Re: I think my book may be wrong about cardinality, but I'm not sure
Date: Wed, 25 Jul 2007 12:50:07 -0000
On Jul 25, 1:33 pm, "David Cressey" <cresse..._at_verizon.net> wrote:
> "JOG" <j..._at_cs.nott.ac.uk> wrote in message
> > A relationship is a row in a table. Nothing more, nothing less.
> I disagree.
> Unfortunately, the word "relationship" is used in both E-R modeling and
> Relational Data modeling. These two usages have a lot of overlap, but they
> are not exactly the same concept.
> A relationship in E-R modeling is NOT a row in a table. It's an association
> asserted to exist between 2 or more entities, in the real world that the
> database describes.
> The association would exist, even if there were no database in which to
> record it.
> In relational modeling, your comment is correct.
> People who see the world exclusively through the lens of the relational data
> model see all real world relationships as instances of relational data
> relationships. But that's subjective.
Hi David, I have to take umbridge with you there - I certainly don't see the world through a relational model lens.
When I was once struggling with the problems of a network model which I was researching I had to take a step back. I tried to analyse first what information and data /was/ and this led me on to having to look at exactly what entities, relationships, etc, were. And I ended up not being able to find any distinction whatsoever between what E/R denotes as "associative entities" and "fundamental entities". My conclusion was that they are both internally n-ary relationships (or both associative entities depending on how you want to frame it)
The fact that RM also makes no such distinction, and the lack of query bias that results, was interesting to me - and hence I made a (painful) transition to favouring that model. But this all came about after my own investigations, especially helped by William Kent's writings, and not because of any prior interest in RM whatsoever.
I also understand that this view is not shared by a large proportion of the business world, who still like to think in terms of objects, which I appreciate as it once seemed intuitive to myself too. I also take your point that ER modelling can be neutral to different approaches.
Regards. Received on Wed Jul 25 2007 - 14:50:07 CEST