Abstract
There are many different views of quality. These different views can cause conflicts between different stakeholders in the software process. Further, they can cause difficulties for the modelling of software quality. Early models clearly emphasized a technical view of quality at the expense of criteria of quality important in the user's view. This paper outlines an approach to modelling within a localized environment. The approach seeks to reconcile the needs of the different stakeholders. It produces a model which goes beyond the simple hierarchical classification of criteria and metrics, seeking to define the relationships and conflicts that exist between the quality criteria. The technique, known as locally defined quality modelling, has been developed by the author through the participation of six major UK companies in a range of market sectors.