Unformatted text preview: jects, object identity, behavioral aspects of objects, re-usability (of designs or software code), and modularity and implementation independence of subparts of the (software) design objects. With the advent of new database applications such as geographical, engineering design, multi-database, o ce automation, etc., the database modeling requirements have dramatically changed. These modeling applications demand sophisticated abstraction mechanisms to capture complex organization of data, intricate inter-object relationships, means for variety of constraint speci cations, etc. Similar to object-oriented systems, semantic data models1 (SDM) are very powerful in specifying database structures and schema at varying levels of abstractions, and hence have a strong potential to be the object-oriented data model. SDM features have been found to be extremely useful in the modeling and development of data and knowledge based systems, as a design technique and documentation tool. SDMs are also very powerful in expressing the relationships among data and their semantics at a higher level of abstraction. Several forms of constraints can also be expressed in these models. As opposed to behavioral encapsulation in object-oriented paradigm, semantic models tend to encapsulate structural aspects of objects. In a di erent way they are also capable of modeling complex structures, derived schema components, type constructors, inheritance, etc. SDMs tend to hide the implementation speci c details while focusing on the logical organization of the databases. In that way they serve as a great conceptual design tool as required by the three level architecture of databases. Although current SDMs o er a rich set of modeling features, they lack some of the desired properties for an object-oriented model. Since the object-oriented paradigm already possess some of the useful features essential for next generation data
See 63] for a detailed survey of SDMs. A detailed reading and a tutorial on SDMs may also be found in 39].
1 6 and knowledge base applications, it is believed that semantically rich and expressive conceptual models can be built for future applications by extending the functionality of the SDMs in the direction of object-oriented programming. The resulting model would likely become similar to abstract data types giving rise to the concept of abstract database models (ADM). The idea of abstract data types (ADT) has been extensively used by the objectoriented (OO) programming and database community during the recent years. The need for programming with taxonomically organized data that are encapsulated with operations, and are arbitrarily complex structured has led to the concept of objectoriented programming. The OO paradigm is not only able to manage complex and hierarchically structured data, it is also able to express the dynamic aspects of data with the introduction of the concept of behavior and methods. The notion of inheritance in OO programming made it possible to maximize sharing of objects and re-usability. The notion of encapsulation has given this paradigm the strength of modularity and implementation independence, and hence provided a higher level of abstraction. The introduction of behavior and methods provided another dimension of logical independence and allowed the user to view objects in terms of the characteristics exhibited through their interfaces. Inheritance in OO languages allow sharing, and re-usability of software components, while speci city is facilitated by the idea of overriding. While encapsulation of methods enhances the idea of physical independence by hiding the internal structure, data, behavioral implementation, etc. from the outside world, inheritance provides implementation independence as an orthogonal mechanism. In all, the ADT is the key to all these nice features. We, among others, believe that combining OO concepts with databases will result in clear bene ts. There is currently signi cant interest in various forms of OO database (OODB) systems. Researchers have made a signi cant stride during the last few years in combining these two concepts. Among others, the works reported in 38, 37, 27, 35, 5, 34, 17, 70] are some of the representative ones to which we will focus our attention. The Data and Knowledge (DK) model 38] attempts to extend the ER like models in the direction of OO paradigm and tries to capture dynamic behavior, metaknowledge about data and constraints on the data in a simple setting and appears to be the right approach towards a rich modeling tool. But it lacks quite a few features that are essential for the purpose of abstract database modeling. Namely, it only 7 uses aggregation 2] type objects and is thus value based. Hence it lacks the strength of modeling complex objects and obviously other important features. The DERDL model 37] also captures a few features in our direction. Particularly, it proposed the idea of dynamic objects and relationships, and similar other dynamic properties of database that is close to the spirit of derived schema components in SDM 36, 63, 39]. Since the underlying concept is value based, it also has the shortcoming...
View Full Document