COMM1PCOMM1P Alan Woolrych Temporary Module Handbook now available at: Usability Evaluation 3

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 The Rule of HCI: Reminder 09:03 Good Design … … demonstrably fits … … its context of use This lecture explains what we mean for HCI by “demonstrably”

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton Fundamentals: Reminder 09:04 Context Design èEvaluation

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 How Evaluation Fits In 09:05 Evaluation Constrains  Informs  Tests Fits Implies Context Design Implies

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Forms of Evaluation 09:06 Formative What’s wrong (and what’s good!) Earlier rather than later Summative How well it has (not) done At the end of each significant design iteration Before going live Before getting venture capital

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Context of Use 09:08 “of use” is a specialisation, a restriction to environments of usage, success and failure evaluations should (re-)create such contexts

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Cover the Context in Evaluation 09:10 Typical end-users and key stakeholders Use real tasks within real activities wherever possible ensure realistic motivation Test in real usage settings where possible

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Approaches to Evaluation 0912 Analytical deduction, inference, constructing arguments based on inspection of web-site Empirical factual, evidence gathered from real usage by real people

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Inspection Methods 0913 Heuristic Evaluation Keith Instone’s Web Interpretations Cognitive Walkthrough Novice users, learning site for first time “Scent” analysis (semiotic engineering) How strong are link “affordances”? How suggestive (and accurate) is a link?

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Heuristic Evaluation Visibility of system status (Nielsen) The system should always keep users informed about what is going on, through appropriate feedback within reasonable time. Instone Probably the two most important things that users need to know at your site are "Where am I?" and "Where can I go next?" Make sure each page is branded and that you indicate which section it belongs to. Links to other pages should be marked. Since users could be jumping to any part of your site from somewhere else, you need to include this status on every page.

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Heuristic Evaluation Match between system and the real world The system should speak the users' language, with words, phrases and concepts familiar to the user, rather than system-oriented terms. Follow real-world conventions, making information appear in a natural and logical order. Instone On the Web, you have to be aware that users will probably be coming from diverse backgrounds, so figuring out their "language" can be a challenge.

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Heuristic Evaluation User control and freedom Users often choose system functions by mistake and need a clearly marked "emergency exit" to leave unwanted states without having to go through an extended dialogue. Support undo and redo. Instone Many “emergency exits” are provided by the browser, but there are still things open to designers, especially taking away the user control that is built into the Web! “Home” button on every page, care with forcing users into certain fonts, colors, screen widths, browser versions, and "advanced technologies” e.g., animated GIFs. Until browsers let users stop and restart the animations, they can do more harm than good.

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Heuristic Evaluation Consistency and standards Users must not wonder whether different words, situations, or actions mean the same thing. Follow platform conventions. Instone Word content and buttons consistently. Check titles and headers against links. Users jump onto (and off of) sites from others. Custom link colors may work well for one site but could conflict with the rest of the Web, and make a site hard to use. Follow HTML and other specifications. Deviations are opportunities for unusable features to creep into a site.

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Heuristic Evaluation Error prevention Even better than good error messages is a careful design which prevents a problem from occurring in the first place. Instone Limitations of HTML forms mean that inputting information on the Web is a common source of errors. Until full-featured, GUI-style widgets arrive, use JavaScript to prevent errors before users submit, but still double-check after submission.

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Heuristic Evaluation Recognition rather than recall Make objects, actions, and options visible. The user should not have to remember information from one part of the dialogue to another. Instructions for use of the system should be visible or easily retrievable whenever appropriate. Instone If users can recognize where they are from the current page, without recall of their path there, they are less likely to get lost. Server-side image maps are invisible. Client-side maps (need java/script) are better iff images make links salient. Good labels and descriptive links are also crucial for recognition.

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Heuristic Evaluation Flexibility and efficiency of use Accelerators - unseen by novices - may speed up interaction for experts so that systems can cater to both inexperienced and experienced users. Let users tailor frequent actions. Instone Bookmarks! Make site easy to bookmark, and corners of interest too. Don’t let frames prevent users from bookmarking effectively. Avoid temporary URLs with a short lifespan. Make sure a URL lives on, even after content is taken down. GET instead of POST on forms. GET attaches parameters to URL, so users can bookmark the results of a search. Easy to link to lets others create specialised views of your site (e.g., Amazon.com's associates program)

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Heuristic Evaluation Aesthetic and minimalist design Dialogues should not contain information that is irrelevant or rarely needed. Every extra unit of information in a dialogue competes with relevant units of information and diminishes relative visibility. Instone Use progressive levels of detail. Let users drill down to details. Provide a way to go "up" to get the bigger picture for when users jump into the middle of a site. Get to rarely needed information via links. Write content for the Web. Break information into chunks and link to relevant chunks to support different uses of content.

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Heuristic Evaluation Help users recognise, diagnose, and recover from errors Express error messages in plain language (no codes), precisely indicate the problem, and constructively suggest a solution. Instone Errors will happen, despite all efforts to prevent them. Every error message should offer a (link to a) solution on the error page. If a search yields no hits, do not advise broadening a search. Provide him a link that will broaden the search.

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Heuristic Evaluation Heuristic Evaluation Help and documentation Even though systems are best used without documentation, it may be necessary to provide help. This should not be too large, be easy to search, focused on user tasks, listing concrete steps to be carried out. Instone Some basic sites need little documentation, if any. But complicated tasks need help. Don’t just slap up help pages, integrate documentation into sites. Link from your main sections into specific help and vice versa. Help could even be fully integrated into each page so users never feel assistance is too far away.

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Break 0940 Back at 09:55 Prompt!

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Planning Inspections 0955 Analyst training Method Site and Context (especially usage activities) Problem Reporting Fixed Format: Summary, Context, Assumed Cause(s), Likely Difficulties Improves Predictions, Supports Merging

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Report Format 1000

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Model-Based Methods 1003 GOMS Goals, Operators, Methods, Selectors Expert Error Free Use Keystroke Level Model/CP-GOMS Time estimates Cognitive Modelling (examples) Gaze Prediction How suggestive (and accurate) is a link?

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Empirical Methods 1005 Paper prototype testing Lab Testing Field Testing Remote observation and instrumentation Site Feedback

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Planning Tests 1008 Who? Typical Users, Key Stakeholders Evaluator present or remote, real-time or retrospective What - (de)briefing, appropriate activities? Where - Lab or field? Instruments - recording, notes, logs (careful!)

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Paper Prototype 1011 Fidelity Investment’s Rapid Testing Print outs of page designs Typical users asked 'Where do you look?' "Where would you find?" 'Where would you click?' 'What do you remember/did you learn?' Can you guess what this is?" (false text)

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Lab. Testing 1013 Pros Controlled, recorded Can focus on areas of concern (from inspections) Cons Out of Context Unrealistic, user’s motivation may be unnatural, learning effects with fixed tasks

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Our Lab

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Our Lab

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Field Testing 1016 Pros In context Realistic, natural motivation within real work settting and users’ own tasks Cons Difficult to control and record Much information gets missed/lost May get no data on areas of concern

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Remote Observation and Instrumentation 1018 Pros As field testing plus Cheap to cover many users and contexts Cons Significant loss of data (Activity Theory, recall binoculars example) Can be intrusive

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Site Feedback 1020 Pros As remote observation Cons Unreliable Tendency to vent by loud mouths Performance does not correlate to preference

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Content 1: Reminder 1021 Human-Computer Interaction for the Web Context Research èUsability Evaluation, Information Architecture Conceptual Functionality Navigation Structures, Display Design, Interaction Techniques

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Web-Page Module Handbook (Temporary) First Assignment Slides & Exercise Resources & Links

COMM1P3COMM1P3 SCET MSc EC/ECA © Gilbert Cockton 2001 Summary 1025 Plan Evaluations Formative - when, where, how Summative - what, who, why Use Evaluations If the design is broken, fix it