Re: database design method

From: Leandro Guimarães Faria Corsetti Dutra <lgcdutra_at_terra.com.br>
Date: Thu, 31 Oct 2002 14:49:14 +0100
Message-ID: <aprccr$4bud9$1@ID-148886.news.dfncis.de>

Jan Hidders wrote:

> Leandro Guimarães Faria Corsetti Dutra wrote:
>

>> Jan Hidders wrote:
>> 
>>> Leandro Guimarães Faria Corsetti Dutra  wrote:
>>> 
>>>> Now I want to see it.  References, please.  You may want to be 
>>>> informed of articles against your view in
>>>> http://dbdebunk.com./, for instance.
>>> 
>>> Dbdebunk is not the gospel.
>> 
>> No, but it is a nice reference site and eye-opener.

>
> Sure, but that doesn't mean that everything they say is true. So
> pointing to an article of theirs and say "and therefore you are
> wrong" is not a good argument.

        I don't think I did. I only pointed to it for interesting, useful arguments in case you weren't aware of them.

        I said "you were wrong" yes, and then asked myself to be proven wrong.   But I never offered DBDebunk as any kind of proof, even because I beware of the argument from authority.

>>> The fact that you can use FOL to describe constraints in the ER
>>> model is so plainly obvious to people who know about logic and ER
>>> models that no serious researcher would write an article about
>>> that.
>> 
>> What about Fabian Pascal as "serious"?  What about 
>> http://dbdebunk.com./fp4a.htm as an article?  Or take Date and 
>> http://dbdebunk.com/kimball1.htm.

>
> Where exactly in those articles do they claim that you cannot use FOL
> to describe constraints in the ER model?

        No, they claim that ERDs can't represent all possible constraints. They are not arguing limitations on first-order logic, but on ERDs.

        Now, I never saw ERDs proposed as a complete, practical representation of neither logic nor a relational database. I would love to be proved wrong, because this would make the RM itself more palatable to GUI weenies that currently dominate all over the world.

>>> If you think there is an inherent problem there then I suggest
>>> you share it with us and I will be happy to explain why it isn't
>>> a problem. :-)
>> 
>> OK.  How do you express non-RI constraints?  How one does express
>> each domain, attribute, relation and database integrity
>> constraints?  Or in a different taxiology, take transition
>> constraints, how would ERDs

> represent

>> them?  I'm not saying you mightn't have a partial answer, but I
>> very much doubt you could give (or point to) a complete answer.

>
> All of the above can be done in first-order logic with the obvious
> limitation that you cannot express what you can express in
> higer-order logics but not in first-order logic. Again, what do you
> think is the problem? You do know what first-order logic is, do you?

        Again, how can all of first-order logic be represented in ERDs?

-- 
  _
/ \ Leandro Guimarães Faria Corsetti Dutra        +41 (21) 216 15 93
\ / http://homepage.mac.com./leandrod/        fax +41 (21) 216 19 04
  X  http://tutoriald.sourceforge.net./      Orange Communications CH
/ \ Campanha fita ASCII, contra correio HTML      +41 (21) 644 23 01

Received on Thu Oct 31 2002 - 07:49:14 CST