Extensions to UML Using Stereotypes

Abstract

The Unified Modeling Language (UML) allows to visualize, to specify, to build and to document the devices of a system that involves a great quantity of software. It provides a standard form for writing the models of a system, covering so much of the conceptual aspects (such as processes of the business and functions of the system) as the concrete ones (such as the classes written in a specific programming language, schemas of databases and software components). In 1997, UML 1.1 was approved by the OMG becoming the standard notation for the analysis and the design oriented to objects. UML is the first language of modelling in which a metamodel in its own notation has been published. It is a strict subset called Core. It is a self-referential metamodel. It is a very expressive language that covers all of the necessary views to develop and to deploy systems. UML is a language that provides three extension mechanisms (Booch, Rumbaugh, & Jacobson, 1999): stereotypes, tag values, and constrains. The stereotypes allow to create new types of elements of model based on the elements that form the metamodel UML extending the semantics of the same one, the tag values are an extension of the properties of an element of UML, allowing to add new information to the specification of the same one, and the constrains are an extension of the semantics of UML that allow to add new rules or to modify the existent ones. The organization of this overview is given in the following way: first, we present the stereotypes according to the standard of OMG; second, we expose the analysis of works that extend UML using stereotypes in diverse real domains; third, we make an analysis of the stereotypes of UML; and we finish giving a general conclusion where we focus ourselves in the distinction of the works according to their inclusion or not of the created stereotypes in the metamodel of UML.