The process of logical database design consists of four phases: view modeling, view integration, schema optimization and schema mapping. View modeling is defined as the modeling of the usage and information structure perspectives of the real world from the point of view of different users and/or applications. The view integration phase combines these views into a single community view which is subjected to further optimization and mapping. As a result, instances of users' model may be altered and application programs transformed. This paper proposes a scheme for view representation which will facilitate the process of view integration. This is done by enhancing the data abstraction framework proposed by Smith and Smith. It takes into account the instance-level interrelationships among data and the identification of instances via these interrelationships. The usage perspective is incorporated as rules and assertions about schema- and instance-level insertion and deletion. The problem of view integration is briefly addressed. Valid transformations of views are indicated as a part of the integration process.