Encapsulating knowledge for intelligent interactoin object selection

Encapsulating Knowledge for Intelligent Automatic

Interaction Objects SelectionJean Vanderdonckt, François Bodart

University of Namur, Belgium

[email protected]

in Proceedings of ACM Conference on Human Aspects in Computing Systems InterCHI'93 (Amsterdam, 24-29 April 1993), S. Ashlund, K. Mullet, A. Henderson, E. Hollnagel, T. White (Eds.), Addison Wesley, Reading (Massachusetts), pp. 424-429.

Interaction Objects SelectionIntroduction

1. Is environment independent

2. Is included in an automatic generator

3. Involves application semantic

4. Requires a dialog model

Interaction Objects Selection (2)5. Requires a user model

6. Considers screen space

7. Uses explicit rules

8. Groups related objects

Selection Requirements1. Is environment independent

2. Is included in an automatic generator

3. Involves application semantic

4. Requires a dialog model

Abstract Interaction Objects Different presentations Same behaviours Abstract (AIO) versus Concrete (CIO)

– Abstract Interaction Objects are platform-independent

– Concrete Interaction Objects are platform-specific

Taxonomy of AIOs 6 sets : action, scrolling, static, control,

dialog and feedback

Abstract Interaction Objects (2)Generic name, definitionNatureTypeAggregation, inheritanceOperations = (causes, effects)Abstract attributes, events and primitives

– PushButton_TriggeredFunctionName

– PushButton_OnSelection

– PushButton_TriggerFunction

Selection Requirements1. Is environment independent

2. Is included in an automatic generator

3. Involves application semantic

4. Requires a dialog model

TRIDENT Approach OverviewSpecification editor

ERA, FCG databases

AIO selector

Selection rules

AIO specifications

AIO to CIO mapper

CIO specifications

CIO placer

UIL objects

Presentation editor

Specification Editor

Abstract Interaction Object Selector

UIDL Specifications

Selection Requirements1. Is environment independent

2. Is included in an automatic generator

3. Involves application semantic

4. Requires a dialog model

Application Data ModelizationDomainData typesValues to chooseDefault valuePrincipal valuesSecundary values

Application Data Modelization (2)Granularity: low-medium-high

Known values: domain values

Ordered list: yes/no

Expandable list: yes/no

Continuous range: yes/no

Selection Requirements5. Requires a user model

6. Considers screen space

7. Uses explicit rules

8. Groups related objects

Application Data Modelization (3) User level : Beginner

Novice

Intermediate

Expert

Master Selection preference Constrained screen space

Selection Requirements5. Requires a user model

6. Considers screen space

7. Uses explicit rules

8. Groups related objects

Selection RulesData input, data display8 data types : hour, date, logical, integer,

numeric, real, alphabetic, alphanumericSimple AIO for elementary dataComposite AIO for grouped data (list,

group, array)

Selection Rules (example)Integer input data, known domain, Nvc > 1Nsv Exp Npv AIO= 0 no [2,3] Npv check boxes

[4,7]Npv check boxes+group box[8,Tm] List box[Tm+1,2Tm] Scrolling list box> 2Tm Scrolling drop-down list box

= 0 yes Combination box> 0 List box

Decision Trees2 trees for input/displayData type on first nodeOne simple condition by nodeBranching nodesConclusion nodes

Decision Tree (example)Nsv=0 Exp=no 2ŠNpvŠ3

Npv check boxes

Nsv>0 Exp=yes 4ŠNpvŠ7

Npv>2Tm

Npv check boxes+group box

8ŠNpvŠTm

List box

Tm+1ŠNpvŠ2Tm

Scrolling list box

Combination box

Scrolling drop-down list box

List box

Decision Tree : Conclusion Visibility Easy backtracking Easy explanation Fast selection Modifiability Refinement Rule redundancy Excessive size Predefined order

Pro

Contra