Robin Bloor and relational database
Published 21 Feb 2017
Table of content
“ When you put your car in a garage, you put it as one thing complete with its capabilities. You don’t put your car in the garage and store its steering, transmission, signaling and lighting functions somewhere else”. (Robin Bloor, para 4)
Robin Bloor, in his article, “ The failure of relational database, the rise of object technology and the need for the hybrid database” cites a very common example, to demonstrate the state of relational database. In this article he categorically analyses the evolution of relational database and the object oriented technology. He discusses the advantages and disadvantages of both the technical advancements. At the end, after citing the shortcomings of both, he puts forward the new development which tries to overcome the shortcomings of both these technologies in one product known as ‘ object-relational database’ or ‘cache’.
This essay , is an effort to opine about the validity of whatever Robin Bloor has to say in this article, and decide if his judgments are purely commerce oriented or have technical justifications also. Background It is a matter of general knowledge that the need to analyze and sort the data of American census, gave birth to the modern day computer in the 1940s. Since then, the way databases have been handled has constantly changed for the better. Till the personal computers came in 1970s, databases were analyzed by batch processing method, which was very slow, cumbersome and expensive. In 1970, American scientist Edgar F. Codd presented the ‘relational model’ to analyze and sort data, when he was working at IBM. Later, in the last decade of the twentieth century, the concept of ‘object’ evolved. Robin Bloor examines the evolution of relational database and object oriented technology, and proposes a solution which overcomes the limitations of these two technological developments. Contents of Bloor’s paper in brief While discussing ‘relational database’, Bloor mentions that “ one of the foundations of relational theory is that the data and the program that uses it should be independent of each other.
This was and is at odds with the whole idea of object technology”. He gives the example of a car taken as an object, and refutes the idea of data and its associated processes being independent of each other. He strongly advocates that data and its associated processes cannot be and should not be separated. Bloor cites one more limitation of relational database. He says that , “ Relational databases are more constrained in their capabilities than most would suspect. Storing and representing some fairly common data structures, can be very difficult”. He takes the example of a “simple ordered list of bus stops. Relational databases only hold tables as unordered lists and can retrieve an ordered list only if a specially built index is added. An object database does not need index.” ‘Performance overhead’ is another concern of Bloor. He says that, “ in relational database, to assemble data, developers have to JOIN one table to another to another to another. While retrieving, database runs optimization routines, to determine the best way to gather data and then retrieves it. This process often takes a long time.
This does not happen with object database.’ “Impedance mismatch’, is another problem cited by Bloor. He says, “ The problem surfaced when OO languages rose to dominance. The differences between the way data was handled by OO languages and the relational database. In reality most databases are not completely normalized when they are implemented , but even so, the problems of impedance mismatch occur”. Conclusively, Bloor suggests a possible solution, by mentioning that, “ the approach involves providing the databases, with a mapping laye r through which developers access the database. The mapping layers should be based on open standards to resolve the impedance mismatch problem. Database calls could then be made either in SQL or direct requests to an object class or collection of classes. The mapping layer would translate these calls, into physical data requests, to the database to retrieve the data. This would obviate the impedance mismatch.”( Robin Bloor)
Relational database A relational database is a database that conforms to the relational model. It could also be defined as a set of relations or a database built in RDBMS. An RDBMS is many times called a relational database, but strictly speaking, Oracle, Microsoft SQL server, MySQl are not relational databases. …. they do not fully conform to the relational model.”(relational database) It is common knowledge that query languages which were developed purely on the relational model, never found the commercial acceptance. Object database In object oriented database, information is represented in the form of objects. When database capabilities are combined with object programming language capabilities, the result is an object database management system. Object databases are generally recommended when there is a business need for high performance processing on complex data. Most object databases also offer some kind of query language, allowing objects to be found by more declarative programming approach.
Access to data can be faster, because joins are often not needed. The pointer based techniques of object oriented databases proves to be slow for general purpose queries. Another disadvantage object oriented database is lack of interoperability with great number of features. ( OODBMS) A closer look at Bloor’s article and the definitions of relational database and OODBMS in an encyclopedia reveals that Bloor is very correct in whatever statements he has made about relational database are correct and they are very much reflected in other sources also. The biggest proof of Bloor’s claim on failure of relational database is that even after 40 years since the invention of relational model by Codd, there is no query language strictly adhering to it. Whatever languages were developed in such a way, never met with commercial success. ( Edgar Codd) An interesting point to be made here is that, even when Codd presented his relational theory, IBM , where he was working, never gave much attention to it unless, companies like Oracle and others used Codd’s relational model to develop query language. (Edgar Codd) This suggests that IBM foresaw the difficulties in pure implementation of Codd’s model. 40 years of history has proved this right. Bloor’s article and technology guide of Intersystems Bloor harps upon the shortcomings of relational database and Object based database in his article. He synthesizes these difficulties and proposes a solution in terms of ‘object- relation database’. His company, Intersystems, in its technology guide, cites the same reasons as Bloor has cited in his article, especially impedance mismatch, and names the solution as ‘ cache’. The article fully echoes the technology guide.( technology guide)
MSIMT courses and commerce All academics, including this course, are turning towards commerce, because the whole world is in the grip of ‘creation of wealth, still more and still more of it.” The ultimate aim of all activities is ‘generation of wealth’. In the history of mankind, never before, was such a mad rush for wealth witnessed, as witnessed in the later part of 20th century. This is perhaps the reason why academics is also getting closely related to commercial applications and its feasibility. Conclusion Robin Bloor is a smart technocrat with a commercial vision. He has carefully analyzed the weak points underlying the relational database and object based database, and shown a practical workable solution to the difficulties faced in implementation of the two. His paper is not just a commercial for cache, but a very legitimate guide to analysis of an existing problem, and development of technically feasible and commercially viable solution, ultimately leading to opening of new business opportunities and helping the IT community. The future history of IT may take the innovation of ‘cache’ by intersyatems, as a ‘landmark discovery that changed the way databases were handled’.
- Robin Bloor, The failure of relational database , the rise of object technology and the need for hybrid database, cache technology, cache innovations, Intersystems
- Relational database, wikipedia the free encyclopedia, 8-11-2006 Retrieved on 4-12-2006
- Object Database, wikipedia the free encyclopedia, 23-11-06, Retrieved on 4-12-06 < http://en.wikipedia.org/wiki/Object_database>
- Edgar F.Codd, Wikipedia the free encyclopedia, 4-12-2006, Retrieved on 4-12-2006 < https://en.wikipedia.org/wiki/Edgar_F._Codd>
- Technology guide, cache technology guide, cache innovations, retrieved on 4-12-2006