Home > Articles > Software Development & Management > UML

  • Print
  • + Share This
This chapter is from the book

This chapter is from the book


The Unified Process has its roots in the work that Ivar Jacobson did at Ericsson in the late 1960s. Jacobson and his colleagues modeled a very large telecommunications system using layers of "blocks," with the lower layers serving as the foundation for subsystems at the higher layers. (These blocks have parallels with what are now known as components.) The team built those low-level blocks by exploring what they called traffic cases (now called use cases in the UML); all subsequent analysis, design, implementation, and test work was driven, to a greater or lesser extent, by those traffic cases. Other diagrams that are now key aspects of the UML evolved from the Ericsson work, including sequence diagrams (called object interaction diagrams at the time) and activity diagrams (which Ericsson called state transition graphs). A document called the software architecture description contained the key material that facilitated communication among developers and between developers and customers.

Jacobson went out on his own in 1987, starting a company he named Objectory AB. He and his associates spent several years developing Objectory, which was both a process and a product, as is the Rational Unified Process (discussed next). His book Object-Oriented Software Engineering (Addison-Wesley, 1995), which described the Objectory process in detail, was regarded as a landmark in the object-oriented (OO) community; it was the first book by a major methodologist to put forth the idea that the customer's requirements, as expressed within use cases, should be the most important driving force in software development. The genesis of the Unified Process's emphasis on architecture (see "Architecture-Centric," on the next page) were also on display in the book. The full on-line version of the process appeared in conjunction with the book in 1995.

Not long after that, Rational bought Objectory AB. Jacobson and his considerably larger set of colleagues set about expanding the areas that the Objectory process didn't address in depth, such as project management and development tools. Grady Booch and Jim Rumbaugh were already on board at Rational—Booch almost from the beginning, and Rumbaugh from late 1994—and the gentlemen who became known as the "three amigos" were among the leaders of the effort to build what eventually became the Rational Objectory Process (ROP), in parallel with their expansion of the Unified Method into what became the Unified Modeling Language.


See Chapter 1 of UML Explained for a brief history of the UML.

While the work on the ROP and UML was going on, Rational was busy acquiring and merging with a number of other companies that made software development tools. These tools added value to the ROP product in the form of requirements management (Requisite's tool became RequisitePro), general-purpose testing (SQA's software has been expanded into several discrete tools), and other areas such as performance testing, configuration management, and change management. In 1998, Rational changed the name of the product to the RUP; the differences between the Unified Process (which is at the conceptual core of the RUP) and the RUP as a product are described in Appendix A.

  • + Share This
  • 🔖 Save To Your Account