Re: By The Dawn's Normal Light

From: Dawn M. Wolthuis <dwolt_at_tincat-group.comREMOVE>
Date: Sat, 23 Oct 2004 22:39:15 -0500
Message-ID: <clf85i$8s6$1_at_news.netins.net>


"Alfredo Novoa" <anovoa_at_ncs.es> wrote in message news:2o4mn0l9nu46phruo9kjon8ngrjj08hc6t_at_4ax.com...
> On Fri, 22 Oct 2004 17:29:02 -0500, "Dawn M. Wolthuis"
> <dwolt_at_tincat-group.comREMOVE> wrote:
>
> >I found a paper that I bought from Date last year specfically on 1st
normal
> >form. He starts with a quote from Codd
> >
> >"A relation is in first normal form if ...none of its domains has
elements
> >that are themselves sets. An unnormalized relation is one that is not in
> >first normal form."
>
> But why is he quoting Codd?
>
> To correct him?
>
> And what is the conclusion of Date's article?

Yes, I think I mentioned in a subsequent post that by the end of the article, he claims that by his def of relation, all relations are necessarily in 1NF

> >So, I think we can get passed the flawed statement that a relation is
> >necessarily in first normal form.
>
> This is astonishing even coming from you.

I know you aren't a fan, Alfredo. It is usually easy to tell whether what I say is logical or a jump in logic based on intuition or experience. I hope you don't discount the former due to the latter. Your writings have a similar quality ;-)

> See this:
>
> 1NF: As we explain in PRACTICAL DATABASE FOUNDATIONS papers #1, What
> Normal Form Really Means, and #2, What First Normal Form Means Not,
> atomicity is not precisely definable and, therefore, neither would be
> a 1NF definition based on it. A table faithfully representing a
> relation variable (relvar) is by definition in 1NF and, therefore, has
> exactly one attribute value of the pertinent data type in each cell;
> any types of value are permissible, as long as they are associated
> with applicable operators.
>
> http://www.inconcept.com/JCM/September2004/Pascal.html
>
> I always thought that you were a troll, but this is too much. You are
> misquoting the article to show that it supports contrary of the
> article's conclusion.

I found my copy of the pdf file and started reading at page one until I found a definition, so I sent it. I didn't finish the article until later, at which point I corrected my error. I'm not a troll. I even go by my real name here. If I were a bit younger, I would have recalled the article in full from having read it last year, but, alas.

> > A normalized relation is in 1NF, but a
> >relation, by its definition need not be. I'll read further to get what I
> >think is Date's most recent def of 1NF.
>
> So you didn't read the article ...

I read it last year and then I re-read it, but not before sending this quote from the start of it. I also have answers to all of the questions that Date poses at the end related to multivalue databases. --dawn Received on Sun Oct 24 2004 - 05:39:15 CEST

Original text of this message