[AccessD] Re: On DB Bloat, Bad DB Design, and various

Francisco H Tapia my.lists at verizon.net
Tue May 25 10:46:32 CDT 2004


Not ADD, but I do always ask that question, can there ever be more than 
one.  I generally have always stopped at 3rd Normal Form + Boyce Codd, 
where parts of the database do get fully normalized and others (for 
performance) do not.


Robert L. Stewart wrote On 5/25/2004 7:37 AM:

> Arthur,
>
> I agree in part with you.  One thing that no one has said that I think 
> is essential to "good" database design is imagination.  And, you are 
> correct in that there really is only one "perfect" design, with many 
> that are close, and even more that are flawed.  I think that the 
> imagination is where the "art" comes in.  There is nothing more 
> artistic than a nice E size plot of a good database design.  ;-)
>
> Most db designers will not go past 3rd normal form.  Sorry guys, but 
> that falls into the "close" category.  There is a reason that there is 
> 6 normal forms defined.  1 - 5 and Boyce-Codd.
>
> Sometimes the "art" is in how we ask the questions to determine the 
> correct design.  I have a friend that is extremely good at OLAP 
> design.  He was working with me on an OLTP system.  When I sketched 
> out the design, he looked at it and said, "Yeah, it looks right, but 
> how did you get there?"  When I explained all the steps that I went 
> through in my head, he was amazed at the jumps that I make in the 
> design mentally.  Art? Maybe, or maybe just a different way of thinking.
>
> The other thing that I told him is that I had reduced all of the 5 
> numbered normal forms to a single question.  "Can there ever be more 
> than one?"  If the answer is not "No," then you need another table.  
> Following this faithfully will get you fully normalized.  He thought 
> about this over the weekend, thinking that I had simplified it too 
> much.  He went and read the requirements for each normal form.  When 
> he came back on Monday, he said that he could not find an instance 
> where my way of thinking about design would not work.
>
> Maybe the art is in the questions we ask to get the design we get.  I 
> like to think that at least a little of what I do is art and not just 
> pure logic.  :-))  Or maybe it has to do with the fact that I am ADD.
>
> Might be some good questions...
>
> How many of you that consider yourselves good programmers are ADD?
>
> How many of you that consider yourselves good db architects are ADD?
>
> Robert
>
> P.S.  I think that this has been a great discussion.  It shows the 
> diversity of opinions from an educated group of developers.
>
> At 06:21 AM 5/25/2004 -0500, you wrote:
>
>> Date: Mon, 24 May 2004 22:46:44 -0400
>> From: "Arthur Fuller" <artful at rogers.com>
>> Subject: RE: [AccessD] On DB Bloat, Bad DB Design, and various
>> To: "'Access Developers discussion and problem solving'"
>>         <accessd at databaseadvisors.com>
>> Message-ID: <000401c44202$83eef560$6601a8c0 at rock>
>> Content-Type: text/plain;       charset="us-ascii"
>>
>> This is equivocation pure and simple, Drew... Though if you wish to
>> persist with it go ahead. There is IMO no comparison between designing a
>> bridge and writing a sonnet. Not least because in the former case if you
>> f**k up many people will die, while in the latter case no one will read
>> you.
>>
>> I grant you that GUI design is art. But database design is pure science.
>>
>> Arthur
>
>
>


-- 
-Francisco





More information about the AccessD mailing list