Re: One-To-One Relationships

On Nov 30, 5:29 am, David BL <davi..._at_iinet.net.au> wrote:
> On Nov 30, 5:09 pm, Marshall <marshall.spi..._at_gmail.com> wrote:
> > On Nov 29, 10:50 pm, David BL <davi..._at_iinet.net.au> wrote:
> > > On Nov 30, 2:54 pm, Marshall <marshall.spi..._at_gmail.com> wrote:
> > > > On Nov 29, 9:20 pm, David BL <davi..._at_iinet.net.au> wrote:
>
> > > > > How does one interpret a tuple as a proposition about the real world
> > > > > if one avoids any conception of entities?
>
> > > > How does one interpret a proposition as a proposition about
> > > > the real world?
>
> > > I'm not sure what you mean.
>
> > > I didn't state myself clearly. I wasn't intending the formal meaning
> > > of "proposition". Please substitute a less formal word like "fact" in
> > > my previous post.
>
> > Okay. Let me rephrase my question according to your specifications:
>
> > How does one interpret a proposition as a fact about the real world?
>
> > I believe that the answer to this question is the same as the answer
> > to the question you asked, which was:
>
> > "How does one interpret a tuple
> > as a proposition about the real world
> > if one avoids any conception of entities?"
>
> Agreed, but what do you think the answer is?

I think the answer is that we use words, and we know what the words mean, and the meaning is the fact about the real world that we wanted.

I'm sorry that my answer is so primitive, but you asked a really primitive question.

> > The formalism that gives us propositions, and that lets us use them
> > to state facts about the real world, and infer new facts from the
> > stated facts, does not have any conception of entities.
>
> Agreed
>
> > Is it clear what I'm saying?
>
> No. It would help if you answer the question!

What I was saying is that entities, as a logical concept, don't add anything that we don't already have with propositions.

Suppose I say

M(s)

What does it mean? Nothing, without an interpretation. My interpretation

s = Socrates
M(x) = x is a man.

The interpretation is in your head. The interpretation is a mapping from the formally stated concepts to words, ideas, thoughts, whatever is in your head.

> Let me ask it a different way: Without any notion of entities in the
> real world, isn't a formal proposition just a meaningless formula?

I don't think so. Can't you have formal propositions about abstractions?

  There is no largest integer
  !ExAy(x>y)

> I understand how it is important that the formalism itself does not
> have any conception of entities. It seems to me there is an important
> line in the sand, with
>
> formalism,
> propositions/tuples,
> predicates/relations,
> extensions
>
> on one side, and
>
> informalism,
> the real world,
> entities,
> intensions,
> natural language,
> facts
>
> on the other.

That seems like a very heterogeneous list to me.

> > > > > It seems to me that some entities are inevitable. Can't the distrust
> > > > > of entities be stated without throwing the baby out with the bath
> > > > > water?
>
> > > > Counter-question: what do you get from entities that you
> > > > don't get from propositions?
>
> > > I would have thought we need both.
>
> > That's no excuse for not answering my question. :-)
>
> > > By "proposition" do you mean formula from the propositional calculus?
>
> > > The propositional calculus is a formalism and doesn't come with some
> > > mapping back to the real world. I don't understand how any mapping
> > > could be understood without any conception of entities.
>
> > Does that mean you understand the term "entity" to be referring
> > to the conceptual layer? I don't think that's the usual sense of
> > the word.
>
> I have read a number of posts from cdt talking about conceptual vs
> logical layers but I'm afraid I still don't understand the
> distinction.
>
> My understanding of conceptual layer comes from Date's 3 level
> description in chapter 2 of An Introduction to Database Systems, and
> I'm not sure how that is relevant to this discussion. I understand it
> as the intermediary level providing both physical and logical
> independence. Can you elaborate?

I'm probably not the best person for that. But, to my way of thinking, the conceptual layer is what's in your head.

My Occam-inspired view: you see stuff and you hear stuff, and you get ideas in your head. You'd like to operate on these ideas in a formal way. You decide to use relations to do so. So you need to map what you think into relations. You could go though any number of intermediate steps to get there. But don't! Just write down what goes in the darn tables.

Some time back someone proposed "people understand tables just fine" as a law.

I have also on occasion referred to "Table/Attribute Modeling."

You just write down the darn tables.

It's easy enough that pretty much anyone can learn how to do it fairly well, directly. So I see no point in any further methodology.

> > > Isn't this mapping related to intensional definitions? These are
> > > typically stated in natural language because they can't be
> > > formalised. It seems to me that the instantiations of natural
> > > language intensional definitions are the counterpart to propositions
> > > being instantiations of predicates, and such natural language
> > > instantiations always refer to things that are assumed to exist in the
> > > real world (assuming the RDB is meant to model reality somehow).
>
> > Sure.
>
> > So, given all this, do we need to separate out the concepts of
> > entity and relationship at the logical layer? It seems to me the
> > answer is a clear "no."
> > Since, as you say, the mapping to the real world cannot
> > be formalized, I figure anyone can go ahead and think
> > about it however they like.
>
> Yes, however I would have thought the intensional definitions, which
> may be expressed in natural language, or inferred from the relation
> and attribute names, are critically important to anyone who wants to
> use an RDB successfully.

Oh my yes.

Marshall Received on Fri Nov 30 2007 - 20:44:17 CET