CS 5150 Software Engineering Lecture 11 Usability 2 Administrivia • • CS 5150 Feasibility studies Surveys canceled • • Anonymous feedback Team feedback 2 First Presentation Contents • Motivation, Scope & Goals • • • Whatever you’ve done... Detailed requirements, mock-ups, prototypes, system architecture Schedule & Plan • • CS 5150 DO NOT assume your audience read the feasibility study Progress to date • • • Quick -- no more than a few minutes How are you doing relative to your plan in your feasibility report? Revised plan 3 SE in the News • Recent history of patents in America changed dramatically in 1982 • • • CS 5150 One appeals court took over a lot of power Crazy stat: Microsoft patents received in ... • 1980s:5; 1990s:1,116; 2000s:12,330 Fitocracy: Will every UI become a game? 4 Interface Design • • • • CS 5150 Most human interaction with software now is some kind of GUI • Windows desktop, tablet, etc Relatively easy to learn and discoverable Support tight feedback loop Hard to automate tasks 5 Design for Direct Manipulation • Visual components are “actors” that have state and can be manipulated independently • • • • • • CS 5150 Conflict with usual programming language focus on procedures metaphors and mental models: Conceptual models, metaphors, icons; may or may not be intuitive “Navigation” in a conceptual space can engage spatial reasoning Use existing conventions whenever you can look: Appearance idioms that convey meaning feel: User input and feedback idioms 6 Design Elements: Menus • Pros: • • • • CS 5150 Certain categories of error are avoided Enables context-sensitive help Challenges: • • • • • Easy for users to learn and use Scrolling menus (e.g., states of USA) Hierarchical Associated control panels Menus plus command line Users prefer broad and shallow to deep menu systems 7 Help System Design • • • Help system design is harder than many programmers assume Must test with inexperienced users Kinds of help systems: • • • • • • • CS 5150 Manual Context information Tutorials Cook books & wizards Emergency Must have many routes to help information Never blame the user 8 Information Presentation • • • CS 5150 Simple is often preferable to pretty Text • • precise, unambiguous fast to manipulate and transmit Graphical interface • • Can be easier to learn colors and animation can convey information 9 Separation of Content from Presentation CS 5150 10 Remember the Mantra: Design and Evaluate CS 5150 11 Evaluations • • • • Quick, cheap iteration when you’re exploring a new interface idea (early stage development) Test usability with real users before launch/deployment (late stage development) Continue refining with user feedback after launch/deployment (maintenance) Categories of evaluation: • • • • CS 5150 Informal testing Expert review Measurement on the actual system Empirical evaluation with users 12 Measuring Usability • • • CS 5150 Effectiveness • Can the users complete their goals (accurately) with the software Efficiency • Effectiveness per unit time/effort Satisfaction • Subjective evaluation of how pleasing or easy the interface is 13 Evaluation based on Measurement • • Basic concept: log events in the users' interactions with a system Examples from a Web system • • • • • • CS 5150 Clicks (when, where on screen, etc.) Navigation (from page to page) Keystrokes (e.g., input typed on keyboard) Use of help system Errors May be used for statistical analysis or for detailed tracking of individual user 14 ... Sometimes Called Metrics • Analysis of system logs • • • • • When was was the help system used? What errors occurred and how often? Which hyperlinks were followed (click through data)? Human feedback • • • CS 5150 Which user interface options were used? Complaints and praise Bug reports Requests made to customer service 15 CS 5150 16 Evaluation with Users • • • “Test the system, not the user” Prepare, evaluate, analyze Preparation • • • CS 5150 Decide on goals (what activities do you want to test?) Write user tasks Recruit participants 17 Usability Lab CS 5150 18 User Observation Considerations • Is environment realistic? • Record • CS 5150 • Video, audio, I/O actions Standardized feedback 19 Analyzing User Test Results • • • CS 5150 Pay close attention to • • • Any frustrations Long pauses Incomplete tasks Resist impulse to blame users Statistics can be useful, but most users studies aren’t large enough to justify strong statistical conclusions 20 Evaluation Example: Eye Tracking CS 5150 21 Good for Complex UIs CS 5150 22 Refining Designs • • • • CS 5150 Critical to keep the roles of designer and evaluator separate Designers tend to assume their existing designs are good enough too quickly Evaluators do not understand the constraints of the project as well Negotiation is key! 23 UI Design Tension: Control of Appearance • • • • CS 5150 Does the designer control the UI down to the pixel level? Not practical in modern web apps Can be important in modern mobile apps Supporting multiple platforms well can be expensive 24 Usability and Cost • • • • CS 5150 Cost increases dramatically with degree of UI innovation Use a multi-tier user testing strategy • office mate, expert review, small external testing Testing is hard to automate Releasing hard-to-use software can be fatal 25