Re: Operationalize orthogonality

From: x <>
Date: Mon, 5 Jun 2006 15:18:13 +0300
Message-ID: <e6177v$3b3$>

"Tony D" <> wrote in message
> x wrote:
> > "Tony D" <> wrote in message
> >
> > >
> > > x wrote:
> > >
> > > > How complicated! It would not be easyer to follow Mr. Codd advice ?
> > > > Domains, not types.
> >
> > > And the difference between a domain and a type is what ? Precisely ?
> > The "standard" answer would be "educate yourself". :-)

> Alternatively, you could attempt a non-standard answer ?

I did. Have you noticed the "" and the :-) ?

> > I don't know how "precisely" you want me to be.

> As precise as you need to be to point out what you believe the
> difference between a domain and a type is.

It was precise enough ? I wondered if you asked for mathematical symbols, that's all.

> > I would say that my current understanding of this (because I have not
> > read Codd's 1990 book - which I strongly recommend because I browsed it)
> > that domains do not include any kind of operator.

> So, if domains include no operators, what can you do with them ?

Anything you want. For example you can define relations on them. Of course I excluded from the exclusion equality, membership and enumeration or something like these.

> > What is your opinion ?

> In this context, I equate the terms domain and type - as mentioned in
> the presentation pointed to elsewhere on this thread.

Sorry, I've forgoten about that. Good work. What I do not include in the domain (or type if you want to call it that way) are the operators like +, -, *,/, type conversions, etc. I think you first need a domain to be able to define a type - or viceversa :-)

<(I don't,
> however, equate classes with domains or types - principally because
>(a) I'm not 100% clear on what a class is exactly,

I think you have been a member/participant of/in a class, so the percent is greater than 0.
Can you give an approximation of the percent ? :-)

> (b) from what I do
> understand about classes & objects, there is a dynamic element to them
> that I wouldn't expect to find in a domain or type. I am open to
> persuasion on these points.)

How about object = set of relations + operators (or object = db) ?

Sorry about these ":-)" Received on Mon Jun 05 2006 - 14:18:13 CEST

Original text of this message