The Engineer's Curse: Knowing Exactly How Many Ways Something Can Go Wrong
There is a peculiar moment in an engineer’s life when a perfectly ordinary sentence stops being ordinary.
“Don’t worry, it’ll be fine,” someone says.
A non-engineer may hear reassurance. An engineer hears an uncompleted fault tree.
Fine under what conditions? Fine for how long? Fine if the mains supply drops out, the Wi-Fi router updates itself at 3am, the battery was fitted backwards, the person who configured it has left the company, and somebody has written the password on a Post-it that is now in a bin somewhere near Slough?
This is the engineer’s curse: not pessimism exactly, but an inconvenient ability to see the chain of dependencies holding a thing together. It is hard to admire a simple solution once you can see the cable tie, undocumented setting and hopeful assumption beneath it.
It starts with useful questions
Engineering teaches a healthy suspicion of the word “works”. A system does not simply work or fail. It works within a set of conditions, with particular inputs, tolerances, maintenance arrangements and people around it.
Take a temperature-monitoring system. The sensor may be accurate. But is it positioned where it measures the product temperature rather than the warm air near a door? Is its clock correct? Will it store readings during a network outage? Does an alarm reach a person who is both available and able to act? What happens if that person is on holiday, asleep, or has understandably muted their phone after receiving 47 meaningless alerts the previous night?
Each question is mundane. Together, they are the difference between a useful system and an expensive electronic ornament.
That habit of questioning is valuable because failures rarely arrive wearing a little badge saying Hello, I am the single point of failure. They emerge at the joins: between hardware and software, supplier and customer, policy and practice, or one person’s assumption and another person’s interpretation.
The engineer is not necessarily expecting disaster. We are just aware that a surprising amount of civilisation rests on connectors being plugged in properly.
The trouble is that the habit follows you home
Once you have learned to think this way, you cannot easily turn it off.
You look at a restaurant’s QR-code ordering system and wonder what happens when the mobile signal is poor. You notice the smoke alarm chirping and calculate how long it has been ignored. You hear somebody promise that they have backed up their photos and develop the careful expression normally reserved for unexploded ordnance.
Planning a trip can become a small systems exercise. There are bookings, travel times, access needs, weather, charging, medication, documents, contingency plans and the remote but oddly persistent possibility that an app will require a verification code sent to a phone with no signal. Technology has done wonderful things for travel, including ensuring that a delayed train can now notify you digitally that it remains delayed.
For disabled people, this kind of forward planning is often not optional. Access arrangements can be fragile. A lift may be out of service, a promised seat may not be usable, a venue may describe itself as accessible while quietly meaning that there is one ramp around the back near the wheelie bins. Seeing the failure modes is not neurotic fussiness. It is often the practical work required to participate on equal terms.
That is one reason casual advice to “just be spontaneous” can land rather badly. Spontaneity is easier when the environment has already done some of the planning for you.
Risk is not the same as fear
There is a misconception that engineers are miserable people who assume every bridge is moments from collapse. Some undoubtedly are, and meetings need all sorts.
Good engineering is not a competition to imagine the most apocalyptic outcome. It is about proportion. A loose label on a cable is annoying. A misleading label on an emergency shutdown circuit is rather more significant. The response should match the consequence, the likelihood and the ability to detect a problem before it becomes serious.
A useful risk assessment asks practical questions:
- What can fail?
- What happens if it does?
- How quickly will anyone know?
- Can the system fail safely?
- Who is expected to respond, and do they have what they need?
That last question deserves more attention than it gets. Organisations often design processes around an imaginary person: permanently available, fully trained, calm under pressure and mysteriously delighted to receive an alarm at 2:17am. Real people have shifts, families, limits and lives. A design that depends on heroics is not robust. It is merely borrowing reliability from someone’s goodwill.
The human side of redundancy
People sometimes treat redundancy as waste: a second sensor, a backup power supply, an alternate route, a manual procedure. Yet redundancy is often an act of kindness.
It acknowledges that devices fail, networks disappear and people make mistakes. It does not demand perfection from any one component or person. A well-designed system says, in effect, “We anticipated a difficult day, and you should not have to carry it alone.”
That applies outside engineering too. Clear instructions help the person who did not attend the original meeting. A spare key helps the person whose hands are full. A backup plan for a family gathering can make space for different energy levels, dietary needs and travel constraints. None of this is glamorous, but neither is spending Boxing Day attempting to revive the heating with a YouTube tutorial and inappropriate confidence.
The best systems are rarely the cleverest-looking ones. They are the ones that remain understandable when something has gone wrong and everyone is slightly tired.
Living with the curse without becoming exhausting
There is, admittedly, a social cost to being the person who asks what happens if the internet goes down. You do not want to become a roaming audit report at a birthday party. Most people are not asking for a structured review of their barbecue logistics, however tempting the discussion of gas-cylinder storage may be.
The trick is to use the instinct selectively. Raise the concern that matters. Explain it in human terms. Offer a workable improvement rather than simply presenting a catalogue of doom.
And leave room for the fact that not every uncertainty needs eliminating. Life cannot be fully fail-safe, nor should it be. Relationships, travel, work and family life all require trust, judgement and the willingness to proceed without a complete datasheet. That is not bad engineering. It is recognising that people are not components, even when they do occasionally require rebooting after a difficult week.
Knowing how many ways something can go wrong can be tiring. But it can also make us more considerate designers, colleagues, partners and neighbours. The point is not to fear every loose connection. It is to notice the ones that matter, fix what can be fixed, and make the world a little less dependent on luck.