Re: Is this bad design ?

Dawn M. Wolthuis wrote:

>>> Hand data with such relationships to any secretary in the 1950's
>>> and they will know how to "model" it and store the data for
>>> easy retrieval in their paper system.
>>> Thank goodness they were not constrained by some desire
>>> to "normalize" the data, eh?

That databank would not qualify as large and shared. There would be no basis for such a desire. For small, shared databanks I prefer paper and pencil above computerized administrations. Of course this is only feasible if all the people sharing this administration have physical access to it, and all people that do can be trusted to care for the consistency and actuality of it.

>>Dawn, what would you do if we extended the example and said that there
>>is also another table called Father, and that every Child must also
>>have a Father? The Child can't be modelled as an attribute of both
>>tables, can it? And then another table called School, and another
>>called Gang, ..., all with similar 1:M relationships. What then?

> You are absolutely correct that if the requirements were different, then I
> would model it differently.  I would also write all other aspects of the
> applications involved differently.  The requirements rule! smiles.  --dawn

They also change. Received on Wed Mar 10 2004 - 15:28:53 CET