System Boundaries and Problem Scope

What’s the opposite of the word tangential?

Very relevant to my primary focus in this newsletter (problem definition and solving) is the idea of a system. One way to think about describing a problem (and solution) is to frame it as an improvement to a system. The definition of a system is the engineering way to identify your “scope,” a perennial challenge in problem-solving. In this article I want to focus on the definition of a system and how fluency in the concept of a system boundary shifts your perspective.

First, what is a system? There are plenty of “traditional” systems you have probably heard referred to:

• Solar system

• Transportation system

• Healthcare system

• Respiratory system

• Automotive system

But when you want to understand what a system is, it takes the right kind of examples to help. This blog post by Filippo A Salustri is what I reference when teaching systems engineering to communicate the fundamentally arbitrary nature of a system. Is the engine in your car a system? Is the powertrain in your car a system? Is the whole car a system? Is the whole transportation network including vehicles and roads a system? The answer in each case is yes, that is a system.

So first we have this nesting-doll problem of systems within systems. But it gets worse than that. Why do we draw the lines where we do in each case? Why is the powertrain a system but not the engine + windshield washer fluid tank + driver-side-front tire? As Filippo put it in his post “Let’s be frank: systems don’t really exist; they’re just modeling constructs that humans use to predict future behaviours and events.” Or in our case, modeling constructs that humans use for segmenting the world.

The way we identify and communicate a system is to define what is within the system by specifying the system boundaries. If an element is on the inside of the boundary, it’s part of the system. If it is outside the boundary, it is part of “the rest of the world.” In engineering we care most about what is inside the boundary because that tells us what we are designing or understanding. The engine team at an automotive manufacturer does not get / have to design the doors of the car. Someone does, but we’ve cut up the overall design into pieces with different teams on each piece.

There are some rules to cutting up those pieces and managing the overall. Nothing about how the engine is designed should prevent working doors… So we try to define those sub-systems that can be handled separately and yet eventually fit together. The engine is a system that is connected to itself enough and disconnected enough from everything else that it can be looked at as a single thing and handed off to a team to design.

More practically, those sub-systems should be defined in such a way that we can define how the sub-system interacts with the system boundary (interfaces) and those boundaries are unlikely to change in dramatic ways that change the design within the boundary. For example, if we tried to split an engine into an “outer half” and an “inner half” you will need a lot more specifications that each connection between the outer and inner systems will be in a specific place. For the engine itself, we have a couple inputs, a couple outputs, and an outer silhouette. This is why we often talk about systems being more connected to themselves than “the rest of the world” as a way to explain how to recognize a system.

All of this feeds back in to problem solving. As I shared in an early post, scope creep is a very real and fundamental challenge. Since systems are arbitrary, there is some truth to a complaint like “how can you worry about System A when the world is burning?” It can feel like a non-sequitur when someone brings up one of these concerns. In your head, System A is a clearly defined thing that is not the whole world. But both as a nesting doll and a “what is your purpose for defining the system” kind of way, it can be easy to negotiate the system boundary to be bigger than just System A.

Everything we do is part of the system of earth / humanity / the universe. While we may also be able to define a smaller system that our actions are focused on, there is nothing stopping us from considering “the rest of everything” too. Or at least, the only thing stopping us is a joint understanding of what our system boundaries are and consequently what our scope is not.