Convention Center Ballroom B-C
Session A
Somewhere Between Anarchy and Monarchy: Choosing the Appropriate Process for Your Next Project.
Moderator: John Daniels, Syntropy, Ltd.
The rise of the agile process movement has received a good deal of attention lately, but many questions remain regarding when such process models are appropriate. Determining which process model is the right one for your project is a critical factor in success. In this panel, the topic of when and why a given process model should be applied will be discussed. To maximize the value of the discussion, panelists will be asked to respond to specific questions based on real-world issues.
Scheduled panelists:
Ron Jeffries, Object Mentor, Inc.
Craig Larman, Valtech
Ivar Jacobson, Rational Software Corp.
Alistair Cockburn, Humans and Technology
Jim Highsmith, Information Architects, Inc.
Bob Marcus, Rogue Wave
Convention Center Ballroom A
Session B
Papers: Languages I
Chair: Joseph Kiniry, California Institute of Technology
The papers in this session are excellent examples of theory successfully applied to OO problems. Findler and Felleisen address contract soundness with respect to annotated Java in its various incarnations. Their results will be relevant to anyone applying formal specification to Java using tools like iContract, JML, and Jass. Ancona, et al., solidify the type theoretic and algebraic foundations of Javas semantics, a highly pertinent area given the popularity of that language. Finally, Bachrach shows how to do macros right in Javaà la Dylan and Lisp, with several new twists.
Contract Soundness for Object-Oriented Languages
Robert Bruce Findler, Rice University
Matthias Felleisen, Rice University and Northeastern University
A Core Calculus for Java Exceptions
Davide Ancona, DISI - Università di Genova
Giovanni Lagorio, DISI - Università di Genova
Elena Zucca, DISI - Università di Genova
The Java Syntactic Extender
Jonathan Bachrach, MIT AI Lab
Keith Playford, Functional Objects, Inc.
Convention Center Ballroom D
Session C
Practitioner Reports: Distributed Systems
Chair: Gerard Meszaros, ClearStream Consulting
Distributed architectures are both a blessing and a curse: With careful planning and design, they can help with reliability and scalability, but getting it right is no small challenge. In this session, two reports describe how distributed systems were successfully developed to provide scalable business applications, and the third author presents an approach for certifying performance of a distributed system.
SPCQuote: The Evolution of a Client Application to Distributed Service Processing
Karen Hope, The St. Paul, Inc.
John Finegan, The St. Paul, Inc.
The Business Foundation System (BFS) writes and manages small business commercial insurance policies for The St. Paul Fire & Marine insurance companies. The original system, deployed in 1996, was a client-based VSE Smalltalk application. By the time of its third release, it supported 120 internal users in three locations. Since its deployment the system has had over 30 subsequent releases to increase and improve its functionality, the most notable of these occurring with BFSs SPCQuote release in April 2000. SPCQuote significantly changed both the system architecture and its development environment and satisfied many disparate key business objectives. Today the system supports over 7500 users in over 3200 locations. It manages over 120,000 policies with policy premiums totaling over $500,000,000. This paper describes the evolution of the smart-client VSE Smalltalk application to an n-tier distributed service model supported by two Smalltalk dialects, Java servlets running within Websphere, and XML as a means of data abstraction.
Business Automation with Distributed Objects
Jason Rogers, Ontario Teachers Pension Plan Board
Dean Mackie, Ontario Teachers Pension Plan Board
Angus MacArthur, Ontario Teachers Pension Plan Board
A large financial institution was faced with the challenge of having to process twice the normal yearly workload without increasing their workforce. The challenge was met by reusing components of an existing object-oriented application in a distributed, fully automated configuration. That solution is discussed, along with subsequent system architecture evolution, resulting improvements to development practices, and the ripple effects of changing the culture of senior management that it was designed to serve.
Certifying Component Performance in Synchronous Distributed Client/Server Systems
Dong-Lih Denq, AT&T-Network Services
Irwin Dunietz, AT&T-Network Services
John Eddy, AT&T-Network Services
Willa Ehrlich, AT&T-Network Services
Don Gerth, AT&T-Network Services
Brian Larson, AT&T-Network Services
Geetha Sivaprasad, AT&T-Network Services
This paper describes an approach for certifying performance of a distributed system that utilizes synchronous interprocess communication. The system is composed of multiple managed objects implemented in object server processes (i.e., components) that communicate via CORBA middleware. Component performance is certified empirically through pairwise testing that emphasizes interactions between a component and its lower-level servers. The applicability of this approach is demonstrated for an AT&T provisioning system.
|