Desktop Interaction C.W. Johnson, Univ ersit y of Glasgo w, - PDF document

Desktop Interaction C.W. Johnson, Univ ersit y of Glasgo w, Glasgo w, G12 8QQ. Scotland. johnson@dcs.gla.ac.uk, h ttp://www.dcs.gla .a c.uk/ � johnso n Octob er 2001 c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001 1

Desktop Interaction Ho w do w e implement a desktop system? � STAGE 1 STAGE 2 STAGE 3 Initial Design Formative Evaluation Summative Evaluation Design Objectives Operational Software Interface Task Analysis Rapid Prototyping Benchmarking Focus On Users User-Defined Interfaces Formal Experimentation Design Guidelines User-Acceptance Testing Structured Walk-Through Iterative development and user-centred design � c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

Desktop Interaction 1. consider users and en vironmen t; 2. design and sp eci�cation; 3. select dialogue st yle; 4. implemen tation issues; 5. do cumen tation issues; 6. ev aluation issues. c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

1. Consider Users and Environment Home: � - cluttered with b o oks, co�ee cups etc; - distractions from radio, other p eople etc. Ac kno wldgemen t: BBC O�ce: � - time p ressure to complete tasks; - need to achieve e ; closur - heat, noise, interruptions, multiple tasks. c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

2. Design and Sp eci�cation T ask analysis lo oks from users' p oint of view. � Sp eci�cations b egin to lo ok at system b ehaviour: � - fo cus on to do not to do it. what how State-transition diagrams often used. � c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

2. Design and Sp eci�cation States: � - situations where system continues to p erfo rm activit y; - change from a state is triggered b y a transition; - rep resented b y no des in the net w o rk. T ransitions: � - edges rep resent change b et w een states; - usually triggered b y user actions; - can b e lab elled b y frequency info rmation. Can do static analysis: � - can y ou get from state A to B (reachabilit y); - maximum of N transitions from A to D (spanning). c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

3. Select Dialogue St yle Dialogue st yles: � - text, fo rms, menus, graphics, mixtures... Dialogue st yles rely on widgets: � - lists, choices and buttons; - checkb o xes and radio buttons; - pull-do wn, tea r-o�, p op-up, scrolling, hyb rid menus c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

3. Select Dialogue St yle Windo ws: � - titles, b o rders and frames, scroll ba rs. Windo wing actions: � - op en, move, close, resize, select, b ring fo rw a rd, - feedback is crucial to all of these actions. Some very complex w eb-b ehaviours: � - sp ontaneously op ening a link ed windo w with a w eb page; - `where on ea rth did that come from?' - lots of securit y issues here - unsigned applets. c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

3. Select Dialogue St yle Windo w managers b et w een applications. � La y out managers within an application. � Placement issues: � - multiple overlapping windo ws (ho w to select a windo w?); - tiling strategies (eg p o w erp oint slide view); - zo oming (eg Acrobat page size); - ca rd cascades etc. c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

4. Implementation Issues Ho w do y ou get a widget in y our co de? � T o olkits such as A WT, Swing (o r JEWL). � /* do NOT memorise this co de! */ /* T ry to understand what it is doing */ b1 = new Button(); // create a new button b1.setLab el("Disable middle button"); // put a piece of text in it b1.setActionCommand(DISABLE); // asso ciate a command with it T o olkits a re lib ra ries of p rograms: � - others have written them to implement common widgets; - y ou don't have to write co de to dra w a button etc; - can b e di�cult to change these p re-co ded widgets. c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

4. Implementation Issues What happ ens when button is click ed? � public v oid actionP erformed(ActionEv en t e) f String command = e.getActionCommand(); if (command == DISABLE) fg // b1 w as pressed so do whatev er y ou need to g c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

4. Implementation Issues User interface p rogramming is skilled: � - can tak e 3-4 months to lea rn the basics. Lots of to ols to simplify the p ro cess. � Direct manipulation interface builder: � - select widget and place it on `screen'; - system automatically generates co de y ou need; - can b e in�exible and ine�cient. Microsoft Visual Basic, Bo rland Delphi. � Also scripting to ols such as Tcl/Tk. � c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

5. Do cumentation Issues Printed manuals: � - seldom read, intimidating and go missing; - must sta y op en at co rrect page while t yping; - must leave ro om fo r the k eyb oa rd and mouse! Key idea of manuals : minimal � - sho rt `get y ou sta rted' summa ries; - lists of available commands and sho rt cuts. c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

5. Do cumentation Issues On-line tuto rials and manuals: � - can b e quite basic (Unix apropos / man ) . On-line and context-sensitive help: � - can b e very di�cult to navigate; - can b e out of synch with soft w a re installation; - can b e pitched at to o simple/complex a level. c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

5. Do cumentation Issues But screen reading is tiring and erro r p rone. � The infamous `O�ce Assistant': � - do es anthrop omo rphism alw a ys supp o rt users? `Tip of the da y' - can b e patronising. � c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

6. Evaluation Issues Recap lecture 4: � - what is fo rmative evaluation? - what is summative evaluation? Recap lecture 4: � - name one fo rmative technique? - name one summative technique? c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001

Summa ry 1. consider users and en vironmen t; 2. design and sp eci�cation; 3. select dialogue st yle; 4. implemen tation issues; 5. do cumen tation issues; 6. ev aluation issues. c CS-1Q: HCI (Lecture 5) � C.W. Johnson, 2001

F urther Reading Shneiderman on: � - soft w a re to ols - pp. 155-179. c CS-1Q: HCI (Lecture 7) � C.W. Johnson, 2001