View article

Although goal modelling is a recognised research area, only few empirical studies are reported. In this work we present an experiment where the quality of two goal languages – i* and KAOS – is investigated by means of the semiotic quality framework. We believed that a high quality language would contribute to effective and efficient modelling, and result in high quality models. But the experiment showed that model quality much depends on the particular language characteristics with respect to a given context. The experiment indicated weak and strong properties of goal modelling languages. For researchers, the findings point out possible language improvements. For practitioners, they can facilitate decisions about language selection and use.