Chris Elliott: 'An engineer would ask, what are mps expenses for – and what would a solution look like?'

What do engineers do? They identify a need and they create something that satisfies it. "Scientists study the world as it is, engineers create the world that never has been," said Theodore Von Karman who was both a scientist and an engineer. Or, as the Prime Minister said recently: "Engineers built our past and they shall build our future".

Engineering is done because someone needs something. In an ideal world, the engineer would simply ask what is needed and then make it. Unfortunately the real world is never that simple. That person with a need is not alone and their needs might conflict with those of their neighbours; they might not be able to afford the solution or willing to wait for it to be built; and the technology to build it might not exist.

Engineering is a compromise, reconciling the conflicting requirements of many stakeholders and the limitations of technology to come up with the least-bad trade-off that everyone can live with. That's why engineering is just as much "the art of the possible" as Machiavelli thought politics to be. And although "stakeholder" gets excessive use, it captures in one word everyone who is affected by or who affects an issue.

A crucial part of every engineer's education is to understand how the specialist expertise of a mechanical, electrical or civil engineer fits within the bigger picture to make that trade-off, to find the least-bad solution. The Royal Academy of Engineering supports visiting professors in many leading universities to share their industrial experience of integrated system design with the students and staff, to help them "think holistic" as well as "think detail".

One of the aims of the visiting professors is to bring experience of the skills that engineers use to help them solve real problems and deliver a better world – not a perfect one but an improved one. Graduate engineers often end up in roles far from spanners, concrete or wires and their skills can help in very different fields.

Look for example at a nagging political problem: MPs' expenses, a subject where debates seem to generate more heat than light. How would an engineer, better still, one with an understanding of integrated system design, set about solving it?

One of the most basic engineering tools is a set of "requirements". A requirement is an unambiguous statement of what the stakeholders want, expressed in terms of what the solution has to do, not how it has to do it. A well-written requirement also tells the engineer how success will be demonstrated – if you can't test it, it's not a requirement. Requirements are stark. They shine a bright light on the problem, forcing every stakeholder to open up and state not only what he wants, but why.

So an engineer would start by asking what MPs' expenses are for and what a good solution would look like. One requirement might be: "MPs should be able to work efficiently". Another might be: "MPs should not be required to subsidise Parliament, nor should they benefit beyond their salaries". These are not too contentious but everyone will have views about what "efficiently" or "subsidise" mean.

Once everyone has said what they want the expenses regime to do and why, the engineer works through a long and often painful process to find a compromise. That's never easy, but, because it's based on what the outcome should be and why that is desirable, it is possible to use rational arguments rather than just shouting. The result is a set of requirements that carry broad support.

The criteria of a good solution might be one that is: fair, robust (against abuse), cheap to implement (avoiding bureaucracy) and credible (perceived to be appropriate).

Now that we know what we're trying to achieve and how to measure success, the engineer would look at how other people have solved the problem – companies that post staff to remote locations or employ temporary workers. Each might provide a potential solution for the expenses regime and the engineer would probably devise others, perhaps as a hybrid of what's emerged.

The candidate solutions can be scored against each of the criteria, calmly and rationally. The heated and emotional debate has already happened and the detailed business of implementation can begin.

When the engineer finally presents the design, she or he can not only say what should be done but also why – a characteristic of good engineering design that's rarely found in the political sphere.