Re: Interesting article: In the Beginning: An RDBMS history

From: x <x_at_not-exists.org>
Date: Mon, 10 Apr 2006 08:26:07 +0300
Message-ID: <e1cq8i$1qe$1@emma.aioe.org>

"Marshall Spight" <marshall.spight_at_gmail.com> wrote in message news:1144567187.961347.27720_at_e56g2000cwe.googlegroups.com...
> x wrote:
> >
> > I remember that the order of the rows is also irrelevant.
> > Why we name attributes with strings and we don't name the rows ? :-)
> > Why not the other way around ? :-)

> You smile, but it's actually an interesting question.

> The columns have an identity that must be unique; there is this rule
> that names identify columns.

Do you think the attributes as defined by Chris Date are "pointers" or not ? :-)

> The rows have an identity that must be unique; there is this rule
> that keys identify columns.

Why we insist the key be of the same degree for all rows of a relation ? :-)

> Coincidence? Are these two examples of the same rule, just with
> different key domains?

How about Codd's entity domains ? :-) Received on Mon Apr 10 2006 - 00:26:07 CDT