From: RR
To: Editor

 

I first learned about relational database systems in the early 80's while in university. One of the first books I obtained was Chris Date's AN INTRODUCTION TO DATABASE SYSTEMS (4th edition). Like most other students, I learned the material but did not fully appreciate it. It wasn't until a few years ago that I began to explore it again, but with a different motivation. The vast majority of my experience has been in the OO arena. Although an improvement over structured/functional programming, it falls far short of where we need to be heading. An approach that really struck me when I first encountered it (again, in university) was Logic Programming. For the first time I saw the possibility of specifying what you wanted rather than how. In my current position I have been developing a language that has a Prolog flavor, but allows one to specify "what" using binary decision trees. Of course, by using binary decision trees I have introduced an awareness of sequence and therefore a degree of "how". But the motivation is the same; we have to move away from writing procedural code. It is just too ad-hoc and therefore error prone. This desire was what led me to my current exploration of the relational model and integrity constraints.

 

I have done quite a bit of reading, mostly Chris Date's books, and really do believe that you guys have it right. My primary question for you is how do we make this vision a reality? Are there any efforts being made to organize (perhaps form a company) to develop the relational system Chris Date has describes in his books (as in THE THIRD MANIFESTO), one that is true to the relational model?

 

 

Fabian Pascal Responds: Good for you.

 

Regarding programming, I don't have much to say as it is in the application, not database domain. I am sure that logic is very useful in most software endeavors. Whatever does not have a scientific foundation (like OO) is generally questionable.

 

We're pretty sure we are right too. Unfortunately, the american culture in general, and the business culture in particular, make it very hard for the right thing to succeed, the propaganda notwithstanding. And we certainly cannot change society or business.

 

So the answer is no. We will document any attempt to do it right--as we've been alluding to a new technology that allows true R implementations, and in On TRDBMS Implementation--but we won't go into products ourselves. That's about the only way we can continue to be right. All those few who also had it right and went into production very quickly stopped being right. As I said, it's systemic and inherent.

 

 

Posted 05/31/02

 

 

 

[ABOUT] [QUOTES] [LINKS]