Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!apple!motcsd!hpda!hpcupt1!hpindwa!donovan
From: donovan@hpindwa.HP.COM (Donovan Hsieh)
Newsgroups: comp.databases
Subject: Re: A few words on the "normaliztion"
Message-ID: <36270005@hpindwa.HP.COM>
Date: 2 Aug 89 22:51:06 GMT
References: <36270004@hpindwa.HP.COM>
Organization: Hewlett-Packard, Cupertino CA
Lines: 23

In reply from <marti@ethz.UUCP>, marti%inf.ethz.ch@relay.cs.net (Robert Marti) 
wrote :

> Huh?  I don't see what's so difficult about stating something like
> empno -->> language (employees, identified by empno, have the ability
> to speak/understand multiple languages).
My point is, it is difficult to formulate all possible MVDs for a given 
relational database scheme in the real world applications. Certainly, finding 
some or several of the MVDs are easier. If the designer choose to use the
synthese algorithm to normalize the database scheme, incomplete set of FDs &
MVDs can lead to incorrect result.

> No DBMS enforces 4NF (or "multi-valued update consistency", as you call
> it).  The database designer enforces 4NF, essentially by defining a
> table per MVD.  Note that this is no differnet from 3NF:  The designer
> enforces 3NF by defining appropriate tables and keys.
You are quite right. It is the DB designer's responsibility to enforce the
NF consistency. I have used a wrong word for saying "database system to 
enforce the consistency".

Donovan Hsieh
Hewlett-Packard
Business Network Division