Looking for something else?
This year marks another milestone for the unified modeling language -- the official release of the UML 2.5 revision. Over the years, some experts say, UML went through some growing pains. At first, the language was deemed a successful way to get developers on the same page, but later renditions, such as UML 2.0, had some industry insiders feeling that UML was too complex.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Regardless of one's perceptions of various versions of UML, it's hard to deny the impact it's had on modern application development. Companies of all sizes have used the standardized language to create model structures.
Leading the UML 2.5 revision task force is Microsoft architect Steve Cook. In this Q&A, Cook discusses the latest elements of the UML 2.5 revision, what architects can expect from the new release and what the update means for the industry.
What prompted the UML 2.5 revision?
Steve Cook: The goal of UML 2.5 is simplification and improvement of the language specification document, rather than any fundamental change in the language itself. Earlier versions of UML 2 suffer from an excessively complex specification, which contains much redundancy and inconsistency and is difficult to understand. As a consequence, UML tool vendors have had difficulty in interpreting the meaning of many areas of the standard, leading to differences in implementation and problems with interoperability between tools. The UML 2.5 effort has greatly simplified and clarified the specification document while leaving the language itself almost entirely unchanged.
What are some of the main changes architects can expect?
Cook: In time, as UML tool vendors adopt the new specification, they will converge on a common understanding of the language, and so, interoperability between tools will improve. UML 2.5 also introduces for the first time a formal specification for the interchange of diagrams as well as models, and as tools adopt this capability, models with diagrams may be interchanged between tools from different vendors.
How will these additions or modifications change the work architects do?
Cook: UML itself has not changed, so architects may continue to use it as before. However, the new interoperability capabilities will, over time, offer increased flexibility in choice of tools, and should help to avoid vendor lock-in.
What do these additions mean for the industry?
Cook: As UML 2.5 is more clearly and precisely specified than its predecessors, there will be fewer differences between different implementations, thus providing a more stable and well-defined platform for model-driven development.
What additional modifications do you think will be made in the future?
Cook: Simplifying the specification of UML 2.5 has allowed the team to identify many areas where the language itself could be simplified or improved. Before 2.5, the complexity of the specification document was such that making improvements would be intractable. It was not the goal of 2.5 to simplify or improve the language, but the work done in 2.5 provides an excellent foundation for future simplifications and improvements.
Further work needs to be done to prioritize such improvements and create a roadmap for enacting them. However, for a standard language such as UML, it is important that it remains stable and backwards-compatible so that existing assets are not devalued as the language evolves.
About the author:
Maxine Giza is associate site editor for SearchSOA.com. She can be reached at email@example.com.