Object-Oriented Methods: A Foundation, UML Edition (2nd Edition) - Hardcover

Martin, James; Odell, James J.

Hardcover

ISBN 10: 0139055975 ISBN 13: 9780139055973

Publisher: Prentice Hall, 1997

Synopsis

Make sense of object-oriented philosphy--and use it in both information systems development and business process reengineering. Completely revised and updated to comply with UML notation, this book gives techniques for serial, concurrent, and synchronized processing specification.

About this title

From the Publisher

This book presents those concepts and techniques that support almost any system development approach--whether it involves computers, people, or machines. It considers object structure, object behavior and more advanced concepts such as composition, structural constraints, rules, using rules and diagrams, meta-modeling, and power types. Shows how to represent OOA constructs--modeling object structure, modeling object behavior, modeling state transitions and event diagrams, scenarios. Outlines considerations for design, discusses object-oriented programming, and considers object-oriented design and "instant" CASE.

From the Inside Flap

PREFACE

What if there were a standard set of core concepts for object-oriented analysis and design (OO A&D)? What if there were a prescribed set of diagrams for communicating these concepts? Imagine how much simpler our OO A&D world would be—not to mention the world of OO-CASE vendors. We could all speak a common OO A&D language, rather than some dialect based on a particular methodologist or OO-CASE vendor. I'm not sure such an event would ever meet everyone's satisfaction. Furthermore, I'm not sure that one fixed set of diagrams can (or should) express everything we need to say. But, a way to express most of our A&D knowledge is now available. With the OMG's Object Analysis and Design Taskforce (OA&DTF) a standard now exists for:
a common meta-model for maintaining our OO A&D knowledge,
a technique for exchanging this knowledge, and
a notation for expressing our OO A&D knowledge.
A Little History

Grady Booch was the first to organize an OO A&D standardization effort. In the spring of 1993, he asked a handful of major methodologists if they wished to create an OO A&D standard. (Jim Rumbaugh, Ivar Jacobson, Stephen Mellor, Peter Coad, and I were included.) However, assembling these people for even one day proved too daunting as we could only agree on a short breakfast meeting at OOPSLA Ô93 in Washington, DC. Since the idea of standardization was considered undesirable by the majority at the meeting, the standards movement floundered—but not for long.

In October 1994, Rational Software Corporation succeeded in hiring Jim Rumbaugh. With the combined force of Booch and Rumbaugh, Rational decided to produce its own standard approach, then called the Unified Method. Its rollout occurred at OOPSLA '95 in Austin accompanied by singing (Jim) and merriment (Grady). When Ivar Jacobson joined Rational about a year later, the resulting unified approach became stronger and broader. Soon, Rational's approach—now called the UML (Unified Modeling Language)—became a unification of more than Booch, Rumbaugh, and Jacobson. Literally dozens of other methodologists and organizations became part of an effort known as the UML Partners.
Enter the OA&D Task Force

By 1995, OO A&D unification and standardization were certainly in the wind and on June 29, 1995, an old special interest group was revived within the OMG (thanks to Richard Soley and Ivar Jacobson). This renewed special interest group became known as the Object Analysis and Design Task Force (OA&DTF) and was co-chaired by Mary Loomis and me. Its bimonthly meetings are usually attended by 30 to 90 individuals representing dozens of companies. After an initial period of data gathering and discussion in 1995, the OA&DTF submitted its first Request for Proposal (RFP) in June 1996. All submissions to this RFP were required to:
define a meta-model that represented the semantics of object analysis and design models.
define the OMG's IDL (Interface Definition Language) interfaces that enable models or model elements created by one tool to be accessed from another tool. The model access facility had to be either an import facility or a connection facility or both.
provide a set of notations for representing models.
include a "proof of concept" statement, explaining how the submitted specifications have been demonstrated to be technically viable.
accommodate evolution of models and incorporation of additional semantics over time.
Who responded to the RFP?

Those organizations that responded include: Hewlett-Packard, I-Logix, ICON Computing, IntelliCorp, IBM, MCI Systemhouse, Microsoft, ObjecTime Limited, Oracle, Platinum Technology, PTech, Rational Software Corporation, Reich Technologies, Softeam SA, Taskon A/S, Sterling Software (formerly Texas Instruments Software), and Unisys. On January 17, 1997, these companies coordinated their responses, resulting in six submissions. In the months that followed, the effort was unified into a single proposal. On September 1, 1997, all 18 organizations presented the UML proposal to the OMG.
Conclusion

In summary, the OA&D facility is expected to implement the interfaces that would enable tool builders to provide facilities for tool users to:
extend and augment a model produced by another conforming tool.
use a model produced by another conforming tool in the course of producing a different but related model.
Additionally, this RFP resulted in a standard notation for users to express their models.
A few years ago, such a facility was hoped for and dreamed about—but hardly considered feasible. Now, there is a standard defining the core concepts for OO analysis and design and a prescribed set of diagrams for communicating these concepts.

While this standard will not solve all our problems, it will reduce the Babel-ization surrounding OO A&D communications. Furthermore, it will provide a kind of Rosetta Stone for communicating various kinds of OO A&D specifications. We can then put more effort into what we're communicating—and not so much into how we communicate it. Imagine a world where an OO developer can spend less time in religious battles about notation and more time on the quality of communication.
This book was revised in this spirit.

James J. Odell

Edition details

• Publisher: Prentice Hall
• Publication date: 1997
• ISBN 10: 0139055975
• ISBN 13: 9780139055973
• Binding: Hardcover
• Language: English
• Number of pages: 408