James Thomas, Manager, Systems Integration Team

1
Indiana University Has Embarked on a Journey to
Create the Next Generation Web Portal

• James Thomas, Manager, Systems Integration Team
• John F. Walsh, Director, University Information
  Systems

2
Attribution

• Jim Gorman, University at Buffalo
• Pioneer
• Michael Handberg and Bob Kvavik, U. of Minnesota
• Marvelous vision
• Randy Ebeling, U. Texas, Austin
• Just do it!
• Ed Lightfoot, U. Washington
• A brilliant moment of lucidity
• Howard Strauss, Princeton
• An integrated thinker and my hero!

3
Summary

• Background
• Buy vs Build
• Getting started
• Methodology
• User centered
• Challenges
• Best practices

4
Background

• May 1998 - IT Strategic Plan for IU
  http//www.indiana.edu/ovpit/strategic/
• Action 37 Common Interface to administrative
  systems
• Action 38 Thin client and multi-tiered
  architectures
• Action 44 User-centered design

5
Background

• March 2000 JAD Session
• Single Sign-on and Authentication
• 24 X 7 availability
• Role-based one stop shopping
• Usability tested interface
• Enterprise Application Integration

6
Background

• Opportunity Knocks
• PeopleSoft ERP implementation for HRMS and
  Student Administration
• Re-engineering numerous legacy applications to be
  web enabled
• Implementing a web services, object-like model
  to enterprise application development under the
  IT Strategic Plan

7
IUs Vision

• Unified front end to IU services
• Single sign-on and authentication
• Role-based customization
• Usability tested personalization
• Application integration (more than just flaming
  logos and spinning icons)
• Adaptive user interface
• Universal access 24 X 7
• Completely user-centered environment

8
Buy vs Build Decision

• Summer 2000 Should we buy a portal?
• Portal market was still maturing
• No complete portal solutions
• Gartner Group white paper (ITExpo 2000, Oct.
  2000) confirmed our analysis
• Portal market is high risk, crowded and due for
  shakeout mid-2001
• Dominated by very small independent software
  vendors
• In contrast, they make up a very small part of
  larger vendors revenues (IBM, Oracle, CA)
• Lack of complete portal solutions

9
Buy vs Build Decision

• Should we build?
• Is it too late? Consider your development
  resources
• Consider building with a partner
• Consider buying a framework (Oracle, Microsoft,
  CA, etc)
• Got Java? Consider the JA-SIG
• Benefit from owning your portal

10
Getting started

• Obtain support from the top
• Communicate vision as broadly as possible
• Establish clear and detailed plan of action
• Engage students, faculty, and staff
• Create process for prioritization
• Focus on big bang for little bucks services
• Plan for iterative development to avoid analysis
  paralysis

11
Why call it next generation?

• Not just an information portal
• Flexible and responsive to change
• Distributed model for service and content
  providers the portal is the framework
• CBD/Web Services methodology and standards
  approach
• Shared component/service infrastructure - EDEN

12
OneStart EDEN
OneStart
User Interface
Customized
Personalized
Adaptable
Desktop
Application Delivered
Channels
Applications
Other Content
Services
Record Keeping
Application Services
Infrastructure
EDEN
13
Review Methodology

• Researched existing methodologies at IU and other
  places
• Needed a methodology that supported the portal
  vision
• Created living methodology document

14
Methodology Overview

• Identify the requirements of the application
  define the business processes or functions
• Map the requirements to business objects and
  usage scenarios.
• Map business objects to components or services.
• Avoid implementation details during this
  process!!
• Focus on business requirements.

15
Methodology Challenges

• Must have well defined business requirements
• May require training of IT staff
• Must implement methodology and standards for
  developing, maintaining, and publishing
  components/services
• Requires cooperation and communication across
  development teams

16
What are Components?

• A component is a specific piece of enterprise
  functionality that can be reused in future
  development and integration (ie Explorer Tree,
  CalcGPA, ValidateAccount, GetLocalAddress)
• A key difference between components and objects
  is that components clearly separate specification
  from implementation. This allows for easier
  re-use.
• Components have published interfaces to
  facilitate integration. These interfaces should
  be independent of their implementation.

17
Benefits

• A CBD and Web Services approach gives you...
• a repository of reusable business functions
• the ability to replace specific business
  functions without affecting the rest of your
  application
• rapid development capability by assembling
  existing components and services
• improved agility, flexibility, and scalability

18
Granularity Decision
19
Build a User-Centered Portal

• 1. User Participation Advisory groups
• 2. Vigorous Usability Testing
• Conducted portal study of MyYahoo, MyFidelity,
  MyExcite
• Conducted navigation testing using portal
  prototype
• Arranged accessibility testing sessions with
  visually impaired students
• Performed testing of personalization features and
  options

20
Portal Challenges

• Single sign-on
• Integrated Workflow
• Getting IT staff buy in (no portal or too many)
• Prioritization of services
• Screen real estate (default channels)
• Portal Taxonomy
• Creating stickiness If you build it, they
  will come. Will they come back?

21
More Portal Challenges

• Myth that browsers are a universal client
  (browsers - IE, Netscape, Mozilla, Opera, etc.
  OSs Windows 95/98, NT, 2000, Mac OS, Unix)
  ARGHHH!!
• Accessibility for a diverse group of clients
  legal and ethical issues
• Privacy policy issues
• Integration of vended software

22
Services

• Email Finanacials
• Calendar News
• Student Self Service Games
• Course Management Library
• Institutional Reporting Links
• Mobile bookmarks Entertainment
• Secure Telnet client
• (create student stickiness)

23
Future services

• Single sign-on Top 10 channels
• Workflow Engine Whats new?
• Faculty services Desktop Tools
• Online Community Wireless access
• HR E-docs Alumni services?
• Service integration

24
OneStart.iu.edu

• It all begins with OneStart
• Staff pilot release May 11th, 2001
• Student pilot release June 28th, 2001
• General availability to students, faculty, and
  staff scheduled for Sep. 28th, 2001

25
Top 10 Best Practices (in Jamess opinion)

• 10. Have Fun!!
• 9. Focus on a few big bang services first with
  some sticky services thrown in
• 8. Define basic interoperability standards
• 7. Assume an iterative development process
• 6. Conduct frequent usability and accessibility
  testing
• 5. Obtain a shared vision across the enterprise
• 4. Get students excited about the project!
• 3. Have a strategic plan
• 2. Work to have only one portal
• 1. Obtain ongoing executive level support

26
Demo

• OneStart.iu.edu
• Want to talk portals?
• jthomas_at_indiana.edu