Re: database design method

Leandro Guimarães Faria Corsetti Dutra wrote:
>Jan Hidders wrote:
>
>> Leandro Guimarães Faria Corsetti Dutra wrote:
>>
>>> Jan Hidders wrote:
>>>
>>>> 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.
>>
>> Of course. (Do you think that the Relational Model can, by the way?)
>
> The RM can represent all machine-enforceable constraints AFAIU. If
>I'm in the wrong, please prove so.

Can you point me to the exact and complete definition of *the* language in which you define constraints in the RM?

>> But I said that you can use FOL to describe constraints in the ER
>> model. So if you extend the ER model with FOL (which is not very
>> difficult) then you can describe those constraints.
>
> What do you mean by "extend the ER model with FOL"?
>
> If it means complementing ERDs with textual representations of
>constraints, then we just have proved ERDs aren't enough.

I never said they were.

>Then it is just a matter of taste or circumstancial expediency to draw
>diagrams for quick visualisation and later flesh them out with test, or to
>go straight for text and later draw parts for presentation.

ERD diagrams can be given an exact semantics and are in that no better or worse than textual table definitions. However, they are often closer to how the user sees the data because, for example, you don't have to add attributes to model certain relationships.

• Jan Hidders