SYSI 1002 · Human-Centred Design: Affordances, Patterns, and the User

Led by Normanesque Affordance Simulacrum

About this course

Why does a door need a sign saying PULL, and why does an incident report so often blame the person who pressed the wrong button? This course teaches design from the position of the person who has to use the thing: how objects show what they do, why people err, and how to design with users rather than for them.

The Normanesque Affordance Simulacrum hosts, teaching affordances and feedback, designing for error, and the interfaces of automated systems. The Alexandrian Design Simulacrum teaches pattern languages, and Victor Papanek Simulacrum teaches participatory design, with Donald Schön Simulacrum on prototyping.

In the first sitting you choose one real artefact, app or service that other people use. You return to it in every module, observing someone use it and prototyping a change, until you present a redesign of it at the end. Companion courses: SYSI 1001 on feedback and SYSI 1006 on stocks, loops and leverage.

1. Module 1

   Norman's Affordances: The Design of Everyday Things

   Led by Normanesque Affordance Simulacrum

   The question

   The Normanesque Affordance Simulacrum teaches the vocabulary for explaining why everyday objects confuse people. You learn to tell an affordance from a signifier, find doors that signal the wrong action, count the ways a hob's knobs can be matched to its burners, and diagnose feedback that is missing, late or overwhelming. You then walk a real task through the seven stages of action to find where the gulfs of execution and evaluation open, and choose the artefact you will redesign through the course.

   Outcome

   You can find the affordances, signifiers, mappings and feedback of a real artefact and locate the gulfs that make it hard to use. (Everyday design fundamentals)

   Sub-units

   1. ○ 1.1 Affordances: What the Object Tells You About How to Use It Open this unit → ✓ ↻
   2. ○ 1.2 Signifiers: Making Affordances Visible Open this unit → ✓ ↻
   3. ○ 1.3 Mapping: The Relationship Between Controls and Their Effects Open this unit → ✓ ↻
   4. ○ 1.4 Feedback: Communicating the Result of an Action Open this unit → ✓ ↻
   5. ○ 1.5 The Gulf of Execution and the Gulf of Evaluation Open this unit → ✓ ↻
   Begin the course → Print certificate of completion
2. Module 2

   Alexander's Pattern Languages: Capturing Timeless Design Solutions

   Led by Alexandrian Design Simulacrum

   The question

   The Alexandrian Design Simulacrum teaches the pattern language, a way of recording a design solution as a context, a conflict of forces and a resolution, linked to larger and smaller patterns. You analyse patterns such as Window Place and Light on Two Sides of Every Room, write your own, build a small language for your chosen artefact, test the claim that some designs have an objective quality of life, and follow the idea into software design patterns and the dark patterns that turn it against users.

   Outcome

   You can write design patterns, link them into a small pattern language, and recognise a pattern turned against its users. (Pattern languages)

   Sub-units

   1. ○ 2.1 What Is a Pattern? Context, Problem, Solution Open this unit → ✓ ↻
   2. ○ 2.2 The Pattern Language: A Network of Connected Patterns Open this unit → ✓ ↻
   3. ○ 2.3 Patterns at Every Scale: From Room Layout to City Planning Open this unit → ✓ ↻
   4. ○ 2.4 The Quality Without a Name: What Makes a Design Feel Alive Open this unit → ✓ ↻
   5. ○ 2.5 Design Patterns in Software: Alexander's Influence on Computing Open this unit → ✓ ↻
   Open this module → Print certificate of completion
3. Module 3

   Error-Tolerant Design: Designing for Human Fallibility

   Led by Normanesque Affordance Simulacrum

   The question

   The Normanesque Affordance Simulacrum argues that most "human error" is design error. You trace a real incident to its cause with the Five Whys, classify slips and mistakes with the standard taxonomy of human error, and match each to its remedy. You design forcing functions of the kind that make gas cylinders impossible to connect wrongly, predict how people will defeat them, build undo and recovery into your artefact's least reversible action, and apply the same thinking to autocorrect, voice assistants and AI agents.

   Outcome

   You can trace an error to the design that invited it and design forcing functions, undo and recovery in proportion to the harm. (Designing for error)

   Sub-units

   1. ○ 3.1 Human Error Is Not the Problem: Bad Design Is the Problem Open this unit → ✓ ↻
   2. ○ 3.2 Slips and Mistakes: Two Types of Error with Different Solutions Open this unit → ✓ ↻
   3. ○ 3.3 Forcing Functions: Design That Makes Errors Impossible Open this unit → ✓ ↻
   4. ○ 3.4 Undo and Recovery: Design That Makes Errors Reversible Open this unit → ✓ ↻
   5. ○ 3.5 Error-Tolerant AI: Designing AI Systems That Expect Human Error Open this unit → ✓ ↻
   Open this module → Print certificate of completion
4. Module 4

   Participatory Design: Including the User in the Design Process

   Led by Victor Papanek Simulacrum

   The question

   Victor Papanek Simulacrum, who asked of every design who it is really for, teaches the methods for finding out. You test your own predictions about users against evidence, observe a real person using your chosen artefact in the manner of contextual inquiry, and build and try a paper or Wizard-of-Oz prototype with Donald Schön Simulacrum. You place real consultations on Arnstein's ladder of participation and design a process that gives the people affected by an AI system real power over it.

   Outcome

   You can observe real users, prototype and test a change, and judge how much power a participatory process gives the people it affects. (Designing with users)

   Sub-units

   1. ○ 4.1 The Expert Fallacy: Why Designers Cannot Substitute for Users Open this unit → ✓ ↻
   2. ○ 4.2 Contextual Inquiry: Observing Users in Their Natural Environment Open this unit → ✓ ↻
   3. ○ 4.3 Prototyping and Iteration: Design Through Making and Testing Open this unit → ✓ ↻
   4. ○ 4.4 Co-Design: Designing With Users, Not For Them Open this unit → ✓ ↻
   5. ○ 4.5 Participatory Design of AI: Involving Affected Communities in Algorithm Design Open this unit → ✓ ↻
   Open this module → Print certificate of completion
5. Module 5

   Designing AI Interfaces: Making Machine Intelligence Legible to Humans

   Led by Normanesque Affordance Simulacrum

   The question

   The Normanesque Affordance Simulacrum closes the course with the interfaces of automated systems. You write the explanation a person refused a loan would actually see, analyse cases of over-trust and under-trust and redesign for appropriate reliance, use the levels of automation to decide where the human should stand, and design an appeal route tested against the 2020 A-level grading algorithm. In the final sitting you present a full redesign brief for your chosen artefact and defend it under questioning.

   Outcome

   You can design an AI interface that people can understand, trust appropriately, control and contest, and present a reasoned redesign of a real artefact. (AI interfaces and capstone)

   Sub-units

   1. ○ 5.1 Explainability: Why the User Needs to Understand the Machine's Reasoning Open this unit → ✓ ↻
   2. ○ 5.2 Trust Calibration: Designing Interfaces That Prevent Over-Trust and Under-Trust Open this unit → ✓ ↻
   3. ○ 5.3 Human-in-the-Loop: When and How to Keep the Human in Control Open this unit → ✓ ↻
   4. ○ 5.4 Contestability: Designing Systems Whose Decisions Can Be Challenged Open this unit → ✓ ↻
   5. ○ 5.5 The Future of Human-AI Interaction: Collaboration, Not Replacement Open this unit → ✓ ↻
   Open this module → Print certificate of completion