Re: Indexes and Logical design

From: Gene Wirchenko <>
Date: Wed, 14 Sep 2005 10:08:41 -0700
Message-ID: <>

On Wed, 14 Sep 2005 10:05:48 GMT, "David Cressey" <> wrote:

>"Gene Wirchenko" <> wrote in message
>> If you want imprecise, talk to end users. They usually do not
>> know computers very well, if at all. There is no need to wonder why
>> it often takes a long time to get things straight with them.
>That is exactly what I want to be able to do. I want to be able to talk to
>end users, in language they can understand. If that means being imprecise,
>so be it.

     But if the end result needs to be precisely defined? At some point, the imprecision must be removed.

>I also want to be able to talk to programmers, and that's what got us
>started on this particular argument. I don't know how many programmers
>you've talked to. I've talked to hundreds. and the majority of them
>believe, correctly or incorrectly, that indexes are part of the logical
>design. Soo.... when I talk to them, I "muddle together" concepts that you
>consider to be disjoint or orthogonal, or whatever. Privately, I agree with
>you. But, for the sake of communicating with programmers, I choose not to
>fight this battle with them.

     So if you were to talk to me (programmer/analyst), you would a) sound as if you do not know what a logical model is, and b) possibly expect something from me that does not belong. (I would turn in a logical model WITHOUT indexes.)

     It hardly need be a battle. "That is part of the implementation. We are still on the logical model."

>Do I try to keep it straight in my own head? You bet! I understand
>perfectly well the differences between the consequences of bad table design
>and the consequences of bad index design. And, I did say, If you will look
>back to the topic starter, that I "very reluctantly" included indexes in
>the logical model, and that there was a great case to be made for excluding
>them from the logical model.

     Yes, but you still did it. If you are going to be informed by other people's ignorance, you are going to muddle. There goes precision.

     In medicine, "femur" is well-defined. It would not be acceptable to use the term as a synonym for "leg bone".

>That's the difference between me and most of the orthodox thinkers in this
>newsgroup. Most of you seem to think that a successful database application
>rests on logical thinking, and nothing more. Logical thinking is crucial,

     No. It is necessary but not sufficient. It is necessary.

>but it's just the beginning. The politics of information, industrial
>psychology, resistance to change, and a host of other human factors are
>critical success factors as well.

     Of course. All the politics will make no difference if the logic is not there. How many projects fail?

>And to me, the reason I include indexes in logical models is precisely due
>to these human factors, and not because logic demands it.

     If anything, logic demands that it not be there.


Gene Wirchenko Received on Wed Sep 14 2005 - 19:08:41 CEST

Original text of this message