From: Perry Valdez

Date: 7 Jul 2005

 

I received an article from SearchOracle.com titled The relational model before Codd. It seemed like they're reconstructing the history of data management!

 

I quote:

 

Following the publication last week of our new Learning Guide to RDBMS Fundamentals, I received some educational feedback from member Edd T. He wrote:

 

Whenever I read articles on RDBMS basics, I'm always surprised to find that most of them begin with Codd's rules and take it from there.

 

In truth, there's a whole body of mathematics--relational algebra--that was invented in the mid-1800's by Augusta DeMorgan that is the real basis for RDBMSs. Of course, both at that time and since, relational algebra is a branch of mathematics that's been used in a number of ways, and not just for RDBMSs. In fact, when it was first postulated around 1845 and continuously enhanced with each succeeding year since then, it was certainly not in the context of database management.

 

What makes Codd's insight so blinding in the late 1960s was his ability to synthesize disparate, cross-disciplinary fields into something of practical use. After being around for a hundred years, this subset of mathematics was joined (no pun intended) with information management.

 

I invite those readers who are interested to ferret out the real fundamentals of relational databases by locating one of the many library books on relational algebra and reading about it. It really puts a whole new perspective on things; you'll never look at a table the same way again, once you know the mathematical underpinnings. We, in our day-to-day work, are the contemporary beneficiaries of those insights."

 

Thanks again to Edd T. for his intriguing historical perspective. If you're interested in more background (and you should be!), check out these links:

 

·   The best introduction to the relational model

·   Learning guide to RDBMS fundamentals

·   Origins of Set Theory

·   DeMorgan's fusion of set theory with relational algebra

 

 

From: Fabian Pascal

 

C. J. Date, who is more familiar with the history of relation mathematics than me, points out that while the relation concept preceded Codd, most of the work was on binary relations. Codd not only shifted the focus to n-ary relations, but invented the relational algebra. Now, it is true that this was done in the context of applying pre-existing predicate logic and set theory to database management, but nevertheless his mathematical contribution to the theory of relations is considerably more than the understatement in the quoted article suggests. But am willing to consider any clear evidence to the contrary.

 

 

From: Perry Valdez

 

So it's Codd and not DeMorgan who invented the relational algebra. Did Codd also invent the relational calculus?

 

 

From: Fabian Pascal

 

According to Date, yes. Take a look at Edgar F. Codd, a tribute and personal memoir. Although the calculus was application of predicate calculus to data management, it nevertheless was, as such,an invention, published in Relational Completeness of Data Base Sublanguages, in: R. Rustin (ed.): DATABASE SYSTEMS: 65-98, Prentice Hall and IBM Research Report RJ 987, San Jose, California: (1972)

For a discussion of the algebra and calculus (there are tuple and domain versions of either, which are equivalent), see chapters 7, 8 in Date's AN INTRODUCTION TO DATABASE SYSTEMS.

 

 

Posted 9/3/05