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: Entity vs. Table

Re: Entity vs. Table

From: Alan <alan_at_erols.com>
Date: Mon, 14 Jun 2004 09:56:27 -0400
Message-ID: <2j5p44Fu27biU1@uni-berlin.de>

"Laconic2" <laconic2_at_comcast.net> wrote in message news:c8-dnU5_7N-PMVDdRVn-uA_at_comcast.com...
> Ben and Alan,
>
> "Alan" <alan_at_erols.com> wrote in message
> news:2j5mp4Ft3gu0U1_at_uni-berlin.de...
> >
> > "ben brugman" <ben_at_niethier.nl> wrote in message
>
> > > The fysical implementation, normaly does have
> > > indexes, errorchecking bits and is often implemented
> > > on a Raid. Therefore the implementation contains
> > > redundancy. The fysical implementation therefore
> > > is not 3NF, although it often is an implementation
> > > of a logical model which should be 3NF.
> > >
> > > (By the way, indexes (which can be rebuild) are
> > > definitely redundant information).
> > >
> > > ben
> >
> > I've been nice up till now, but I'm sorry- you clearly do not know what
> you
> > are talking about. 3NF has nothing to do with RAID, error checking bits,
> or
> > anything else that is not data. Indexes are not redundant information.
> They
> > are references. Indexes are analagous to looking in the index of a book.
> You
> > look there to get a specific place to look in the book. If indexes were
> > redundant, they would serve no purpose. Indexes are part of the physical
> > implelmetation, and have zero to do with the logical model, which is
where
> > you establish 3NF.
>
> I honestly think the above disagreement between the two of you is caused
by
> language confusion rather than substantive difference. In particular, I
> think that the word "redundancy" means different things to the two of you.
>
> I think I understand what Ben is saying about redundancy and indexes,
error
> checking and RAID. I think he means, roughly,
> that an index can be derived from the corresponding table, the error
> checking bits can be derived from the corresponding block, and the RAID
set
> can be derived from one of the copies of it.
>
> Both of you are saying that a system that contains indexes, error
checking,
> or RAID does not necessarily violate 3NF, because 3NF pertains to the
tables
> themsleves, and not to the support structure.

That is ceratinly what I am saying.

>
> I think both of you know what you are talking about, but you aren't
> understanding the other one.

I hope you are correct!

>
> Or am I wrong?
>
>
Received on Mon Jun 14 2004 - 08:56:27 CDT

Original text of this message

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