Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Usenet -> comp.databases.theory -> Re: Conceptual, Logical, and Physical views of data

Re: Conceptual, Logical, and Physical views of data

From: dawn <>
Date: 12 Sep 2005 17:40:15 -0700
Message-ID: <>

David Cressey wrote:
> "dawn" <> wrote in message
> > I've learned plenty from you, David. I suspect you might not suggest
> > you have learned anything (positive) from me.
> I'm frustrated that I have heard from you over and over again about the
> glories of Pick and a variety of other
> products, without a simple illustrative example that will explain why
> someone who can use SQL and the RDM to good effect should be interested in
> the alternative you claim to be superior.

I haven't been trying to convince you to be interested in other approaches. I didn't know you wanted an example. I argue why I still believe something so I can hear all counter-arguments, again hoping a lightbulb goes off either so that I switch to a more popular position or better understand why my own is viable.

I do have it on my to do list to load an mv product and prepare a schema that Lauri & Marshall asked for. I think I have said that the data model is similar to XML implemented as hash files and I do give examples of various things in the "flash cards" that I've pointed you to on my web site. What is it that you want?

> > >From your statement here, I do sense that you think I say "I don't
> > understand" when I really do understand, but disagree.
> Yes.
> > I find it exceedingly irritating that I don't buy into relational
> > theory the way most professionals do.
> > All external factors point to
> > SQL-DBMS tools as being a good way to do business, including IBM, MS,
> > and Oracle and their db customers all investing heavily in these.
> I'm glad to see you finally admit this.

I am unaware that I ever wasn't willing to admit this. By external factors I am referring to factors not within the tools, but about the tools, such as their market penetration.

> You've gone on and on about how
> much IBM invested in one of the MV products you like, but have never, to my
> knowledge, admitted that IBM makes that investment in addition to (not
> instead of) investing in the products of the relational class.

I wasn't aware there was a need for me to admit that IBM spends far more dollars on DB2 than any of their other dbms tools. I would think that was exceedingly obvious. Is there anything else that I would happily admit to that I didn't know was required or even useful?

> > So, why don't I "get it"?
> Everything you have said convinces me that your failure to "get it" is NOT
> due to a lack of intelligence, education, experience, or expertise. What
> you are expressing is "invincible ignorance".

I don't understand how you could consider me intelligent and yet think I am permanently stuck with my ignorance. Again, I don't know what you would have me do, say, learn, ...

> > It isn't that simple. I still think I'm not "getting it" and you are.
> > I would like to wake up at some point and have it all as clear to me
> > why all my standard database application data should be modeled
> > according to relational theory or, if that doesn't happen, have enough
> > understanding of why I disagree that I can give a rationale that is
> > more than just "but my experience has been ..." or "but I think ...".
> >
> I'm going to repeat what I've said before: I'm no theoretician. If you
> need a theoretical proof that the relational model is sound, I simply
> cannot provide that.

I would like that, but I can see now that is not possible. I'm trying to understand where it is tight and where it is not so I can figure out why my experience does not match the theory.

> If you need examples of success stories using RDM from
> experience, they abound in the industry. My own experiences are meaningful
> to me, but you have far better examples than mine to observe.

I certainly have seen RDM success stories. I've seen lots of COBOL success stories too, but it is not where I would suggest our industry head for the future.

> > And I would be very pleased to hear that explanation (although you
> > might tell me that you have already provided such -- I really am having
> > memory issues).
> >
> > > But I have little, if anything, to teach to you.
> >
> > You might feel that way, but I have learned quite a bit from you. If
> > you say something and I disagree, should I pretend to agree? If I
> > don't understand how you arrived at a conclusion and want to, should I
> > refrain from asking? I would think your answers to those questions
> > would be "no", right?
> >
> > > I need not teach you
> > > anything that you already know to be true. I cannot teach you anything
> that
> > > you already know to be false. What's left?
> >
> > I'm hoping you are going to be willing to tell me what you have taught
> > others is the difference between a pointer and a foreign key. But then
> > I suspect I will have a follow-up question and you will think something
> > like "You can't teach that girl anything!" and we will do the same
> > dance that you don't enjoy.
> The stuff I said was very low level introductory stuff. The sort of thing
> that would have been learned by people who worked for you on the Oracle
> projects that you led. I think it's beyond belief that you have not already
> been exposed to everything I could say on the subject.


> All I'm going to suggest is that you go back to some classic work on
> databases, preferably one that predates the widespread acceptance of the
> relational model, and read up on the difference between "pinned records"
> and "unpinned records". I am sure you can, if you will, learn all that you
> need to learn from that.

I have never heard the terms "pinned records" and "unpinned records" so you might want to retract that "beyond belief" statement ;-) Thanks, I'll look those terms up. Cheers! --dawn Received on Mon Sep 12 2005 - 19:40:15 CDT

Original text of this message