Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!batcomputer!munnari.oz.au!manuel!ccadfa!ghm
From: ghm@ccadfa.adfa.oz.au (Geoff Miller)
Newsgroups: comp.databases
Subject: Re: Relational Databases (and why they don't exist)
Message-ID: <2421@ccadfa.adfa.oz.au>
Date: 10 Jun 91 04:15:34 GMT
References: <6822@vela.acs.oakland.edu> <1991Jun07.063749.7904@chinet.chi.il.us> <1991Jun08.124337.29813@uvmark.uucp>
Organization: Computer Centre, University College, UNSW, ADFA, Canberra, Australia
Lines: 46

jim@uvmark.uucp (Jim Todhunter) writes:

>[... stuff deleted ...]

>Just a minor correction to this posting.  A relational database does NOT
>need to be normalized.  In fact, some of the most interesting current
>research is the area of NFNF (non first normal form) relational models.
>This is a very active area of investigation because there are serious
>deficiencies in the first normal form model.  While Codd dismisses NFNF
>and OO model out of hand, most others, including Date, do not share his
>view.  (It is interesting to note that CODD's recent book reads very much
>like a dinosaur's defense of his raison d'etre.)  NFNF models can also use
>indices.  The truth is normalization is merely a convenience for that
>database theorist, making the mathemetical proofs easier.

I have previously put forward the view that, if you consider a DBMS to
be "relational" only if it obeys all of the Commandments of Codd, there 
is no such animal as a truly relational DBMS.  I would also suggest that
this does not matter very much!

The relational database *model* is extremely useful for database design.
However, you can implement a relational model using a number of DBMS
products which are quite clearly not "relational" in the sense described
above, and as Jim points out, DBMS products which offer NFNF capabilities
have a number of uses.  Pick is of course one of the leading examples,
as Jim could have mentioned (but perhaps he didn't want to be accused
of advertising!).

Normalisation does have advantages for more that the database theorist,
though.  Remember that there is a basic law of computing that, if you
give a user the chance to enter the same piece of information in two
different places you will inevitably end up with two different versions.
Normalisation is largely about making sure that each piece of information
is stored once and once only.  However, you can implement a normalised
data *model* in ways which use NFNF structures.

>-- 
> James W. Todhunter, Manager, Software Development
> Vmark Software, Inc., 5 Strathmore Road, Natick, MA  01760, USA
> Internet: uvmark!jim@merk.com, UUCP: uunet!merk!uvmark!jim
> Phone: (508) 655-3700, Fax: (508) 655-8395, Telex: 5101011619 VMARKUNIVERS

Geoff Miller  (ghm@cc.adfa.oz.au)
Computer Centre, Australian Defence Force Academy

A