O'Reilly logo

Metamodelling for Software Engineering by Brian Henderson-Sellers, Cesar Gonzalez-Perez

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

4.2 LINGUISTIC AND ONTOLOGICAL METAMODELS

Since modelling languages such as UML have been designed to be widely applicable, they are necessarily domain-agnostic. As well as diagrams such as Figure 4.3 being valid, modellers might also choose to use instance-of relationships within the domain (SUS). Thus, a class diagram such as that shown in Figure 4.6 is also, arguably, valid. (Although a purist strict metamodeller would argue that we have used an instance-of relationship within a layer, in contravention of the rules of strict metamodelling.)

Image

Figure 4.6: Instance-of relationship within a single domain

The two kinds of model represented by Figures 4.6 and 4.3 have been labelled as ontological (or logical) and linguistic (or physical), respectively [6]. The former is defined in terms only of domain concepts (at different abstraction levels) whereas the latter relates those domain concepts to modelling language concepts such as class, entity and relationship (the dominant mode in UML, for instance). Problems occur when these two kinds of metamodelling are not clearly identified; consequently, Atkinson and Kühne [5] analyze various multi-dimensional representations that combine ontological and linguistic modelling styles in one diagram – an example, called by them the “Orthogonal Classification Architecture” [7], is shown in Figure 4.7.

Linguistic metamodelling focuses the alignment ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required