Re: what data models cant do

From: Kenneth Downs <knode.wants.this_at_see.sigblock>
Date: Tue, 17 May 2005 14:36:12 -0400
Message-Id: <a9qpl2-ivm.ln1@pluto.downsfam.net>

Paul wrote:

> Kenneth Downs wrote:

>>>Non IT people understand relations very well too. IMO business people
>>>tend to understand relations and set theory better than IT people.
>> 
>> Yes, indeed.  I formulated my "law" when I observed that people of any
>> technical background or management level could actually conduct very
>> productive meetings together if they stuck to the detailing of what
>> should
>> be kept in tables.  There is something in the human mind that easily
>> works with tabular data.
>> 
>> I believe, completely without proof, that this human intuition is
>> ultimately
>> behind both the success of spreadsheets and the relational model.  It is
>> what gives us our sense that the relational model is "elegant".

>
> I disagree; I think the relational model is actually counter-intuitive.
>
> If you've ever seen (or been) a database novice, the instinct is to
> stick everything in one big table, because they then feel it is easy to
> get at what they want without having all those complicated joins getting
> in the way.

As the novices mature, they run into troubles and seek a way out, which leads them by one route or another to discover normalization.

Thus questions the student: Master, I thought tables were good, and yet my table is clumsy and difficult.

Thus replies the master: Ah, my grasshopper, if one table is good, more tables are better. Seek thou to have a place for everything and everything in its place.

...and the student was enlightened.

>
> It just goes to show that intuition isn't always right (cf. quantum
> mechanics, relativity, etc.).

Some intuition is learned. I find both quantum and relativity to be very intuitive.

>
> I think the sense of elegance stems from the fact that every fact is
> stored in one place only.
>

See above.

-- 
Kenneth Downs
Secure Data Software, Inc.
(Ken)nneth@(Sec)ure(Dat)a(.com)

Received on Tue May 17 2005 - 13:36:12 CDT