ON THE DANGERS OF REPEATING HISTORY
with C. J. Date and Fabian Pascal

 

 

 

From: FA

To: Editorial

Date: 14 May 2004

 

I am reading for my theory of databases exam, and one of the topics our lecturer suggested to revise is: "What did Date mean by 'the dangers of repeating history', with reference to possible extensions of the relational model (e.g. to accommodate complex data types and temporal data)?" Can you help me out with this one?

 

 

From: Fabian Pascal

To: FA

 

I can give you a general response and will forward your message to Chris for a more specific reply.

 

In general, the IT industry is ignorant of data fundamentals (the foundations of the database field) and operates like the fashion industry, constantly promoting "new" fads which very often are old technologies that failed and were made obsolete. One of the best, but not the only, examples is XML databases, which is nothing but the old hierarchic data model that we discarded decades ago. Other examples are object databases, TSQL2, REF types!!!

 

The point is that if you don't know the history and foundations of your field, you are likely to repeat all the mistakes of the past.

 

 

C. J. Date Responds:  To answer your question fully would be a huge undertaking! Let me just give one particularly egregious example. Among the many reasons why prerelational systems like IMS and IDMS were so difficult to use and error-prone—in fact they couldn’t be directly used at all by end-users, you had to have programming expertise—is the fact that they relied so heavily on pointers-chasing. So Codd very deliberately excluded pointers from the relational model. But now the SQL standards committee and the major SQL vendors, under the influence of the object database community, have added pointers to SQL: a dreadful move, and one that clearly shows that the people responsible don’t know history and don’t know the relational model (or if they do know the relational model, they don’t know why it is the way it is). It’s also one that means that SQL now has to forfeit whatever claim it might ever have had to being truly relational.

 

Good luck with your exam.

 

 

Posted 08/20/04