Cognitiv e mo dels Ov erview � goal and task hierarc hies � linguisti c � ph ysical and device � arc hitectural Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (1) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Cognitiv e mo dels They mo del asp ects of user: � understanding � kno wledge � in ten tions � pro cessing Common categorisati on: � Comp etence � P erformance Computational �a v our No clear divide Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (2) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Goal and task hierarc hies Men tal pro cessing as divide-and-conquer Example: sales rep ort pro duce rep ort gather data . �nd b o ok names . . do k eyw ords searc h of names database further sub-go als . . sift through names and abstracts b y hand further sub-go als . searc h sales database further sub-go als la y out tables and histograms further sub-go als write description further sub-go als Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (3) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Issues for goal hierarc hies � Gran ularit y Where do w e start? { Where do w e stop? { Routine learned b eha viour, not { problem solving The unit task { � Con�ict More than one w a y to ac hiev e a goal { � Error Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (4) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
T ec hniques � Goals, Op erators, Metho ds and Selection (GOMS) � Cognitiv e Complexit y Theory (CCT) � Hierarc hical T ask Analysis (HT A) Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (5) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
GOMS what the user w an ts to ac hiev e Goals basic actions user p erforms Op erators decomp osition of a goal in to Metho ds subgoals/op erators means of c ho osing b et w een Selection comp eting metho ds Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (6) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
GOMS example GOAL: ICONISE-WINDOW [select GOAL: USE-CLOSE-METHOD MOVE-MOUSE-TO-WI NDOW- HEADE R POP-UP-MENU CLICK-OVER-CLOSE -OPTI ON GOAL: USE-L7-METHOD PRESS-L7-KEY] F or a particular user: Select unless Rule 1: USE-CLOSE-METHOD another rule applies. If the applicatio n is GAME , select Rule 2: L7-METHOD . Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (7) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
CCT Tw o parallel descriptions: pro duction rules User generalised transition net w orks Device Pro duction rules are of the form: if c ondition then action T ransition net w orks co v ered under dialogue mo dels Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (8) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Example: editing with vi Pro duction rules are in long-term memory Mo del con ten ts of w orking memory as attribute-v al ue mapping (GOAL perform unit task (TEXT task is insert space) (TEXT task is at 5 23) (CURSOR 8 7) Rules are pattern-matc hed to w orking memory , e.g., LOOK-TEXT task is at %LINE %COLUMN is true, with = 5 LINE = 23. COLUMN Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (9) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
F our rules w ould mo del inserting a space: SELECT-INSER T-SP A CE INSER T-SP A CE-MO VE-FIRST INSER T-SP A CE-DOIT INSER T-SP A CE-DONE (SELECT-INSER T-SP A CE IF (AND (TEST-GO AL p erform unit task) (TEST-TEXT task is insert space) (NOT (TEST-GO AL insert space)) (NOT (TEST-NOTE executing insert space))) THEN ((ADD-GO AL insert space) (ADD-NOTE executing insert space) (LOOK-TEXT task is at %LINE %COLUMN))) When �red, adds to w orking memory (GOAL insert space) (NOTE executing insert space) (LINE 5) (COLUMN 23) Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (10) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Notes on CCT P arallel mo del Pro ceduralisation of actions No vice v ersus exp ert st yle rules Error b eha viour can b e represen ted Measures � Depth of goal structure � Num b er of rules � Comparison with device description Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (11) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Problems with goal hierarc hies � a tec hnique p ost ho c � exp ert v ersus no vice � Ho w cognitiv e are they? Simple extensions p ossible (e.g., closure) Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (12) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Linguistic notations Understanding the user's b eha viour and cognitiv e di�cult y based on analysis of language b et w een user and system. Similar in emphasis to dialogue mo dels � Bac kus{Naur F orm (BNF) � T ask{Action Grammar (T A G) Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (13) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
BNF V ery common notation from computer science A purely syn tactic view of the dialogue lo w est lev el of user b eha viour T erminals CLICK-MOUSE, MOVE-MOUSE ordering of terminals; higher Non terminals lev el of abstraction select-menu, position-mouse Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (14) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Example of BNF Basic syn tax: ::= nonter minal expr ession An expression con tains terminals and non terminals com bined in sequence (+) or as alternativ es ( j ). ::= + points + dr aw l ine sel ect l ine choose l ast point ::= + sel ect l ine pos mouse C LI C K M O U S E ::= choose points choose one + choose one choose points j ::= + choose one pos mouse C LI C K M O U S E ::= + l ast point pos mouse D B L C LI C K M O U S E ::= pos mouse N U LL + M O V E M O U S E pos mouse j Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (15) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Measuremen ts with BNF Num b er of rules (not so go o d) Num b er of + and j op erators Complications � same syn tax for di�eren t seman tics � no re�ection of user's p erception � minimal consistency c hec king Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (16) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
T A G Making consistency more explici t Enco ding user's w orld kno wledge P arameterised grammar rules Non terminals are mo di�ed to include additional seman tic features Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (17) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Consistency in T A G In BNF, three UNIX commands w ould b e describ ed as ::= + + copy cp f il ename f il ename + + cp f il enames dir ector y j mov e ::= mv + f il ename + f il ename + + mv f il enames dir ector y j ::= + + l ink ln f il ename f il ename + + ln f il enames dir ector y j No BNF measure could distinguish b et w een this and a less consisten t grammar in whic h ::= + + l ink ln f il ename f il ename + + ln dir ector y f il enames j Human{Com puter In teraction, Pren tice Hall Mo dels of the User in Design (18) A. Dix, J. Finla y , G. Ab o wd and R. Beale � c 1993 Chapter 6
Recommend
More recommend
