Design
Human-Computer Interaction
17 Aug 2026
Week EIGHT
Part VI of Hornbæk et al. (2025)
Intro
Everything you interact with was designed. Someone decided which functions to include, what the icons look like, and how the pieces fit together. When something “just works” we forget that a great many decisions went into making it that way—and when it is cumbersome, we feel every one of them.
Why design matters
- Design affects how we perceive, experience, remember, decide, and form habits
- Small improvements accumulate into large effects across millions of uses
- Design is a central driver of innovation and business—much of today’s revenue comes from products and services that did not exist five years ago
- Design is more than making things look nice; it pursues exploratory, critical, and artistic goals
Part VI at a glance
This part covers four chapters
- Introduction to design (Ch 30): what design is
- Design cognition (Ch 31): how designers think
- Design practice (Ch 32): how designers actually work
- Design processes (Ch 33): how design is organized
Introduction to design (Ch 30)
- Design is not the privilege of professional designers
- Hornbæk et al. (2025) quotes Herbert Simon: “Everyone designs who devises courses of action aimed at changing existing situations into preferred ones”
- The challenge here is to use our knowledge of people (Part II), needs (Part III), and interfaces (Parts IV–V) to design new interactive systems
Four views of design
Design means at least four different things
- Design as product: the artifact or plan that results—the ultimate particular
- Design as process: the activity of designing, a form of problem-solving
- Research through design: the generation of new knowledge, not just artifacts
- Design as change: design’s transformative, even political, potential
Design as product
- The outcome is an ultimate particular: a concept, plan, or artifact that realizes an idea in a testable form
- Sketches and wireframes count—they materialize the designer’s imagination so it can be assessed
- Gaver: ultimate particulars are about possible worlds
- Natural sciences ask how things are; design asks how things ought to be
Design as change
- Design is not only about objects you can hold—it can shake the world and get people to think and act in new, better ways
- Buchanan (1992) saw design as a means of changing culture and human experience
- Instead of merely fixing problems (solutionism), design can create transformative possibilities
- Example: the Nintendo Wii transformed a video console into a device for physical activity
- The catch: this makes designers’ thinking political—they must take a stance on what is desirable
Aside: research through design
- Curious Cycles is a research-through-design project on the experience of menstruation
- It used a design probe: a kit sent into the wild for participants to engage with, without the researcher present
- Probes purposefully exploit ambiguity—participants interpret their own experiences
- An unexpected finding: participants shared materials with each other, improving self-knowledge
Interaction design
- HCI’s slice of the design world is interaction design, defined by two characteristics
- a focus on interactive technology
- human-centeredness
- Designing a secure payment protocol is not HCI; designing the interface for those payments is
- The point of a design is realized only in the context where it is used
The designer’s fundamental uncertainty
- Simon distinguished the inner environment (how the artifact behaves) from the outer environment (where it is used)
- We can only design the inner environment—and hope it interacts well with the outer one
- So a design is, in a sense, a hypothesis about that interplay
- No matter how well the artifact is controlled, its potential is uncertain until it meets real use—only evaluation and deployment reduce that uncertainty
A simplified view of design—the designer accrues design knowledge as artifacts meet users across many contexts.
Four rationales for design in HCI
- Improving: design can improve usability, accessibility, and experience
- Creating: design can open entirely new opportunities for using computers
- Informing: by showing alternatives to the status quo, design informs decisions and creates the need for change
- Producing knowledge: design does not just apply HCI knowledge, it evolves it
User-centered design
- Human-centeredness: the primary aim of design is to improve human conduct
- User-centeredness: “The needs of the users should dominate the design of the interface”
- Distilled into a handful of principles
Principles of user-centered design
- User focus: serve users’ goals, tasks, and needs; avoid jargon
- User involvement: engage representative users throughout
- Iterative and incremental development: you cannot know the specifics up front
- Prototyping: evaluate ideas before converging
- Evaluation in context: test with real users in real settings
- Holistic design: everything from ads to manuals affects use
- Process customization: avoid rigidity; keep reflecting
Generating creative ideas
- Design is hard because creative ideas are hard—novelty alone is not enough, an idea must also meet objectives
- The design space is enormous: choosing which of \(n\) functions to include gives \(2^n-1\) options
- For just 50 functions, that is over \(1.1\times10^{15}\) possibilities—before you even organize them into a menu (\(50! \approx 10^{64}\) arrangements)
- Design tasks are also often poorly defined: “design a better version” invites the questions better for whom and in what sense
Boden’s three types of creativity
- Combinatorial: a novel combination within a known conceptual space (the Dvorak keyboard optimized letter placement)
- Exploratory: a new conceptual approach revealing previously unthinkable ideas (ubiquitous computing)
- Transformative: replacing a conceptual space entirely (the desktop metaphor)
Design thinking
- A popular, emancipatory idea: creativity is not the gift of a “design hero”—it is something anyone can develop
- Offers methods for divergent thinking (chart the options) and convergent thinking (narrow to the best)
- Emphasizes two activities: knowing what to design and knowing how to design it
- Knowing what to design is arguably the harder and more important part
Aside: the double diamond
The Design Council’s model has two diamonds
- First diamond—Discover then Define: understand and frame the problem
- Second diamond—Develop then Deliver: build and refine the answer
- Each diamond opens with divergence and closes with convergence
- Four principles: put people first, communicate visually, collaborate, and iterate
The double diamond: diverge then converge, twice—first on the problem, then on the solution.
Practicing design
- Creativity rarely happens spontaneously—it follows from deliberate technique and practice
- Design techniques: defined ways of performing design tasks
- Design practices: how designers actually think and work, reflecting a designer’s maturity
The reflective practitioner
- Schön (1983) argued that professional designers develop distinctive ways of knowing through practice—they reflect on their creations
- Real design problems lack the tidy goals of engineering; design is messy
- Buchanan: designers contextualize their work through placements—cycles that let them make sense of what a design is for
- These meta-cognitive skills are the product of training, not talent or DNA
Wicked problems
- A wicked problem is hopelessly constrained or underdefined
- “Innovate something for an automobile company”—how would you even begin?
- Wicked problems have no stopping rule and no clear goal
- Some even have indeterminacy: no definitive boundaries at all
- Reflection and contextual placement let a designer iteratively crack them
Practices of participation
- Participatory design: stakeholders help generate ideas, not just get studied—born in 1970s Scandinavia to democratize the workplace
- The researcher’s role shifts from translator to facilitator
- Co-design: the whole process is rethought as a collective effort; authorship is shared with user communities
- Action research: to change the world, the designer must become part of it—and accept public accountability
What should be designed?
- The key question in human-centered design is not what can be designed but what should be
- Value-sensitive design: build designs rooted in principles users find important
- A startling implication: the designer must be ready to not design when the data suggest negative outcomes for users
- Human-centricity is empirical—it is grounded in user research from early stages to evaluation
Aside: value-sensitive design of AI
- Intelligent systems are often built algorithm-first: grab a dataset, train a model, report accuracy
- But an algorithm deciding insurance policies, tumors, or job resumes lives inside a system
- Zhu et al. applied value-sensitive design to Wikipedia edit-war communities in five steps: understand the people, create prototypes, elaborate methods for working with users, learn from deployment, and assess outcomes
- The lesson: situate an algorithm as part of a system, not merely in relation to data
Summary of Ch 30
- Design reaches beyond the user interface to graphics, concepts, and services
- Design is about changing users’ practices and experiences—natural sciences seek knowledge but do not aim to change the world
- Designers’ work is organized into processes involving techniques and practices
- Design is the nexus of human-centered design where user research, evaluation, and engineering meet
Design cognition (Ch 31)
- Try it: spend ten minutes sketching alternative designs for a remote control. How many are genuinely different from the original? Did you get stuck?
- Design cognition studies designers’ thinking when creating ideas and solving problems
- It focuses on the factors, challenges (fixation, bias), and practices that shape this ability
Four cognitive processes in design
Good designers move between four distinct processes
- Divergent thinking: identify distant, novel solutions (find more solutions)
- Convergent thinking: refine iteratively toward a better solution
- Reconceptualizing problems: rethink goals and constraints to get a better problem
- Reorganizing the design situation: externalize thinking to set up better conditions
The four cognitive processes: diverge for more solutions, converge for a better solution, reconceptualize for a better problem, reorganize for better conditions.
Problems and solutions, together
- Designers rarely “just solve a problem”—they alternate between defining problems and generating solutions
- They “crack the problem” until problem and solution converge—the moment of making sense
- The recurring dilemma is exploration versus exploitation: when to stop diverging and start drilling down
- Shift too early and you miss superior solutions; shift too late and you cannot finish
The paradox of expertise
- Iteration is the most pervasive facet of design thinking
- Being an expert does not mean generating more ideas—experts often generate fewer
- But their ideas tend to be better, because they spend effort redefining the problem first
- They are also better at strategizing when to explore and when to exploit
Cognitive heuristics and biases
- A cognitive heuristic is a rule of thumb for a quick solution—handy, but it hides other options
- The availability heuristic: ideas that come to mind easily get undue weight
- This produces bias: undue attention to a small slice of the design space
A menagerie of biases
- Anchoring: centering the design around a known reference solution
- Decoying: a reference point warps how we see an alternative behind it
- Status quo bias: undue weight to a prevailing or popular design
- Bandwagon bias: following peers’ solution paths
- All of them help us produce ideas quickly—and all limit creativity
Design fixation
- Fixation is being mentally locked into a particular solution, unable to generate alternatives
- Stronger and more constraining than bias—an inability to release an idea
- Even expert designers, taught about fixation, still exhibit it in studies (though they notice it more)
- Breaking fixation: actively create new reference points—collect examples, seek metaphors and analogies, incubate ideas, draw from art and nature
Generating solutions
A well-defined design task has four constituents
- Design decisions: the open decisions toward a design
- Design space: all designs still under consideration
- Objectives: properties an acceptable design must have (ease of use, cost)
- Constraints: hard limitations (all commands must fit in the menu)
Why design is not just optimization
- Optimization is an idealized account of problem-solving—define everything precisely, then search systematically (Ch 39)
- But design is rarely like this
- The designer must overcome bias, fixation, and the exploration–exploitation dilemma
- Quantity drives quality: generate up to 100 low-fidelity candidates per hour, suspend criticism, then select, reflect, and refine
Quantity drives quality: 50 preliminary designs distilled to 25, then to 2 final designs over roughly 30 iterations (a cashless-society project).
Ideation methods
- The eight most common: brainstorming, function/morphological analysis, scenarios, conceptual maps, checklists, analogies, metaphors, storyboards
- All follow expansion (generate) then selection (distill)
- Divided by three considerations
- how far associations are sought (further = more novel, less often valuable)
- how other people are involved (more people \(\neq\) more good ideas)
- which representations are used (visuospatial sketching vs verbal brainstorming)
Aside: how to run a brainstorm
Kelley and Littman (2001) make recommendations, popular in HCI
- Sharpen the focus (better too clear than too vague)
- Playful rules; number your ideas (aim for 100 per hour)
- “Build and jump” on each other’s ideas
- The space remembers—use tables, walls, whiteboards
- Stretch your mental muscle; get physical
- Brainstorm-killers: letting the most senior person speak first, insisting on experience, not being playful
Sketching
- Sketching has assumed a special role in interaction design—quick, disposable, low-fidelity
- It is not merely drawing an idea already in your head—it is a creative act that produces new ideas
- It supports both divergent and convergent thinking
- Greenberg and Buxton warned that usability evaluation can fixate designers prematurely; sketching alleviates this
- Drawing a sketch forces key decisions, reducing the degrees of freedom for the rest
Sketches of shape-changing keys explore the design space—supporting both divergent and convergent thinking (courtesy Miriam Sturdee).
Divergent and convergent sketching
Moggridge and Buxton’s alternation
- Divergent: generate as many ideas as you can; suspend criticism
- Selection: pick a handful, ensuring they differ
- Convergent: create higher-fidelity versions
- Divergent again: try still-newer sketches
- Selection, then convergent again
- Repeat until 2–3 designs you are happy with
- Experts sketch more, produce more ideas per unit time, and impose a tree-like structure on the space
Reconceptualizing problems
- Design briefs are often ambiguous: “innovate a new style for a web page”—what is “new”?
- An ill-defined problem has no clear objectives or too many mutually dependent constraints
- A wicked problem is the extreme case
- Remedy: goal refinement—designers rarely treat the brief as a given
Techniques for reconceptualization
- Metaphor: understand the unfamiliar via the familiar (the desktop maps a physical desk to a file system)
- Analogy: conceptual transfer (Weiser’s “woodwork of everyday life” for ubicomp)
- Conceptual blending: combine known concepts (desktop plus window metaphor)
- Reframings: mindsets that let you see a problem anew
- Concept maps: graphical tools that organize a domain’s knowledge
A concept map turns a raw taxonomy into a prioritized, connected structure of a domain.
Aside: reframing rowdy pub-goers
Dorst’s frame generation example
- Problem: pub-goers disturb the city at 2 a.m.
- Paradox: more law enforcement makes it worse
- Themes: young people go out to have fun and relax
- Frame generation: treat it like planning a music festival—help, don’t control
- Solution creation: better transit, signage, restrooms, places to “sleep it off”
- Pattern retention: rebrand and sustain the new concept
Co-evolving problems and solutions
- Designers consider problem and solution simultaneously
- They start with a default solution idea—a naive or obvious design—to gain a first foothold
- Any partial solution is assessed against this default and can be rejected
- The co-evolution model (Dorst and Cross 2001): like evolution in nature, creative bursts come from moments when the default is challenged
The co-evolution model: the problem space and solution space evolve together over time, each refocusing the other.
Aside: a wastebasket for trains
- Dorst and Cross asked experienced designers to design a newspaper wastebasket for Dutch trains
- One designer, in the 26th minute, decided to do away with bins altogether—put a hole in the floor
- He realized trains already have such a system (the toilets), was “genuinely shocked” it opened onto the rails, and adopted a new goal
- Rather than incrementally improving, designers changed the problem definition
One brief—a newspaper wastebasket for trains—and nine designers who each reinterpreted it into a different solution.
Does design happen in the mind?
- The romantic (hero designer) view: creativity is innate and uncontrollable
- The non-romantic view: creativity is a skill that can be fostered
- The field has moved beyond “the seat of creativity is the designer’s mind”
- Distributed cognition (Ch 5): thinking relies on interactions with the material and social environment
- Tie a designer to a chair and demand a solution, and you will get a poor one—they need the right conditions for thinking
Summary of Ch 31
- Problem-solving is central to understanding design work
- Bias and fixation limit exploration and reduce the quality of outcomes
- Problems must be well-defined; goal refinement gets them there
- Creativity can be facilitated through the systematic use of creative methods
Design practice (Ch 32)
- Do professional designers actually use all these methods, or are they academic pastimes?
- Design practice is how design is actually done, versus how it is prescribed
- Practice includes methods, tools, styles, materials, habits, beliefs, and craft
- Design-as-practice describes what designers do; design-as-process prescribes how it should be done
Why practices exist
Practices help designers manage complexity. Six common causes
- Designers need to renew themselves to stay creative
- Design has multiple objectives and constraints (form and function)
- Design choices are made under uncertainty (the “fuzzy front end”)
- Design is affected by many contextual factors
- Projects involve many stakeholders, materials, and documents
- Design is a multidisciplinary, collaborative effort
Design practice: the individual designer (who reflects, creates, and experiments) is nested within teams, the organization, and communities of practice.
Aside: being user-centered is harder than it sounds
Wilson and colleagues studied a UK department building a customer-query app. To ensure user involvement the team had to
- Convince not just managers but all stakeholders, on their own terms
- Define “representative user”—the most representative are not those who know the old system best
- Nominate a champion of user involvement
- Keep users engaged, informed, and educated
- Manage expectations: users are not treated as designers
Experimentation and prototyping
- Design is partly a craft—designers tinker with digital tools, pen and paper, and materials
- To prototype is to build a model that can be assessed; proto implies incompleteness
- Two purposes
- studying the feasibility of an idea by creating it
- presenting an idea concretely so others can experience and test it
- A prototype needs no code—early on, code may be unwise
Fidelity levels
- Low-fidelity: paper prototypes or rapid-prototyping tools
- Medium-fidelity: adds type, color, position
- High-fidelity: high-effort simulacra covering key details in prime scenarios
- Higher fidelity is more expensive and harder to throw away—so test cheaply first
- Prototyping is not limited to software: 3D printing, conductive ink, metalworking, textiles
Prototypes exploring how children could make electronic payments—testing feasibility and gathering feedback.
Aside: designing for a dollar a day
- Prototypes at HCI conferences can carry price tags in the tens of thousands—mostly accessible to the advantaged
- Kyng (1988) asked how to make prototyping available to “resource-weak” communities like shipyard workers
- His toolkit: True Stories (narratives of how a system succeeded or failed elsewhere), Future Workshops (analysis, goals, actions), and mock-up simulations built from plywood and cardboard
- Watching co-workers use a mock-up built empathy across the team
Design tools
- Digital tools quickly realize a version of a sketch; they catalyze team communication, externalize insights, and let designers reuse past ideas
- An essential aspect is thinking through making
- But tools have drawbacks—they can constrain creativity by offering only certain functionality
- Stolterman and Pierce: designers prefer tools that are not overly prescriptive
- Example: grid snapping helps layout but locks you into grid-adhering designs
Reflection and critique
- Reflective practice is developing and transforming one’s ways of thinking
- Two dimensions of reflection
- in-action versus out-action (reflecting while designing versus stepping away)
- remembering versus gathering (the past versus new materials)
- Napping, jogging, and dinner-table conversations all count—distancing helps incubate ideas
- Design samples (from Behance, Dribbble, Pinterest) support reminiscence and inspiration
Interpretative activities
- Sensemaking: finding connections among disconnected facets of a problem
- Abductive thinking: generating explanations from observations—imagining what might be, not deducing what is
- Design rationale: an explicit statement of why a decision was made (rarely adopted—too much extra work)
- Design judgment [Wolf et al.]: appearance, compositional, framing, service, deliberated off-hand, and navigation judgments
Design critique
- Crits are sessions where designers meet to review designs
- A wireframe is presented, choices explained, and discussed with peers or stakeholder representatives
- Crits grow competence and forge a design identity—“this is how we design”
Design fiction
- A thought experiment is done in imagery, on paper, or in simulation—but not in real life
- Design fiction is a speculative thought experiment: it depicts a possible future with a hypothetical artifact
- Unlike science fiction, it confines ideas to plausible futures
- The most famous is Weiser’s 1991 narrative of “Sal,” giving concrete form to ubiquitous computing
- Fiction offers a temporally consistent, holistic story—and can leave things ambiguous to invite further thought
Participatory practices
- Participatory design: the direct involvement of people in the co-design of the tools that shape their lives
- Workshops: users and designers co-create a vision of a desirable future
- Future workshops have three phases: critique (of the current situation), fantasy (turn issues into positives), and implementation (plan how to realize it)
Co-designing and acting
- Co-designing visions: representations must be concrete—users struggle with abstract descriptions, so low-fidelity prototyping shines
- Participation should be full and equal to that of designers
- Playing and acting: play breaks fixation; acting lets users and designers experience a scenario first-person
- Iacucci’s techniques had players enact future scenarios with a toy character on a map, or use a mock-up device in everyday life
Aside: co-design with children in a pandemic
- Lee et al. ran 10 online co-design sessions with children aged 7–11 during COVID-19
- Comic boards prompted the children to fill in ideas frame by frame
- The researcher became an improviser—adapting to moving cars and low bandwidth
- The children were rarely alone: one child’s mother corrected him mid-session (“you should be drawing right now”)—interdependencies shaped the whole process
A comic board used to co-design with children—they filled in frames 4 and 5 with their own ideas for library programs.
Collaboration
- Design is almost always teamwork—an IT team of 10–20, of which 2–3 focus on user-centered design
- Multidisciplinarity is central; the line between “pure” and “applied” work has blurred
- Members must empathize with different viewpoints, build a shared language, and take on roles flexibly
- Danger: pressure to deliver drives designers to forgo healthy criticism—evaluators confirm what they already know rather than challenge it
Communities of practice
- Groups who share an interest and interact regularly to get better at it
- Design forums: online discussion boards
- Design systems: coherent, organizationally-developed systems with values, rules, samples, and reusable software
- Design patterns: prescribe a structure for successful solutions (context, problem, solution)—e.g. Borchers’s “Flat and Narrow Tree” for kiosks: no more than 5 levels deep, no more than 7 branches per node
- Design portfolios and annotated portfolios: collections that communicate a designer’s thinking
An annotated portfolio: the Photostroller shown alongside the concerns and design moves it embodies.
Growing as a designer
- Nobody is an expert designer straight out of school
- Beyond “learn by doing,” the key is to develop a practice that fits your unique contexts and resources
- Being articulate and reflective about your practice—and scrutinizing it with others—is how it evolves
- Questions to ask: what is “sacred” to you? which methods and tools fit you? what do you bring to a team? whose work do you learn from?
Summary of Ch 32
- The study of design practice looks at what designers actually do, not what they are supposed to do
- Designers not only solve problems but make sense of materials and reflect on their practices
- They experiment through sketches and prototypes at different fidelity levels
- Design fiction helps them take distance from the present to develop possible futures
Design processes (Ch 33)
- A design process defines a structure and practices for carrying out design projects in an organization
- It lays out activities, methods, and conditions for progressing from one to the next
- Purpose: ensure a certain level of quality in execution and outcomes—while regulating, directing, educating, and building a shared identity
The waterfall and its discontents
- The classical model is the waterfall: complete one stage, meet a set of requirements (“a door”), then advance
- It flows from user research to requirements to design, evaluation, and release
- Criticized as parochial and even damaging
- it copes poorly with mid-project discoveries
- it has high sunk costs—early decisions are hard to change
- evaluation comes too late to inform design
What HCI process models share
- A core insight: a design process need not be antithetical to creativity—creativity can be managed like development
- Three shared aspects
- User focus: goals defined in terms of the user, not just economics
- Iteration: perfect solutions in one shot are impossible
- Evaluation with users: goodness is shown by reference to empirical evaluation
- Four core phases: user research → design goals (requirements) → design ideas → evaluation
Normative versus suggestive processes
- Normative: an agreed-on procedure that must be followed (ISO 9241)—compliance signals reliability
- Suggestive: an idealized procedure, with execution left case-by-case (agile, at the extreme)
- Most processes fall in between
- All process models are idealizations—real projects go back and forth and skip stages, because design is hard. Even mathematicians reach proofs iteratively, not in one giant step
Usability engineering
- The usability engineering lifecycle—Nielsen, early 1990s—is perhaps the most successful user-centric process model in software
- It arose because heuristics and guidelines were insufficient: often based on intuition, no help with trade-offs
- Conceived as a lifecycle model: track how skills from earlier versions shape future ones
The ten phases (1 of 2)
- Know thy user: chart characteristics, tasks, and needs
- Competitive analysis: study existing products for guidelines and thresholds
- Setting usability goals: concrete, measurable goals—this is what distinguishes usability engineering
- Design stage: produce a testable, deployable implementation
- Coordinated design: ensure consistency across products and versions
The ten phases (2 of 2)
- Guidelines and heuristics: general, category-specific, and product-specific
- Prototyping: primitive prototypes first, even paper mock-ups
- Empirical user testing: test against the usability goals
- Iterative design: refine, and document why via design rationale
- Collecting feedback from the field: usability engineering does not end at release
Criticisms of usability engineering
- Two serious criticisms
- it offers limited support for design as a creative activity—little help on how to produce ideas
- it can be viewed as glorified trial and error, embracing no theoretical knowledge
- Its theoryless stance was motivated by the limited theories of the early 1990s
- But it risks creating a form of user-centered design detached from HCI research
A standard for human-centered design
- The ISO 9241-220:2019 standard defines a process model for human-centered design
- Usability, accessibility, user experience, and avoiding harm are stated goals
- Key message: usability is an outcome of interaction, not a property you insert into a product
- Almost 100 pages of definitions, checklists, and guidance—and it requires the whole organization, including management, to commit
- Divided into four HCPs: enterprise-level focus, human-centered design across projects, execution within a project, and introducing/operating/ending a system
A standard for ethical system design
- The IEEE Standard 7000 addresses ethical concerns and minimizes potential harm, building on value-sensitive design
- It does not guarantee ethicality—it ensures concerns are systematically considered
- Four main steps
- Concept of operations and context exploration
- Value elicitation and prioritization
- Ethical requirements definition (EVRs)
- Ethical risk-based design
- …followed by transparency management throughout
IEEE Standard 7000: from concept exploration (operations, value elicitation) through the development stage (ethical requirements, risk-based design).
Aside: the ethics of a body scanner
- IEEE 7000 was applied to an airport full-body scanner
- People complained a previous X-ray technology exposed high-quality images of naked bodies to officers
- The core value for passengers: privacy—avoid exposing figures and genitals
- But the highest-ranked ethical value requirement was air safety: no dangerous articles onboard
- Efficiency ranked below both; the privacy EVR was framed both negatively (what must not be shown) and positively (what can be)
Agile development
- Agile arose as a response to waterfall’s shortcomings, treating design as an empirical activity—learning from trials
- Emphasis on design sprints over requirements: the faster a bad idea is rejected, the better
- Tens or hundreds of sprints per project; a sprint may last a week or a single day
- Scrum attends to constraints, time pressures, and learning from prior sprints to define the next ones
- HCI methods that fit a sprint: low-fidelity prototyping, concept designs, rapid observation, heuristic evaluation
Human factors engineering
- Sometimes called cognitive engineering—the design of safe, reliable systems, with heavy emphasis on safety
- Medical device design is the canonical example (tens of thousands of US deaths yearly from medical errors)
- FDA-recommended five-stage process
- Ideation (use cases, personas)
- Requirements (field studies, task analysis)
- Design (sketching plus theories of cognitive load and attention)
- Testing (usability tests, heuristic analyses, cognitive walkthroughs)
- Maintenance (reacting to accident reports)
Design affects psychological variables
- The defining aspect of human factors engineering: design affects psychological variables
- By affecting workload, trust, or perception, design can selectively reduce the possibility of error
- Variables may be physiological (fatigue, stress), cognitive (perception, attention, mental demand), or about human reliability (reasoning errors, motor slips)
- Heavy use of analytical methods: task analysis and performance modeling (Ch 41)
- Outside safety-critical domains these methods are often too costly—motivating “theory-free” methods like usability engineering
Faking the process
- Design processes commit to the rational designer: always a good reason for a decision
- But we know designers struggle with fixation, bias, and ill-defined problems—so why “fake” a process that can never be attained?
- Parnas and Clements (1986) made the case for sticking to processes anyway
- they guide when you face a hard challenge
- they harmonize practices across teams
- they let you measure progress
- “It is very hard to be a rational designer; even faking that process is quite difficult. However, the result is a product that can be understood, maintained, and reused”
Summary of Ch 33
- Design processes define the order and manner in which methods are applied—normative accounts of how design should be done
- Most process models are iterative with four stages: user research, requirements, design generation, and evaluation
- Agile emphasizes fast iteration; usability engineering ensures high usability but has been criticized as glorified trial and error
- Even though processes are idealizations, following one offers many advantages over ad hoc practice
Activities
Activity 1: Sketch until it hurts (~35 min)
Task: Individually, take the design problem from your own project and sketch eight solutions in eight minutes — one minute each, no erasing, no styling. Then swap sheets within a group of 3 to 4 and sort each other’s sketches into piles: which of these are genuinely different approaches, and which are one idea wearing different clothes? Then break fixation deliberately by creating new reference points — a metaphor from outside the domain, an analogy from nature or another industry, an existing product solving an adjacent problem — and sketch four more.
Produce: The full sheet, with the second round marked against the first. State how many materially different approaches you actually have, using the project’s definition: different in the substance of the approach, not in how it looks.
Time: 25 min individual and group work · 10 min share-out
Debrief: Did the reference points produce new approaches, or new decoration on the approach you had already committed to in minute one?
Activity 2: Run a crit (~40 min)
Task: In groups of 3 to 4, run a structured critique of one wireframe or sketch per person. The presenter shows the design and explains the choices; the group responds in three separate passes, and does not mix them: first describe only what is there, then interpret what it seems to be for, then judge whether it achieves that. Rotate until everyone has presented.
Produce: For each design, a written record with the three passes kept apart, plus one decision the crit could not settle by argument. For that decision, specify the cheapest prototype that would settle it and say what fidelity it needs — and no more.
Time: 28 min group work · 12 min share-out
Debrief: Where did your group jump straight to judgment, and what did skipping description cost the presenter? Greenberg and Buxton warn that evaluating too early fixates the designer — did that happen here?
Activity 3: Does the model fixate too? (~40 min)
Task: In groups of 3 to 4, give an AI model the same design problem you sketched in Activity 1 and ask for fifteen solutions. Classify each against Boden’s three types of creativity: combinatorial, exploratory, or transformational. Then apply the same materially-different test you used on your own sketches.
Produce: The classified list, the count of materially different approaches, and a direct comparison against your human round: which produced more ideas, which produced more distinct ones, and which produced the idea you would actually pursue. Include one idea the model gave you that you would not have reached, and one place it was obviously anchored to the way you phrased the prompt.
Time: 28 min group work · 12 min share-out
Debrief: Did reading the model’s list widen your design space or fixate you on its framing? State whether you will use a model in the divergent phase of your own project, and what rule you would set for yourself if you do.
END
References
Buchanan, Richard. 1992. “Wicked Problems in Design Thinking.” Design Issues 8 (2): 5–21.
Dorst, Kees, and Nigel Cross. 2001. “Creativity in the Design Process: Co-Evolution of Problem–Solution.” Design Studies 22 (5): 425–37.
Hornbæk, Kasper, Per Ola Kristensson, and Antti Oulasvirta. 2025. Introduction to Human-Computer Interaction. Oxford University Press. https://doi.org/10.1093/oso/9780192864543.001.0001.
Kelley, Tom, and Jonathan Littman. 2001. The Art of Innovation. Doubleday.
Kyng, Morten. 1988. “Designing for a Dollar a Day.” Proceedings of the 1988 ACM Conference on Computer-Supported Cooperative Work (CSCW ’88), 178–88.
Parnas, David L., and Paul C. Clements. 1986. “A Rational Design Process: How and Why to Fake It.” IEEE Transactions on Software Engineering SE-12 (2): 251–57.
Schön, Donald A. 1983. The Reflective Practitioner: How Professionals Think in Action. Basic Books.
Colophon
This slideshow was produced using quarto
Fonts are Fira Sans, Fira Sans Light, and Victor Mono Nerd Font
Math is set in Fira Math via MathJax 4
