a comparative analysis of task modeling notations
DESCRIPTION
Task Models describe how to perform activities to reach users' goals. Task models represent the intersection between user interface design and more systematic approaches. Task models can be represented at various abstraction levels. When designers want to specify only requirements regarding how activities should be performed, they consider only the main high-level tasks. On the other hand, when designers aim to provide precise design indications then the activities are represented at a small granularity, thus including aspects related to the dialogue model of a user interface (which defines how system and user actions can be sequenced). In this paper a comparative analysis of selected models involving multiple users in an interaction is provided in order to identify concepts which are underexplored in today’s multi-user interaction task modeling. This comparative analysis is based on three families of criteria: information criteria, conceptual coverage, and expressiveness. Merging the meta-models of the selected models enables to come up with a broader meta-model that could be instantiated in most situations involving multi-user interaction, like workflow information systems, CSCWTRANSCRIPT
A comparative analysis of task modeling notations
Dra. Josefina Guerrero García Dr. Juan Manuel González CallerosDr. Jean Vanderdonckt *
Facultad de Ciencias de la ComputaciónBenemérita Universidad Autónoma de Puebla*Université catholique de Louvain
Outline
•Introduction•Overview of task models•A multi-user interaction meta-model•Conclusion
Introduction (1/4)
•Task models describe how to perform activities to reach user´s goals.
•Task models represent the intersection between user interface design and more systematic approaches.
•A number of task modeling notations have been developed in HCI communities, often with different goals and different strengths.
Introduction (2/4)Task Modeling are
the hearth of Model Driven Development of UIs
Final User Interface (FUI)
WindowWindowConcrete User Interface (CUI)
AICfacet=control
AbstractIndividualContainer
Abstract User Interface (AUI)
textInputtextInput buttonbutton buttonbutton
AICfacet=control
AICfacet=control
Task & Domain (T&D)
Introduction (3/4)
•A comparative analysis of selected models involving multiple users interaction is provided in order to identify concepts which are underexplored in today´s multiuser interaction task modeling.
Introduction (4/4)
•The goals are:▫To provide a conceptual understanding of the
most significant models involving multiple users and their related concepts.
▫To establish semantic mappings between the different models so as to create a transversal understanding of their underlying concepts independently of their peculiarities.
▫To rely on these semantic mappings to develop a multi-user model editor that accommodates any type of input.
Overview of task models (1/7)
•Our method for uniforming task models consists of four major steps:1. selection of individual task models, 2. identification of the concepts within each
model, 3. representation of those concepts into a
meta-model, 4. consolidation of these meta-models into
one single meta-model (called multi-user interaction model).
Overview of task models (2/7)
•The following criteria were used:▫The task models should be integrated in a
development methodology as a core or side component and tool supported.
▫The task models should be widespread and accepted within the Human-Computer Interaction (HCI) community.
▫The selected models should be supported by theoretical studies to assess their soundness and experimental studies for effective case studies.
Overview of task models (3/7)
•Each model was decomposed into constituent concepts using an entity-relationship method and UML representation.
•A syntactical uniforming has been conducted to provide a single way of referring to different concepts where possible.
•For concepts having different definitions a semantic uniforming is needed (semantic mappings between concepts having different aims and scopes).
Overview of task models (4/7)
•To maximize the semantic scope of the uniformed task meta-model, the union of the concepts present in each particular task meta-models was preferred rather than the intersection.
•We use Separation of concern principle, which assumes that only concepts relevant to a similar domain of discourse should be kept in a particular model, thus avoiding mixing different concepts into a single model.
Overview of task models (5/7)
•Task modeling notations reviewed
AMBOSS GTA (Groupware Task Analysis)
ANSI/CEA-2018 HTA (Hierarchical Task Analysis)
CTT (ConcurTaskTree) TKS (Task Knowledge Structure)
Diane+ TOOD (Task Object-Oriented Description)
GOMS UsiXML
•ANSI/CEA-2018 meta-model
Task Model
+about: URI
Binary
Task
+id: ID+description: string+precondition: string+postcondition: string
input
+name: string+type: string+modified: string
output
+name: string+type: string
concept
+verb: string
step
+name: string+task: string+requires: string+ordered: boolean
applicable binding
+slot: string+value: string
TaskModelScript
+task: QName+model: URI+init: boolean+platform: token+deviceType: token+applicable: string
Task Relationship
sequential Choice
role
+name: string+type: ceaDataType+slot: ceaSlotName
optional iteration
Unary
TemporaryRelationships
agent
theme
instrument
location
Decomposition
Concurrency
1..* 1..*
TaskScript
+init: boolean+platform: token+deviceType: token+applicable: string
1..*
0..*
0..*
0..*
Overview of task models (6/7)
•CTT meta-modelTaskModel
Task
+id: string+name: string+type: string+category: string+frequency: string+platform: string+description: string+precondition: string+duration: time
TaskRelationshipsRole
+id: string+name: string
TemporalRelationshipsDecomposition
Object
+name: strig+type: string+class: string+accesMode: string+cardinality: string
Action
+id: string+name: string+type: string
Binay Unary
Choice OrderIndependence Inerleaving
Synchronization
Disabling
SuspendResume
SequentialEnabling
SequentialEnablingInformqtionPassingCooperative
1..* 1..*
Optional Iterative
0..1
0..*
1..*
1..*
Overview of task models (7/7)
Task Modeling Operators
CTT
Diane+
AMBOSS
ANSI/CEA-2018
GOMS
GTA
HTA
TKS
TOOD
UsiXML
Abstraction
Multi-user interaction meta-model
UsiXML
MDDUI:
A multi-user interaction meta-model (1/8)
GUI
AR
3DUI
VocalUI
HapticUI
PhysicalUI
MultimodalUI
None
Moderate
High
DynamicityColor is useful for identifying the concepts used in the design and implementation of the workflow editor
A multi-user interaction meta-model (3/8)
Marco de Referencia
A multi-user interaction meta-model (4/8)
Paso 1
A multi-user interaction meta-model (5/8)
Paso 2
A multi-user interaction meta-model (6/8)
Paso 3
A multi-user interaction meta-model (7/8)
A multi-user interaction meta-model (8/8)
Conclusion (1/2)
•Task models show a variety of concepts and relationships, we analyzed them in order to identify concepts which are underexplored in today´s multi-user interaction task modeling.
•After the analysis of task models, a multi-users interaction meta-model was generated in order to cover the principal characteristics required to work with multiplicity entities playing a role.
Conclusion (2/2)
• Our meta-model tries to cover the principal aspect required to support group work, it include process, tasks, task operators (including collaboration relationship), actions, objects, resources, groups (as an attribute), organizational units, jobs, agendas, goals and rules (both of them as attributes).
• In a future work, we would like to integrate in our comparative analysis other task models that are focused on multi-users interaction. Also, it would be interested to integrate a task analysis part, until now our meta-model is devoted to task modeling.
Join us
BUAP
http://www.facebook.com/UsiXML
@usixml
For more information and downloading,http://www.lilab.eu
User Interface eXtensible Markup Languagehttp://www.usixml.orghttp://www.usixml.euRegister as a member of the UsiXML End-User Club at
http://www.usixml.eu/end_user_club
Q & A
THANK YOU
Contact
Dra. Josefina Guerrero-García [email protected]
Dr. Juan Manuel González-Calleros
[email protected] @Juan__Gonzalez