Exam 2
Standardization
Fundamental of desperation when no other design works. Simplifies life, but makes future development harder.
Activity Centered Design
Controls for multiple activities that need to be done simultaneously are placed together. This could be problematic in exceptional cases.
Interlock
Forces operations to take place in proper sequence (i.e. locking the oven door while it's working)
Peak End Rule
Humans judge their past experiences on the peak of that experience (whether pleasant or unpleasant) and how it ended. The net pleasantness or unpleasantness or the length of experience is disregarded.
Reporting Error
If errors can be caught, many problems can be avoided, but they're not always easy to detect. Social pressures also cause people to have difficulty admitting or reporting error.
When Good Design Isn't Enough
If humans lose competency (i.e. due to sleep deprivation) good design may fail. This should be fixed by the system, not machines.
Uncanny Valley
If it's really similar to a human but not close enough, it's worse than something completely different
Measuring Goals
If requirements are truly falsifiable, then we should be able to design user studies that prove (based on data) that our prototype is sufficient.
Sensibility Checks
Make sure the action is not extreme (i.e. transferring 1 million dollars)
Miller's Law of Short Term Memory Load
Measure's short term memory in chunks. Discovered, among other things, that users are generally able to remember 7 plus or minus 2 'chunks' of information. Miller concluded that humans can only categorize stimuli into about 7 categories.
Error and 7 stages of action
Mistakes are errors in setting the goal or plan and comparing results with expectation. Slips happen in the execution of a plan, or in the perception or interpretation of the outcome. Memory lapse can happen at any of the 8 transition between stages.
Memory- Lapse
Necessary action is not take due to a memory lapse.
Deliberate Violations
Occur when people knowingly take risks. This is unwittingly rewarded when as a result of the violation, the person is successful.
Knowledge-based Procedures
Occur when unfamiliar events occur, where neither existing skills nor rules apply. Applying knowledge based behavior is slow and reflective.
Rule-based Behavior
Occurs when normal routine is no longer applicable but the new situation is one that is known, so there is already a well prescribed course of action. (i.e. if...then..)
Mistake
Occurs when the wrong goal is established or the wrong plan is formed.
Skill-based Behavior
Occurs when workers are extremely expert at their jobs so they can do the everyday routine tasks with little or no thought or conscious attention. This results in slips.
Waiting
A little waiting is okay. Consistent and numerous waits are frustrating. Upgrades should only be done when necessary or they should be put in the background.
Prefrontal Cortex
1. Controls emotions 2. People with damage to it could not make even the simplest decisions 3. Emotions help us make decisions
Emotions and Design
1. Develop products that elicit positive responses such as feeling at ease, enjoying the experience, etc 2. Eliciting specific types of emotions such as motivation to learn, be creative, play, be metacognitive, etc 3. Eliciting trust (e.g., making users feel comfortable with divulging their payment info).
Krug's Visibility Laws
1. Don't make me think 2. It doesn't matter how many times I have to click, as long as each click is a mindless, unambiguous choice 3. Get rid of half the words on each page, then get rid of half of what is left
Categories of Prototype
1. Feel - What does it look and feel like? 2. Implementation - How does it work? 3. Role - What is the experience like?
Four Types of Frustrating Interfaces
1. Gimmicks 2. Error Messages 3. Waiting/Upgrading 4. Appearance
Error Handling Examples
1. Jidoka: stop assembly line when error discovered and find out the root cause 2. Poka- Yoke: Add signifiers/forcing function to controls to ensure proper action 3. Semi anonymous reporting: Remove the name of the reporter from the report once investigation is done
Types of Prototypes
1. Low Fidelity 2. High Fidelity
7 Usability Laws
1. Miller's Law of Short Term Memory Load 2. Fitt's Law 3. Keystroke Level Model 4. Hick's Law 5. Power Law of Practice 6. Pareto and Zipf's Laws 7. Peak End Rule
Cost of Change Over Time
1. Prototypes are MORE expensive to change the more high-fidelity they become 2. The most common estimate is that it's 100 times cheaper to make a change before any code has been written than it is to wait until after the implementation is complete [Jakob Nielsen]. 3. Twenty years of usability engineering experience uniformly indicates that the biggest improvements in user experience come from gathering usability data as early as possible in a design project [Jakob Nielsen].
Lessons of prototyping
1. Prototypes are nearly ALWAYS incomplete 2. Goal is to SIMULATE specific aspects of the design and acquire knowledge regarding these targeted aspects
Prototyping Rules
1. Prototypes should NOT be required to be complete 2. Prototypes should be easy to change 3. Prototypes should be disposable (mental prototype)
Error Messages
1. Reduce error messages 2. Speak the User's language 3. Don't burden users with extraneous tasks when errors do occur 4. Keep it simple
Expressive Interfaces
1. Simplest Form: feedback meant purposefully to illicit emotions 2. Doesn't necessarily involve fancy algorithms, emotion detection, etc 3. Just simple emotional designs meant to illicit specific emotions 4. Includes: Specific Forms of Audible Feedback Animations Spoken Messages Dynamic Icons
Paul Ekman
1. Studied facial muscles and how they relate to emotions 2. Discovered micro expressions 3. Most emotional detection software based off of Ekman's work. 4. Mapping facial expressions helps with: Lie Detection, Divorce, Emotions in general
Potential Applications of Norman's claim
1. Systems can adjust how rigid they are based on a user's emotions Hard (but not impossible) to detect emotions Often more complicated than necessary 2. Design more 'serious' applications more rigidly - Should design be more error prone just because the system is involved with a less stressful user?
Norman's Emotional Design Model
1. Visceral: responds rapidly, makes judgments regarding what is good, bad, safe, etc - Make products look and feel pleasurable 2. Behavioral: Well learned activities such as typing, talking, walking, etc - Apply design principles appropriately 3. Reflective: Conscious thought; generalizing across situations, etc. - Take into account the meaning and personal value of the product (very popular among our business savvy colleagues). Claims that affective state changes how we deal with particular situations (e.g., stress causes our muscles to tense, our glands to sweat, and improves focus)
Task Analysis
A design process that results in a good fit to the actual ways the tasks get performed.
Keystroke Level Model
A model that provides numerical predictions of user performance. Sum of sub-tasks when typing something gives total time required to type something.
Chunking
A psychological phenomenon whereby individuals group responses when performing a memory task
Conventions
A type of cultural constraints that determine what person will be inclined to do with a system (i.e. traffic rules). May make it difficult to enact change.
Swiss Cheese Model of Accidents
Accidents are most often caused when the mistakes/errors (holes) line up. Accidents can be most prevented by : 1. Adding more checks (more slices of cheese) 2. Reducing the number of critical safety points (make holes smaller) 3. Alert human when an accident is close to occurring (holes have lined up)
Cultural constraints
Actions allowed in social situations that limit how people interact with the system.
Error
All wrong actions.
Resilience Engineering
An approach used to deal with extreme weather or disaster situations (i.e. computer designers shut down entire systems to test disaster scenarios).
Unknown Unknowns
Aspects of a design that you don't know are open issues
Known Unknowns
Aspects of a design that you know you don't understand and wish to learn
Physio - Pleasure
Bodily pleasures connected to sensory experiences (E.g., the pleasurable feel of holding a really sleek phone)
Sound as Signifier
Can provide information not available in any other way. Can also tell us if things are working properly or need repair. Should convey something about the actions we just performed. Should be natural but distinct.
Human Centered Design
Careful observation of how the tasks being supported are actually performed, followed by a task analysis.
Human Error
Categorized as deviance from appropriate behavior. Most often caused by bad design. Designers don't understand or consider mental limitation (i.e. getting bored).
Hick's Law
Describes the time it takes for a person to make a decision based on the number of choices available. Essentially, users quickly categorize and halve the number of choices iteratively until they select what they want. T = b* log2(n+1) T = b H H = sum (i^n p_i \log2(1/p_i + 1))
Designing for Error
Design should especially consider what happens when things go wrong. The following should be done: 1. Understand causes of error and design to minimize causes 2. Do sensibility checks (does action make sense) 3. Make it possible to undo action or make it harder to do irreversible action 4. Make errors easier to discover, and make them easier to correct 5. Don't treat action as error, but try to help the person complete the action properly
Mode-error slips
Device has different states in which the same controls have different meanings.
Legacy Problem
Devices use the existing standard rather than choosing the better design to keep it familiar for the user.
Hindsight
Easier to understand/analyze the outcome of an event logically once it's done. Understand which factors are relevant and why.
Dealing with Error
Errors occur when humans have to complete task done by machine. Human error should not be used, since often times it's due to incompatibility between humans and machines. Can be dealt with the following ways: 1. Put operational knowledge in the world 2. Use physical, logical, semantic, and cultural constraints. 3. Bridge gap between Gulfs of Execution and Evaluation: make things visible, provide feedforward (execution) and feedback (evaluation)
Prototyping Goals
If you perform your user and task analysis well, then these are represented by your usability goals and requirements. Thus we can select the usability requirements (probably a subset) that we wish to test with a given prototype.
Visible affordances
Include cylinders, holes, etc. Indicates how something fits together or works.
Knowledge in the Head
Includes conceptual models, cultural, logical, and semantic constraints on behavior and analogies between current and previous experiences.
Knowledge in the World
Includes perceived affordances and signifiers, mappings between controls and what they control and the action they perform, and physical constraints that limit actions.
Social and Institutional Pressures
Includes the pressure of keeping the system running even when dangerous. Usually involves the expense of stopping a system or public embarrassment.
Skeuomorphic
Incorporating old, familiar ideas into new technologies, even though they no longer have a functional role. Helps with transition into new technology and modifies existing conceptual models.
Checklist
Increases accuracy of behavior and reduces error, especially slips and memory lapse. If too many people involved, it increases the chance of error since checking relaxes people too much.
Capture Slip
Instead of the desired activity, a more frequently or recently performed activity gets done instead. Occurs with experienced people more since the more frequent action becomes second nature.
Root Cause Analysis
Investigate the accident until the single underlying cause is found. With design, this involves finding out why humans made errors and try to fix it.
Prototype Process
Is NOT a process in which you haphazardly create designs that you believe are various levels of "awesome". Is about defining questions regarding your designs, and building something that answers those questions. 1. What are your goals for the prototype? What do you wish to learn? 2. How can you measure whether or not that goal has been achieved? How can you measure which of multiple prototypes is superior? 3. What is the MINIMUM amount of work necessary to produce, measure, and learn from your prototype?
Minimizing Slips
Is not done by ensuring people pay attention. Can be best accomplished by providing perceptible feedback when action is being performed and the new resulting state. Also allow undoing.
The Five Whys
Keep asking why until you discover the underlying cause. This is only a partial analysis, and doesn't guarantee success.
Lock-in
Keeps an operation active, preventing someone from prematurely stopping it (makes sure process needs to be exited). This could also include a company making their products incompatible with those of other companies.
Physical constraints
Limits what fits together or what can be done with the system physically. Also makes desired actions more obvious.
Collaboratively Followed Checklist
Made by two people (usually the pilots). Many people resist because it makes them feel incompetent. Ideally are iterative, continually being refined. Many people also skip over items thinking they'll get back to it. Can be fixed with electronic lists.
Pareto and Zipf's Law
Pareto observed that wealth was unequally distributed in Italy. He noted that 80% of the land and wealth was owned by 20% of the people (early 1900's). By fixing the top 20% of the most reported bugs, 80% of the errors and crashes would be eliminated. Zipf - As it turns out, the most frequently used command used in a piece of software is used about twice as often as the second most frequently used.
Why do errors occur
People are forced to behave in unnatural ways such as being alert for hours or multitask with interruptions.
Psycho - Pleasure
People's emotional and cognitive reactions to a product (E.g., The emotionally satisfying experience of shopping on the web).
Rule-based Mistake
Person has appropriately diagnosed the situation but decided an erroneous course of action: wrong rule is being followed. These occur in multiple ways: 1. Situation is mistakenly interpreted and wrong goal or plan is invoked and the wrong rule is followed. 2. The rule itself is faulty either because it was formulated improperly or conditions are different. 3. The correct rule is followed but the outcome is incorrectly evaluated.
Ideo - Pleasure
Pleasure involving people's values (similar to Norman's reflective level). E.g: The Prius provides ideo-pleasure to users because it's best buds with planet earth. Can include functional items that people use as art.
Detecting Error
Prevents harm if done quickly. Action slips are easy to discover compared to mistakes. Memory lapses are the most difficult. Lack of knowledge is also hard to do because it's done after much consideration.
Lockout
Prevents user from entering a space that is dangerous. Usually used for safety reasons (i.e. baby locks).
McCarthy and Wright's Technology as Experience Framework
Proposed four core threads that make up our holistic experiences: 1. Sensual Thread - our sensory engagement with a situation (visceral) 2. Emotional Thread - Emotions are intertwined with the situation in which they arise (e.g., people become angry because the website won't register their clicks) 3. Compositional Thread - The internal structure and narrative of an experience that we create (Online shoppers expect coherent steps that lead to a purchase) 4. Spatio-temporal Thread - Refers to the space and time in which our experiences take place and their effect upon those experiences Provides a means to think about emotional interaction in terms of various distinct threads of emotion
Jordan's Pleasure Model for Product Design
Proposes four conceptually distinct types of pleasure: 1. Physio Pleasure 2. Socio Pleasure 3. Psycho Pleasure 4. Ideo Pleasure This model does not attempt to describe: How these types of pleasure work at the biological level How to apply the framework directly for the purpose of interaction design Model does not assume that any product could necessarily support all four types of pleasure
Hick - Hyman
Reaction Time =Movement Time +log2(n)/Processing Speed This law only applies to groups of items that can be reasonable categorized.
Semantic constraints
Rely on the meaning or significance of the situation to limit the set of possible actions. This is dependent on our knowledge of the situation and the world.
Warning Signals
Require a lot of work since they have to be noticeable but not annoying, otherwise people will turn them off.
Interruptions
Should be avoided, especially during critical phases since they cause the most memory lapses.
Confirmations
Should emphasize the object the action is being acted on.
Explaining Away Mistakes
Small but crucial discrepancies are dismissed as inconsequential, but ended up costing a lot.
Fitt's Law
States the time it takes to reach a target with a pointing device given it's distance and size from the current pointer position. An example of a predictive model because it predicts things without the need for users. Says that things done most often should have bigger button or have button closer to average cursor position. The top, bottom, and sides of the screen are infinitely targetable because of the boundary created by the edges of the screen (unless a virtual screen exists). They should be fully utilized. T = a + b log2( 1 + D/W) T = time to move to target D = distance between point and target W = size of target
Memory-lapse Mistake
Takes place when there's forgetting at the stages of goals, plans or evaluation. Goals and plans become wrong since decision making as to be done again. Can be fixed by having the system displaying the plan.
Description-Similarity Slip
The action is performed on a object similar to the intended target.
Socio - Pleasure
The enjoyment derived from our relationship with others. Types: Products that enable social interaction Products that facilitate social interaction Products that facilitate social notice or comment
Power Law of Practice
The logarithm of the reaction time for a particular task decreases linearly with the logarithm of the number of practice trials taken. It is an example of the learning curve effect on performance.
Knowledge-based Mistake
The problem is misdiagnosed because of erroneous or incomplete knowledge. For this, a new procedure must be developed. This could be solved by intelligent machines, but right now they're rudimentary.
Prototyping
The rapid creation of an approximation to a design idea for the purpose of retrieving feedback and knowledge Used to: 1. Gain insights into user behavior 2. Communicate ideas to other teammates and stakeholders 3. Collect data for arguing the best design choice
Affective Computing
The study and development of systems and devices that can recognize, interpret, process, and simulate human affects
Action based slip
The wrong action is performed or action applied to the wrong object.
Slip
Type of error that occurs when a person intends to do one action and ends up doing something else. With a slip, the action performed is not the same as the action that was intended. Occurs more often with skilled people than novices.
Forcing Functions
Type of physical constraint; situations in which the actions are constrained so that failure at one stage prevents the next step from happening.
Gimmicks
Unnecessary imagery or actions. May include: 1. Scrolling through agreement before clicking I agree 2. Superfluous and slow animation 3. Featuritis - too many buttons
Logical constraints
Used to identify what makes sense to do with any leftover parts or parts that don't have an obvious meaning.
Don Norman's primary corollary
Users who are happier are MORE likely to cope with minor errors
Minimum Work Necessary
What is the absolute minimal working prototype with which we can perform our user study? Strip out every feature that is not absolutely necessary.
Paradox of Automation
When it works, automation vastly reduces human errors (i.e. self driving cars). When it doesn't, the accident will be huge (power grid shut down). Often fails with complex tasks.