Re: Does Codd's view of a relational database differ from that ofDate&Darwin?[M.Gittens]
Date: Fri, 8 Jul 2005 17:34:51 -0400
I appreciate your spending time to educate me on intricacies of alternative data models.
"Jan Hidders" <jan.hidders_at_REMOVETHIS.pandora.be> wrote in message
> VC wrote:
>> Let's try another approach. *When* does one need "conceptual objects"
>> and "semantic domains" as opposed to just objects and domains ? On what
>> specific occasions ? I hope you agree, since you've excised this part of
>> my message, that U.o.E vs. U.o.D contradistinction does not make much
>> sense. If you do, the may question stands: what is a "conceptual object"
> When you are making a data model there are certain things you want to
> describe, and certain things you do not want to describe.
As I said before, the stuff you do not want to describe simply does not exist in your hypothetical model, we do not need to talk about it. Why do we need a term(s) for something we do no care about (within some context) ? When we start caring, we'll just begin using simple words like "domain" or "object" and be done with it.
> That is the distinction between U.o.D and U.o.E. What you seem to be
> asking is if there is somehow an objective quality that distinguishes the
> objects in the U.o.D from those that are not in it. But since this is a
> relative notion that depends upon the purpose of the model, there is no
> such objective quality.
If we cannot define the distinction even informally, how come one can use those words in a presumably scientific publication ? Be as it may, what's the *purpose* of this separation ? its practical value ? Could you give a rel life modelling example where the distinction would be useful ?
P.S. I presume a "semantic domain" is composed of "conceptual objects", right ? So, when I figure out, with your kind help, what a "conceptual object is", the "semantic domain" must be easy ;)
vc Received on Fri Jul 08 2005 - 23:34:51 CEST