Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> comp.databases.theory -> Re: Noninferential vs. inferential DBMS

Re: Noninferential vs. inferential DBMS

From: x <x-false_at_yahoo.com>
Date: Tue, 18 May 2004 09:28:23 +0300
Message-ID: <40a9ac37@post.usenet.com>

"Paul" <paul_at_test.com> wrote in message
news:R55qc.4379$wI4.499382_at_wards.force9.net...
> x wrote:
> > Or is projecting the tuple
> > ('Fred Smith', 123, 10, 10,000)
> > onto
> > ('Fred Smith', 123, 10)
> > an example of inference?
> >
> > The answer is:
> > This depends if the projection was done by the DBMS or by the
application
> > programmer/database user.
> > If the DBMS had a rule that said (a, b, c , d, e) -->(a, b, c) and this
rule
> > was used, then projecting the tuple is an example of inference.
>
> But by itself the tuple ('Fred Smith', 123, 10, 10000) is meaningless.

Yes. By itself.
But this does not mean the DBMS cannot do anything with it except store and retrieve.

> The DBMS just does does the mechanics, it can't know that the first
> column is a name, etc. In other words it can't know the real-world
> interpretation of the underlying predicate. So I'm still not sure what
> is meant by the DBMS inferring things.

Yes. The DBMS does the mechanics.
It don't have to KNOW the real-world interpretation to do that. It only need to have some rules that tell the DBMS what it can or cannot do.

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

Received on Tue May 18 2004 - 01:28:23 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US