Re: The original version
Date: Thu, 25 Jun 2015 00:55:56 -0700 (PDT)
Message-ID: <80a30061-f3bb-459b-89d1-92ae77fb4630_at_googlegroups.com>
Dana petak, 6. kolovoza 2010. u 12:06:16 UTC+2, korisnik Bob Badour napisao je:
> vldm10 wrote:
>
> > On Aug 5, 2:41 am, Bob Badour <bbad..._at_pei.sympatico.ca> wrote:
> >
> >>vldm10 wrote:
> >>
> >>>The authors of Anchor Modeling state that their model is based on the
> >>>Sixth Normal Form. However, 6NF cannot secure binary structures. Let
> >>>us consider a relation with five attributes that are mutually
> >>>independent. This relation is in 6NF , thus it cannot be further
> >>>decomposed. If we wish to get the history for each of its attributes,
> >>>things become very complicated.
> >>
> >>Huh? Do you know what 6NF is?
> >
> > I guess I know?
> > Do you know what 6NF is?
> >
> > 6NF is defined by the authors of the “Anchor Modeling” as follows:
> > A table is in 6NF iff it satisfies no nontrivial join dependencies at
> > all.
> > ( http://www.anchormodeling.com/tiedostot/6nf.pdf )
> >
> > Maybe you can explain your thinking by using this definition?
>
> How does a relation with 5 mutually independent attributes meet the
> definition of 6NF?
I think now I can fully answer this question. Question 6NF becomes a serious issue and it spins at user groups in the past ten years I would like to say that 6NF can not solve anything. I think so because the authors of 6NF do not provide any solution or procedure by which we can put a relvar in 6NF. But let's start from the beginning. Your question, "Huh? Do you know what is 6NF? "Is related to your next question question: „How does a relation with 5 mutually independent attributes meet the definition of 6NF?“
Your note is OK. Your intention was as folows:
(i) JD * {A,B,...,Z} is trivial iff at least one of A,B,...,Z is equal to
entire heading of pertinent relvar R.
(ii) Relvar R is in 6NF iff it satisfies no nontrivial JDs at all.
(iii) So your question „How does a relation with 5 mutually independent
attributes meet the definition of 6NF?“ is OK.
Let me explain what I want to say „with 5 mutually independent attributes“.
My Simple Form is not based on FD. It is based on quite different grounds,
that enables the decomposition of data structures in atomic (binary) data
structure.
For example, I can work with binary files. Of course, the intention of the
Simple form is to work with binary entities and binary relations.
In contrast to 6NF, my Simple Form gives the conditions under which we can
get binary (atomic) data structures and "simple key" in these structures.
This refers to the databases that maintain the current state. The main
requirement is that these data structures have attributes which are mutually
independent.
In my General theory of databases I use other tools. Thus decomposition into
atomic data structures in my approach is divided into two different
theories. Several times I presented it on this user group. In short, this
other theory I named General db theory which has its General Form.
Here are a few additional remarks. I do not use non-defined terms such as "metadata". But the question is how here goes predicate calculus for relvar R with "metadata". Note metadata are not attributes of objects. So, how to explain the semantics, that is the relationship between the (formal) language and objects from the real world that the language denotes. In other words, how to explain the connection between our (logical) model and the corresponding fragment of the real world, which is modeled. Here I want to emphasize the distinction between semantics and logic.
Now, let's take the example from Anchor modeling:
(AnchorKey, attribute1, metadata1, ..., metadataN), key is K = (AnchorrKey,
metadata1).
Note that here some "metadata" can be a kind of "interval-type". Here, it is
necessary to do the following analysis:
When the above "relvar" is in 5NF? When is in 6NF?
Note that in main paper of Anchor Modeling, mentioned relvar is presented only in the very reduced form: (AnchorKey, Attribute1, T) and key K = (AnchorKey, T). Of course, this kind of presentation does not look like a serious analysis.
As for plagiarism, which is done in Anchor Modeling, which is related to 6NF, I can say the following:
1. Anchor Key je plagijat ključa iz moje Simple Form. Note that the key from Simple Form is simple. Note that Anchor key is slightly different from Codd's surrogates.
2. Authors of Anchor modeling inserted "Historized" data structures in 6NF. But these "Historized" structures I presented on this the user group in 2005, and the Anchor modeling was presented in 2009. It is important to understand that here 6NF can help. "Historized" data structures that the authors of Anchor modeling put in 6NF, solve this problem. 3. Anchor Modeling works with the entities, but suddenly the authors of Anchor modeling work in relational model and here they prove that their "Historized" structures are in 6NF.
Since the mapping between data models is made only in my data model, I explained on this thread that the authors of Anchor modeling did not do a mapping between these two aforementioned data models. In addition to other posts, I wrote about this serious error, in my posts to you, in this thread. After my comments, authors of Anchor modeling published entirely new paper in 2010, where they plagiarized my theory of states, which is the basis of my theory and my data model.
With this post, I'll end up writing about my papers. Of course, if someone wants to write something or ask a question, let put it on this user group, I will respond to him.
(more details that are relating to this post you can find in my thread: The anatomy of plagiarism that was made by the authors of "Anchor Modeling")
Vladimir Odrljin Received on Thu Jun 25 2015 - 09:55:56 CEST