Re: 1GB Tables as Classes, or Tables as Types, and all that refuted

From: Ja Lar <jalar_at_nomail.com>
Date: Tue, 14 Dec 2004 10:04:00 +0100
Message-ID: <cpma4h$g21$1@news.net.uni-c.dk>

"Alfredo Novoa" <alfredo_novoa_at_hotmail.com> skrev i en meddelelse news:41bdcadc.25349859_at_news.wanadoo.es...
> On Sun, 12 Dec 2004 10:47:51 +0100, "Ja Lar" <ingen_at_mail.her> wrote:
>
> >Another example to support the notion of 2GB, but flawed as most of them
as
> >beeing based on using the wrong operators or methodology to handle
> >"identity" and "equality".
> >
> >To be equal is not the same as to be identical. Please read Mr. Gittens
> >paper (again) ...
>
> Gittens is very confused.

Please make a refusal by logical deduction, not by opinion.

> To be the same is not the same as to be identical :-)
Right.

> If two objects are identical this means that they are two different
> objects and not the same object :)

Nonsense.

> The identity of a value is simply itself.
Yes, in D&D's view, but that's a limited view supporting a specific aim. It doesn't distinguish equality and identity. The value is a property of the object, which have an identity. The 2GB "OID" is the "hack" that makes identity by equality of value possible. Thus, OID is _not_ a pointer, but a "feature" (a value) to give "identity" by value (as demanded in RM), _instead_ of by reference. It is a surrogate key.
>
> Two values can't be identical because if they are identical then they
> are the same value.

The values of two objects can be identical. Remember, OO has capabilities beyond RM - for the good and the worse. That's like in the real world. Received on Tue Dec 14 2004 - 03:04:00 CST