degrees cooler than the air. Households knew this long before anyone wrote down a theory of evaporation. They also knew that a thin unglazed pot cools better than a thick glazed one, that a wet cloth wrapped round it helps, and that the trick fades in the humid weeks of the monsoon. That is real knowledge, built by use. Nobody needed physics to benefit from it.
Now compare the two questions one can ask about the pot. A scientist asks why the water cools: which molecules escape from the wet surface, why their escape carries heat away, how the rate changes with humidity. The answer is general. It holds whether or not anyone ever makes a pot. An engineer asks something else: how do I get cold drinking water for a hundred people at a village well, in April, without electricity, for very little money? That answer is tied to a purpose, a place, a budget and a deadline. It borrows from science where science helps, and from habit, trial and rough guessing where it does not.
So it makes more sense to treat engineering as its own store of knowledge, which borrows from science, than as science with a spanner. A pot, a pump or a bridge must work as one whole, out of doors, made of imperfect materials and used by careless people. A list of the laws that act on clay and water does not say that a pot is worth making, how thick its wall should be, or where to stand it. Someone had to think of letting water seep and evaporate on purpose. Physics came later, to measure the effect and to help improve it.
Two consequences follow. First, engineering is judged by whether the result works and is worth its cost, not by whether the explanation is elegant. A scientific theory that fails one careful test is in trouble. A design that performs well in the field is a success even if nobody understands why. Second, engineering must act before the science is finished. A bridge cannot wait for a perfect theory of steel. The rest of this module is about how engineers manage that: by working with rules of thumb, and by drawing on several kinds of knowledge that science alone does not provide.
water creeps through the pores to the outer wall
2 Evaporationeach kg that evaporates takes about 2,400 kJ of heat from the pot and the water
3 Heat drawn from the waterthe stored water cools until heat leaking in from the air matches heat lost to evaporation
Outputs Cooler drinking waterabout 5 K lower after 210 kJ removed Water vapour in the airabout 90 gRound numbers for illustration. Real pots depend on wind, humidity and wall thickness.
- Kosky et al., Exploring Engineering (Academic Press, 2013) · ch. 1, 'What Engineers Do'
- Vincenti, What Engineers Know and How They Know It (Johns Hopkins University Press, 1990) · ch. 1, 'Introduction: Engineering as Knowledge'
1.2 Engineers act on tested shortcuts when the facts are incomplete
A rule of thumb is a shortcut that usually lands near a good answer without a full analysis. Builders, mechanics and cooks carry hundreds. Engineers prize them for an awkward but practical reason: a rule that has worked many times is often better than a calculation built on numbers nobody trusts. What matters is knowing where a rule came from, because a rule is reliable only inside the kind of job that produced it.
A concrete floor slab shows this. India's concrete code, IS 456, lets a designer skip a detailed deflection check for a slab of modest span and ordinary load if its overall depth is at least a set fraction of the span. For a slab simply resting on two walls, reinforced with the usual high-strength bars, that works out at about one twenty-eighth of the span; for a slab running continuously over several supports it is about one thirty-second. The rule says nothing about the grade of concrete or the exact load, and it is written only for spans up to 3.5 m. It gives a first figure to draw, which the engineer then checks, and adds to where the floor will carry stored grain or machines.
Notice what the rule is and is not. It is quick. It shifts the odds strongly toward a floor that does not sag, without promising it. It can pull against other rules, such as the minimum concrete cover the steel needs against rust, and the designer has to reconcile them. And outside its home ground it is silent: for a 12 m hall it would be no guide at all.
Rules also fade without being refuted. Indian builders long sized load-bearing brick walls by thickness rules handed down on site, and many such houses still stand. Those rules were not proved wrong; they dropped out of use as reinforced concrete frames took over the load and walls became thin infill. A scientific theory usually leaves the stage when a better theory beats it in a test. A rule of thumb more often leaves when the job it served changes. That is also why there is no fixed recipe for engineering. The skill lies in choosing, joining and sometimes inventing rules for the problem in front of you, and one of the most useful is itself a rule about method: cut a large problem into pieces small enough to solve.
| Case | Rule applied | Slab depth | Comment |
|---|---|---|---|
| 3.5 m room, slab resting on two walls | span / 28 | 3,500 / 28 = 125 mm | Inside the rule's range |
| 3.5 m room, slab continuous over a middle wall | span / 32 | 3,500 / 32 = about 110 mm | Continuity stiffens the slab, so the rule allows less |
| 4 m room, slab resting on two walls | span / 28 | about 143 mm, built as 150 mm | Just beyond the rule's range: check deflection properly |
| 12 m hall | No ratio applies | Needs analysis | Long spans call for beams, ribs or a different structure |
Ratios from the simplified IS 456 clause for slabs up to 3.5 m with ordinary loads and high-strength bars (35 and 40 for mild steel, each multiplied by 0.8). Worked depths are calculated. An illustration, not a design method.
Source: Bureau of Indian Standards, IS 456:2000 Plain and Reinforced Concrete: Code of Practice, as applied in Central Institute for Educational Technology (CEC) course notes, cl. 24.1 span to overall depth ratios for slabs up to 3.5 m, https://cec.nic.in/webpath/curriculum/Module/Arch/Paper16/30/downloads/faq.pdf- Hammack, The Things We Make (Sourcebooks, 2023) · ch. 1, 'The Invisible Method: How to Build a Cathedral without Mathematics, Science, or a Yardstick'
- Hammack, The Things We Make (Sourcebooks, 2023) · ch. 6, 'Science: The Gold Standard for Rules of Thumb'
- Bureau of Indian Standards, IS 456:2000 Plain and Reinforced Concrete: Code of Practice, as applied in Central Institute for Educational Technology (CEC) course notes · cl. 24.1 span to overall depth ratios for slabs up to 3.5 m, https://cec.nic.in/webpath/curriculum/Module/Arch/Paper16/30/downloads/faq.pdf
1.3 Best available resources and a safety margin turn uncertainty into a decision
Every design is made with what is at hand: materials, tools, skills, money, time and knowledge. As these change, the best answer changes too. Imagine crossing a 20 m stream. A village with timber and rope will build a suspended walkway. A district with a steel works and a road will build a steel truss. A city with a precast yard will lay concrete beams on piers. All three are good engineering because each answers the same problem with different resources. Trade-offs sit inside each answer: the rope walkway gives up capacity and lifetime for speed and cost, the truss gives up cost for strength, the concrete bridge gives up speed of construction for durability.
A real design has dozens of choices that affect one another, and some inputs are known only roughly, so the single best answer usually cannot even be defined. Engineers therefore aim for a design that clears every bar that matters and stop there. The economist Herbert Simon named this satisficing, and Vincenti found it described design offices better than the language of optimising. It is not laziness; it is how a finite team finishes a job.
The tool that makes satisficing safe is the safety factor, also called a factor of safety: design the part to withstand some multiple of the load you expect. It covers the unknown strength of a material, the unknown size of loads, workmanship and ageing. It is a rule of thumb dressed in numbers, and it is paid for in weight and money. Codes set it by how bad a failure would be: lift standards in Europe, for example, require the suspension ropes to be at least 12 times stronger than the load they carry when there are three or more ropes, and 16 times when there are only two. Too little and things fail. Too much and a bridge or aircraft becomes too heavy or too dear to build. Module 13 of this course returns to this balance. Module 12 shows how failures teach engineers where the margin was wrong.
Notice that the method hides a form of humility. The engineer does not claim to know the true strength of this particular steel bar or the true load on this particular day. The engineer claims only that, with these margins and these checks, the chance of failure is acceptably small.
| Resource at hand | Design | What it gives up for what |
|---|---|---|
| Timber, rope, labour | Suspended walkway | Gives up load and lifetime for low cost and speed |
| Steel works and a road | Steel truss | Gives up cost for strength and long life |
| Precast yard and cranes | Concrete beams on piers | Gives up speed for durability and low upkeep |
Illustrative. A real choice would also weigh flood risk, the span and who will use the bridge.
- Hammack, The Things We Make (Sourcebooks, 2023) · ch. 4, 'Resources: From Mesopotamia to the Islamic Empire to Space'
- Vincenti, What Engineers Know and How They Know It (Johns Hopkins University Press, 1990) · ch. 7, 'The Anatomy of Engineering Design Knowledge'
- Lift and Escalator Library, paper on suspension rope safety factors under EN 81 · minimum factors of 12 (three or more ropes) and 16 (two ropes), https://liftescalatorlibrary.org/paper_indexing/papers/00000145.pdf
See it in the table
H₂O WaterSt SteelTm TimberEl ElectricityCheck your understanding
A village potter makes cooler pots by trial and keeps what works, with no theory. In the sense of this part, this is best described as:
Engineering knowledge is judged by whether it helps make something that works. Many working designs were found long before their science existed.Why can two rules of thumb for the same problem disagree and both still be useful?
A rule of thumb is a context-bound shortcut that improves the odds. Two of them can pull in different directions and both be right in their own settings.A designer applies a factor of safety of 5 to a 1,000 kg lift load. What is the main reason for the factor?
The factor is a margin against what cannot be known. It does not guarantee safety, and it costs material.