Relational: Free at Last

There is no relational database management system on the market today.

Isn't there a $20 billion market consisting of Oracle®, DB2®, SQL Server®, and their brethren? What of them?

They are pretenders. No expert in relational theory says anything on the market delivers in code what E. F. Codd described on paper.

Why not? What would a true RDBMS be capable of that Oracle and its cohort are not? What can we do to make it happen?

Glad you asked.

The progress toward a true RDBMS long ago ceased, and forces of darkness — ignorance and profit, to name two — threaten to turn it back. The promise of Relational Theory will never be realized without effort on the part of those who understand it. If you are such a person, or would like to be, or just want to know what all the fuss is about, I hope what you find here interesting, even provocative.

These pages describe the miserable state of the RDBMS world today. The problem is so big that it's hard to see. It touches many areas: the IT market, the typical level of technical knowledge among buyers and users of RDBMSs, the design of client/server libraries and protocols. Even worse, there is a growing belief that these problems are somehow inherent in the relational model, when in fact it they are the product of widespread ignorance and the proprietary model of software development.

One can hardly escape the loud and pointless acclaim accorded to would-be successors to SQL. A few years ago it was XML. Today it goes by varying “No SQL” names. But replace how, and with what? They do not even attempt to match what SQL databases have been doing for two decades. They offer no theoretical foundation, little if any data integrity, and a more complex, less powerful data model (to the extent they can be said to have one). That is not progress.

On this we agree: SQL is a problem. But inside every SQL DBMS is a true Relational DBMS yearning to breathe free! The way forward is not to discard the relational model. It is to implement it, finally.

Relational Theory offers something unique in database management: a mathematical foundation, an algebra that permits provably correct inferences from the data. Tutorial D offers a way to express Relational queries using a language based, at long last, on Relational Algebra. It has seen scattered implementation, but not, so far, by any widely used RDBMS, proprietary or free.

Open source RDBMSs continue to grow in function and capability. World Domination ™ is possible! But, what kind of world? If we're not careful — and there's no evidence we are being careful! — the open source RDBMSs will only reproduce the same incompatible cacophony brought to us by the proprietary model. It need not be so! A better, more flexible, more powerful, more relational future awaits us. If only we realize the opportunity, and seize it.

Comments? Send email.