Re: Relational Databases and Their Guts

From: Paul Vernon <paul.vernon_at_ukk.ibmm.comm>
Date: Wed, 25 Jun 2003 18:31:43 +0100
Message-ID: <bdcmeg$cq8$1_at_gazette.almaden.ibm.com>


"Bob Badour" <bbadour_at_golden.net> wrote in message news:yrjKa.505$My6.68419676_at_mantis.golden.net... [snip]
> > Let me reword it as:
> >
> > The ability to interact with a [subset of a] database via any one of
> > multiple information equivalent[subset] database schemas defined to the
> RDBMS
> >
> >
> > "base relations" are in the eye of the beholder.
> >
> > Logically, all equivalent schemas are equal, there is no requirement to
> make
> > one more equal than the others by making it 'primary'. Other than each
> user
> > choosing one as their current schema that is.
> >
> > Physically, one schema will be picked as the basis for physical
> > implementation. This schema can be called "base" if you wish, but users
> should
> > not care.
>
> I still do not see how one will specify the information equivalent schemas
> without views. With updatable views, users should not care either.
>
> What is different from views in what you are suggesting?
>

Well obviously you use something (very) like views, it's just that you need a bit of extra other stuff to enforce the information equivalence and enable the switching.

Regards
Paul Vernon
Business Intelligence, IBM Global Services Received on Wed Jun 25 2003 - 19:31:43 CEST

Original text of this message