Re: Proposal: 6NF

From: dawn <dawnwolthuis_at_gmail.com>
Date: 1 Oct 2006 07:27:27 -0700
Message-ID: <1159712847.070859.84120_at_k70g2000cwa.googlegroups.com>


David Cressey wrote:
> "Karen Hill" <karen_hill22_at_yahoo.com> wrote in message
> news:1159558507.190840.68280_at_m7g2000cwm.googlegroups.com...
> > 6NF would be a database that uses no Nulls.
> >
> > This way, one could specify that they demand their relations in 6th
> > Normalized Form, and have a database design that uses no Nulls.
> >
> > Those that like nulls can have thier 3NF, while the rest of us reach
> > for higher standers like 6NF.
> >
> > regards,
> >
> > Karen
> >
>
> First, let's find out where we all agree, then work our way towards where
> there are different viewpoints.
>
> Here's a starter: A schema in DKNF will have no nullable columns.

Unless this is only a discussion about SQL-based DBMS's, the term "null" might refer to a value (the empty set) if used in conjunction with two-valued logic (e.g. IBM UniData). Since most computer languages are two-valued, it might be best if you clarify that these "bad nulls" are nulls used with a 3VL language.

On a side note, I would suggest that the industry move away from use of a 3VL with DBMS tools as soon as feasible. I understand that from a practical perspective, the elimination of nulls is a tactic to help mitigate the problems with the language.

For future databases and systems, we should be looking at whatever will take the place of SQL so that we get past the 3VL NULL issue without bandaids. Even if we eliminate NULLS from base tables, are we going to eliminate NULLS in views too? Otherwise we are still stuck with 3VL until we move beyond it. Ultimately we need to address the cause, the 3VL language SQL.

--dawn Received on Sun Oct 01 2006 - 16:27:27 CEST

Original text of this message