Re: Object-oriented thinking in SQL context?

From: Bernard Peek <>
Date: Tue, 9 Jun 2009 18:35:34 +0100
Message-ID: <>

In message
<>, writes

>> >> At the moment relational theory seems to be so effective at handling
>> >> low-level database management that I think that its practitioners are
>> >> quite right in considering themselves an essential component of
>> >> efficient systems design.
>> >What is low level database management?  What is high level database
>> >management?
>> In this context high-level database systems are object-oriented. The
>> analogy is with high and low-level programming languages.
>I am having difficulties understanding you. Could you state
>according to what principles you can qualify a system as being high
>level or low level. In database theory, low level = physical level =
>procedural languages. What is high level according to OO mindset?

Higher level languages deal with more abstract concepts that are closer to the way non-specialists view them. So I see the relationship between OO and RM as essentially similar to that between COBOL and Assembler.

Whether or not you agree with this characterisation it's the prevailing view in software engineering now.

I'll extend the analogy a bit further. In the past assembler languages were used to coax the last bit of performance from limited hardware. Now it's usually easier to upgrade the hardware than to rewrite code in assembler. But the people who write compilers still occasionally need to rewrite critical sections of code in assembler.

By analogy, most developers use OO analysis and coding techniques. They aren't concerned with RM at all. But there are still some people who need to get maximum performance from database systems. Those are the DBAs and the compiler writers who write the driver software that interfaces between OO languages and relational databases.

>> >> On the other hand from an OO practitioner's point of view relational
>> >> theory is a quite little backwater that doesn't have much applicability
>> >> in the real world.
>> >There seem to be  a contradiction between this statement and the
>> >previous.  How can you claim that relational theory is effective into
>> >handling some database management and then denounce its lack of
>> >applicability.
>> Please re-read what I actually said.
>I believe I did. I just asked a question. How can something that is
>applied on a daily basis have a lack of applicability ?

That wasn't what I said. I said that *from the OO practitioner's point of view* RM has little applicability. So I don't see any conflict between the two viewpoints.

Bernard Peek
Received on Tue Jun 09 2009 - 19:35:34 CEST

Original text of this message