Stop Programming and Start Modeling: Developing WorkCentered Semantic Applications

1
Stop Programming and Start Modeling Developing
Work-Centered Semantic Applications
Andrew Crapo Amy Aragones, Jeanette Bruno, Marc
Garbiras GE Global Research

• Semantic Technology Conference
• May 23, 2007

This work was partially funded by the Air Force
Research Laboratory, Wright Patterson Air Force
Base, under contract F33615-03-2-6300.
2
Introduction
Problem How to develop, maintain and extend
work-centered decision support applications
with less effort Approach Use semantic
technology to drive adaptive UIs based on models
of users, their work, and HCI
3
What if we wanted to create application UIs like
these
4
Logistics Planning Application
Customizable displays and learned user preferences
Rich, complex UI design
Integrates with decision support algorithms
Domain Forecasts
Status data, trends
Planning Algorithms
5
COTS Integration with Custom Information

• Synergize COTS app with custom information
  displays

Enterprise Logistics System
6
What would be

• The required level of development effort?
• The risk of user non-acceptance?
• The required level of effort for maintenance and
  extension as the business evolves?

What if we could reduce these by 25-50?
7
What we did to achieve the vision

• Built a semantic model of general work-centered,
  user-interface, and information processing and
  integration concepts
• Created an architecture supporting model-driven
  user-computer-services interactions
• Provided templates and tools to allow
  domain-specific application creation by extending
  the general model

8
Demo (Part I)

• Create New Course 101 Tutorial Application
• Professor Home Vantage
• Simple Student Grade Table
• Demonstration of Learning of User Preferences

9
What we just did
Create new application using ACUITy
Semantic Models Core functionality and properties
are inherited from upper-level models of
human-computer interaction
Declare the contents and properties of the
application
Relate application content to procedural scripts
Deploy application
Users refine content, presentation, layout
ACUITy learns and updates application model
10
ACUITy Adaptive Work-Centered User Interface
Technology
11
Adaptive Work-Centered Support
provide an integrated and tailored support
system that offers support to work in a
flexible and adaptable manner (Eggleston
Whitaker, 2002)

• Open-ended, dynamic decision environments, e.g.,
  logistics
• Principles
• Problem-Vantage-Frame
• Focus-Periphery Organization
• First-Person Perspective
• Approach
• Semantic models captured in the ACUITy
  Problem-Vantage-Frame Ontology

(from Eggleston Whitaker, 2002)
See Recent Papers at http//acuity.sourceforge.n
et/technical_details.html
12
Semantic Model Structure
Work Model

• Work at Hand
• Work processes
• Context
• Information

Problem Focus
User Model
Interaction Model

• When to present information
• What information to present
• How to present information
• When to ask for clarification or guidance

Vantage
Vantage

• User Profiles
• Expertise
• Preferences
• Peers

User
Frame
Frame
13
Semantic Concepts in Professor Vantage
14
Extending the Application Using Templates

• An ACUITy template is a comma-separated (CSV)
  file that
• Captures a portion of the semantic model
• Can be constructed from design patterns
• Is easily viewed and edited, e.g., in Excel
• Is source code control (e.g., CVS) compatible
• Allows reconstruction of a modified model with
  minimal effort

15
Template for Professor Vantage (Example)

• Professor Vantage in App Frame
• Student Grade Table in Professor Vantage

16
Demo (Part II)

• Extend Application Using Templates
• Student Home Vantage
• Student Grade Time Series Graph

17
Semantic Concepts in Student Vantage
18
Demo (Part III)

• Cross-Vantage Student Summary Table
• Properties Mapped Data Table of Name/Value Pairs

19
Semantic Concepts in Student Summary
20
Demo (Part IV)

• Add Student Selector Interaction Object
  (Implicit)
• Modify Student Grade Table to be Student-Specific
• Add Grade Input Form to Professor Vantage

21
Semantic Concepts in StudentSelector
22
Semantic Concepts in Student Grade Input Form
23
Demo (Part V)

• Add an Overview Vantage
• Show Students Grade Average, Number of Grades
• Highlight Averages in Green, Yellow, Red
• Provide Drilldown on Student ID to the Student in
  the Professor Vantage

24
Semantic Concepts in Overview Vantage
25
Many things still do to

• Productization
• e.g. Performance, Java Server Faces

Refactor and extend models e.g. Model user
intents problems, web service as info source
Extend/create richer UI e.g. AJAX, Autolayout,
Rich client
Exploit instance data e.g. Extent peer group
learning, use trends, patterns
Plugin framework e.g. External tools, roles and
permissions, ad hoc data access
26
ACUITy Recap

• Enables automated reasoning about what to present
  and how
• Intelligent info display
• Dynamic content
• Platform for interaction with remote services
• Simplifies UI development and maintenance
• Developers extend the design
• Users finish the design

27
ACUITy on SourceForge http//acuity.sourceforge.n
et ACUITy Forum http//tech.groups.yahoo.com/grou
p/acuity-user Contact us Andy Crapo
crapo_at_research.ge.com Amy Aragones
aaragones_at_research.ge.com Jeanette
Bruno bruno_at_research.ge.com Marc
Garbiras garbiras_at_research.ge.com