Introduction Object Oriented Databases Management Systems example essay topic

Position Paper: How can Object Oriented Design help the Migration from Relational to Object Oriented database architecture? Introduction Object Oriented Databases Management Systems (OODBMSs) add database functionality to object programming languages. They bring much more than consistently storing programming language objects. OODBMSs extend the principals of object programming languages to provide full database programming capability, while keeping the original language compatibility. The best benefit of this approach is the unification of the application and database development. As a result, applications require less code, use more natural data modelling and code is easier to maintain.

Object developers can write complete database applications with a small amount of additional effort. In contrast to a Relational DBMS where a complex data structure must be flattened out to fit into tables or joined together from those tables to form the structure, OODBMSs have no performance overheads to store or retrieve a hierarchy of related objects. This one-to-one mapping of object programming language objects to database objects has two benefits over other storage approaches: it provides higher performance management of objects, and it enables better management of the relationships between objects. This makes object Dbms better suited to support applications such as world wide web document structures for example, which have complex relationships between data. So how easy is it to migrate to using OODBMSs and how can the principles of object-oriented analysis and design help us with our database design?

This paper looks at research aimed at relational and object-oriented databases and how the latter is being groomed from object-oriented methods. This new wave of database technology is developing day by day and I hope to highlight the current and future issues and the work that still has to be done. Section 1 the current major type of database in industry - relational databases. Section 2 looks at object-oriented design Section 3 relational to object-oriented databases 1 Relational Database Arguments Since the publishing of (Codd 1970) a large amount of research into the theory and implementation of relational databases (RDBs) has taken place and most of the databases implemented have been RDBs. They allowed users to understand a databases' contents more clearly, based on normalisation. Dr. E.F. Codd, the originator of the relational data model, even published a two-part article in 'ComputerWorld' (Codd, 1985) that listed 12 widely used rules for how to determine whether a DBMS is relational and to what extent it is relational.

However, during the 1980's, the limitations of RDBs have become apparent. They failed to provide facilities for representing the complex structured objects that come about in software engineering. They failed to provide facilities for representing data such as images and fail to support integrity constraints effectively (Celle and Bertossi 2000). Although relational databases are still adequate for the majority of todays business needs, how long can big organisations continue to ignore them before they are left behind? 2 Object Oriented Design "The key principle of object-oriented design is that each program object should correspond to an object in the real world".

(Biddle et al. 2002). This always has been and, as can be seen above, still is, the underlying principal of object-oriented design. However, since its inception in 1960, the depth of knowledge of object oriented technology has matured over the years. In fact, object-oriented application development is now the mainstream of information technology. Most notable of these standards is UML (Unified Modelling Language) for object oriented analysis and design (Bennett et al.

1999). The UML specification represents the best practices in the object technology industry. Release 2.0 is currently being worked on by the Object Management Group (OMG) which is the main organisation for object technology standards. Although precise details have proved elusive, browsing the UML section of their website gives the impression that advances in UML development do not remain static for long. 3 Relational to Object Oriented Databases The first OODBs appeared in the late 1980's. See (Martin 1993) for a full list.

Since then there has been much debate on the best way to develop them forward. Many attempts have been made to standardise the OODB (e.g. Atkinson et al 1995) but a clear guideline has yet to emerge. Most applications have yet to leave RDBs behind in favour of OODBs and there is much feeling that, while applications would certainly benefit from OODB support, they are impeded by limitations. For example, many OODBs are closed and restricted to one language unlike relational database systems (Carey et al 1994).

However, the biggest and probably the most obvious factor is cost. If you switch a large-scale application to read from an OODB then all the old RDB code will have to be re-written. Have the factors counting against OODBs meant that they will never be seen as the main type of database in industry? I have seen at least three recent articles in Newspapers and Magazines that are suggesting that they have had their day and failed to deliver (e.g. Wilson, 2002) On reading (Bertino et al) I found the best journal in my opinion that would be very beneficial to someone looking on how to migrate. There is talk of database schema, object evolution, relationships and classes that I believe can be related to RDB principles. Conclusions and Future Work Simply by definition, object oriented design and object oriented databases, it is obvious the two are not mutually exclusive.

The fundamentals are the same but at what stage do the two separate themselves? "In many class-based object-oriented systems the association between an instance and a class is exclusive and permanent. Therefore these systems have serious difficulties in representing objects taking on different roles over time. Such objects must be reclassified any time they evolve (e. g., if a person becomes a student and later an employee). Class hierarchies must be planned carefully and may grow exponentially if entities may take on several independent roles. The problem is even more severe for object-oriented databases than for common object-oriented programming.

Databases store objects over longer periods, during which the represented entities evolve" (Gottlob et al. 1994) With this in mind, it would be ideal to research at length how the work of (Bertino et al.) enhances or rejects the questions that this statement poses. Can object-oriented design theory produce a mechanical process that everyone can follow to convert their relational database model to an object-oriented one? It took (Codd 1985) fifteen years to produce his famous twelve rules of measurement after his first paper, are we not well overdue an attempt at similar guidelines?

Bibliography

o Atkinson et al. (1995) The Object-Oriented Database System Manifesto.
University of Leeds. o Bennett, S. and Mc Robb, S. and Farmer, R. (1999) Object -Oriented Systems Analysis and Design using UML.
McGraw-Hill o Bertino, E. and Guerrini, G. and R usca, L. (Date Unknown) Object Evolution in Object Databases. o Biddle, R. and Noble, J. and Temper, E. (2002) Metaphor and metonymy in object-oriented design patterns.
Victoria University of Wellington. o Carey et al. (1994) Shoring up Persistent Applications.
University of Wisconsin. o Celle, A and Bertossi L. (2000) Querying Inconsistent Databases.
University of Chile. o Codd, E.F. (1970) A Relational Model for Large Shared Data banks.
Communications of the ACM, 6, p. 377-387 o Codd, E.F. (1985).
Is Your DBMS Really Relational? and Does Your DBMS Run By the Rules? ComputerWorld, October 14 and October 21. o Gottlob, G. and Sch refl, L. and Rock B. (1994) Extending Object-Oriented Systems with Roles o Martin, J.
Principles of Object-Oriented Analysis and Design. Englewood Cliffs, NJ: Prentice Hall, 1993.
o OMG Object Management Group website at web o Wilson, E (2002) Object Databases find a New Lease.