Systems & Hidden Structure

Engineering That Respects How People Work

Also called: Human-centered engineering

Personal interest
Observed in experiments
Philosophical question
Working interpretation

A mathematically elegant answer can still fail if it ignores how people actually work. Good engineering respects people's skill, the friction they face, their judgment, and the dignity of everyone who uses a system.

Whose real experience is missing from what I am optimizing?

Why it attracts me

I distrust answers that are only elegant on paper. A schedule can be optimal and still fail on the floor because it ignores how long training takes, how people coordinate, or what they need before they trust a plan. When people work around a tool, that is usually information about the tool, not a failing of the people.

The idea

Human-centered engineering starts from the people inside a system. It treats their skill, judgment and dignity as design requirements, not as noise to be removed. It takes friction seriously: the small costs of each step that look trivial in a model and add up over a day. And it remembers that people are the point of engineering, not just inputs to it (Love Is Not Something to Maximize).

An example

A Voice Before Words, a side quest on this site, explores a communication aid for people who communicate with picture symbols. A language model can turn a short symbol sequence into a full sentence, but it could also put words in someone's mouth. So the design offers several readings and lets the person choose the one that is truly theirs. Fluency must never cost authorship.

Where it connects

This is why I count accessibility as part of what makes a system successful (Designing So More People Can Take Part). It is also why automation needs oversight that accounts for real people and real consequences (Using Automation Responsibly), and why I check every optimization (Finding the Best Plan Within Limits) against the question of whose experience it leaves out.

Questions I'm still exploring

How do I bring the knowledge of the people doing the work into a model without reducing it to a number?
When should a tool make a choice for someone, and when should it only offer options?
How can I tell whether a system is respectful, not just efficient?

Sources and further reading

Don Norman, The Design of Everyday Things (revised and expanded edition, 2013)

Working interpretation: drafted from my notes and interests for review. It is not a direct quotation, and I may still change it.

In short

Human-centered engineering designs systems around the people who use and run them, not around an imagined ideal user. It means watching real work, learning where people struggle or work around a tool, and treating their skill as information rather than an obstacle. A plan that is best on paper but ignores fatigue, training or trust will often fail in practice, or succeed at a cost to the people inside it.

Why this matters to me

A schedule or a tool is only useful if people can live with it. I want my models to include the skill and the friction that numbers miss, and to leave real choices with the person rather than take them away. It matters especially in accessibility work, where careful design can decide whether someone gets to say what they actually mean.

Connections

Love Is Not Something to MaximizeValues

People are the reason engineering exists, not just inputs to be optimized, and love at home makes that easy to remember.

Using Automation ResponsiblyOversight

Responsible automation has to account for the real people who use it and live with its consequences.

Designing So More People Can Take PartValues

Engineering that respects people counts accessibility as part of success: a system some people cannot use has not fully worked.

Seeing Connections, Not Just PartsPart of

Engineering that respects people reminds systems thinking that the parts include humans, whose skill and dignity matter, not just their output.