1. Models of the User in Design IACT 403 IACT 931 CSCI 324 Human Computer Interface Lecturer:Gene Awyzio Room:3.117 Phone:4221 4090 Email:gene@uow.edu.au

2. Overview zgoal and task hierarchies z linguistic z physical and device z architectural.

3. Cognitive models zThey model aspects of user: yunderstanding yknowledge yintentions yprocessing zCommon categorisation: yCompetence yPerformance zComputational flavour zNo clear divide.

4. Goal and task hierarchies zMental processing as divide-and-conquer zExample: sales report produce report gather data. find book names.. do keywords search of names database further sub-goals.. sift through names and abstracts by hand further sub-goals. search sales database further sub-goals layout tables and histograms further sub-goals write description further sub-goals

5. Issues for goal hierarchies zGranularity yWhere do we start? yWhere do we stop? yRoutine learned behaviour, not problem solving yThe unit task zConflict yMore than one way to achieve a goal zError

6. Techniques zGoals, Operators, Methods and Selection (GOMS) zCognitive Complexity Theory (CCT) zHierarchical Task Analysis (HTA)

7. GOMS zGoals ywhat the user wants to achieve zOperators ybasic actions user performs zMethods ydecomposition of a goal into subgoals/operators zSelection ymeans of choosing between competing methods.

8. GOMS example GOAL: ICONISE-WINDOW [select GOAL: USE-CLOSE-METHOD MOVE-MOUSE-TO-WINDOW-HEADER POP-UP-MENU CLICK-OVER-CLOSE-OPTION GOAL: USE-L7-METHOD PRESS-L7-KEY] For a particular user: Rule 1: Select USE-CLOSE-METHOD unless another rule applies. Rule 2: If the application is GAME, select L7- METHOD.

9. CCT zTwo parallel descriptions: yUser production rules yDevice generalised transition networks zProduction rules are of the form: yif condition then action zTransition networks covered under dialogue models.

10. Example: editing with vi zProduction rules are in long-term memory zModel contents of working memory as attribute- value mapping (GOAL perform unit task (TEXT task is insert space) (TEXT task is at 5 23) (CURSOR 8 7) zRules are pattern-matched to working memory, e.g., LOOK-TEXT task is at %LINE %COLUMN zis true, with LINE = 5 zCOLUMN = 23.

11. Example: editing with vi zFour rules would model inserting a space: SELECT-INSERT-SPACE INSERT-SPACE-MOVE-FIRST INSERT-SPACE-DOIT INSERT-SPACE-DONE (SELECT-INSERT-SPACE IF (AND (TEST-GOAL perform unit task) (TEST-TEXT task is insert space) (NOT (TEST-GOAL insert space)) (NOT (TEST-NOTE executing insert space))) THEN ((ADD-GOAL insert space) (ADD-NOTE executing insert space) (LOOK-TEXT task is at %LINE %COLUMN))) zWhen fired, adds to working memory (GOAL insert space) (NOTE executing insert space) (LINE 5) (COLUMN 23).

12. Notes on CCT zParallel model zProceduralisation of actions zNovice versus expert style rules zError behaviour can be represented zMeasures y Depth of goal structure y Number of rules y Comparison with device description

13. Problems with goal hierarchies zA post hoc technique zExpert versus novice zHow cognitive are they? zSimple extensions possible yclosure

14. Linguistic notations zUnderstanding the user's behaviour and cognitive difficulty based on analysis of language between user and system. zSimilar in emphasis to dialogue models zBackus-Naur Form (BNF) zTask-Action Grammar (TAG).

15. BNF zVery common notation from computer science zA purely syntactic view of the dialogue zTerminals lowest level of user behaviour yCLICK-MOUSE, MOVE-MOUSE zNonterminals ordering of terminals; higher level of abstraction yselect-menu, position-mouse

16. Example of BNF zBasic syntax: ynonterminal ::= expression zAn expression contains terminals and nonterminals combined in sequence (+) or as alternatives (|). ydraw line ::= select line + choose points+ last point yselect line ::= pos mouse + CLICK MOUSE ychoose points ::= choose one | choose one + choose points ychoose one ::= pos mouse + CLICK MOUSE ylast point ::= pos mouse + DBL CLICK MOUSE ypos mouse ::= NULL | MOVE MOUSE+ pos mouse.

17. Measurements with BNF zNumber of rules (not so good) zNumber of + and | operators zComplications ysame syntax for different semantics yno reflection of user's perception yminimal consistency checking

18. TAG zMaking consistency more explicit zEncoding user's world knowledge zParameterised grammar rules zNonterminals are modified to include additional semantic features

19. Consistency in TAG zIn BNF, three UNIX commands would be described as zcopy ::= cp + filename + filename | cp + filenames + directory zmove ::= mv + filename + filename | mv + filenames + directory zlink ::= ln + filename + filename | ln + filenames + directory zNo BNF measure could distinguish between this and a less consistent grammar in which zlink ::= ln + filename + filename | ln + directory + filenames

20. Consistency in TAG zIn TAG, this consistency of argument order can be made explicit using a parameter, or semantic feature for file operations. zfile op[Op] ::= command[Op]+ filename + filename command[Op]+ filenames + directory

21. Consistency in TAG zcommand[Op = copy] ::= cp zcommand[Op = move] ::= mv zcommand[Op = link] ::= ln

22. Other uses of TAG zUsers existing knowledge zCongruence between features and commands zThese are modelled as derived rules

23. Physical and device models zBased on empirical knowledge of human motor zsystem zUser's task: acquisition then execution. zThese only address execution zComplementary with goal hierarchies zThe Keystroke Level Model (KLM) zBuxton's 3-state model

24. KLM zSix execution phase operators zPhysical motor yK keystroking yP pointing yH homing yD drawing zMental yM mental preparation zSystem yR response

25. KLM zTimes are empirically determined. zTexecute = TK + TP + TH + TD + TM + TR zExample GOAL: ICONISE-WINDOW [select GOAL: USE-CLOSE-METHOD MOVE-MOUSE-TO-WINDOW-HEADER POP-UP-MENU CLICK-OVER-CLOSE-OPTION GOAL: USE-L7-METHOD PRESS-L7-KEY]

26. KLM zAssuming hand starts on mouse: zUSE-L7-METHOD USE-CLOSE-METHOD zOperator T (sec) zH[to kbd] 0.40 zM 1.35 zK[L7 key] 0.28 zTotal 2.03 zOperator T (sec) zP[to menu] 1.1 zB[LEFT down] 0.1 zM 1.35 zP[to option] 1.1 zB[LEFT up] 0.1 zTotal 3.75.Human{Computer Interaction, Prentice Hall zA. Dix, J. Finlay, G.Abowd and R. Beale c 1993 zModels of the User in Design zChapter 6 (24) zArchitectural models zAll of these cognitive models make assumptions zabout the architecture of the human mind. z Long-term/Short-term memory z Problem spaces z Interacting Cognitive Subsystems z Connectionist z ACT z.Human{Computer Interaction, Prentice Hall zA. Dix, J. Finlay, G.Abowd and R. Beale c 1993 zModels of the User in Design zChapter 6 (25) zDisplay-based interaction zMost cognitive models do not deal with user zobservation and perception. zSome techniques have been extended to handle zsystem output (e.g., BNF with sensing terminals, zDisplay-TAG), but problems persist. zLevel of granularity zExploratory interaction versus planning