View article

Despite several years of mashup practice and research, it is still hard to find high-quality, useful mashups on the Web. While this can be partly ascribed to the low quality of the components used in the mashups or simply to the lack of suitable components, in this paper we argue that this is partly also due to the lack of suitable quality models for mashups themselves, helping developers to focus on the key aspects that affect mashup quality. Although apparently easy, we show that – if taken seriously – mashup development can be non-trivial and that it deserves an investigation that specializes current web quality assessment techniques, which are not able to cater for the specifics of mashups. In fact, we believe a mashup-specific quality model is needed.