Re: some information about anchor modeling
Date: Wed, 27 Nov 2013 10:12:52 -0800 (PST)
Message-ID: <0e8a6f11-7140-49f8-86ba-6209ae366cb3_at_googlegroups.com>
Hi Derek
With this post I would like to conclude this discussion about the conceptual
model. In fact, very little has been written about the conceptual modelling. In
my opinion, the E / RM is not conceptual model at all.
In this post I will show that E / RM is bad based on inaccurate or naive
statements. The two most important things in the E / RM were not done. It is not
done definition of the concept and Conceptual model is not built as a fully
(well established) theory
In my opinion Anchor Modeling is used to correct large gaps of several major db
theories.
In the subsections 1, 2, 3, 4, 5, I'm going to show some serious errors of E / RM:
1.
In the paper, The Entity-Relationship Model - Toward a Unified View of Data,
Chen says, in section 2.2, at the conceptual level, the following: "There is a
predicate associated with each entity set to test whether an entity belongs to
it.
It appears that P. Chen is not clear enough about predicates and what is not a predicate. Note that predicates do not belong to conceptual level. Note that at the conceptual level, predicates do not test anything. At the conceptual level there is one other testing. For example, there is testing whether a particular attribute "falls under the corresponding concept." The predicates are on the sentence level; more precisely they are a kind of grammatical constructs. The concepts are on the level of thoughts.
2.
There are no "conceptual objects," as Chen it claims. In fact, there is a big
difference between concepts and objects (entities). Therefore, the following can
not be: "The entities, relationships, and values at level 1 are conceptual
objects in our mind," as Chen wrote in section 2.3.
It seems that Chen believes that the entity (object) is some kind of a concept? Let me quote Frege: "A concept - as I understand the word - is predicative. On the other hand, the name of object, a proper name, is quite incapable of being used as grammatical predicate. "
3.
Chen did not write the definition of the concept, which is a fundamental thing
in the conceptual modeling.
Chen does not explain what it is that connects the predicate with the concept. He do not understand the nature of the wholeness that consists of the concept and the predicate. So the fundamentals are not defined in the E / RM. Note that here we are speaking about the most important things related to mental activities. Therefore definition of the basic terms is important. This is about concepts. Concepts are also important for set theory. This matter is also related with the following crucial issue: How a man realizes the meaning of a word or expression?
4.
I looked again at Chen's paper about the E /RM . Chen introduced entity /
relationship relations (see section 2.3.2). It should clearly say, that this was
already done in RM, by Codd . More specifically these entity / relationships
relations are a special case in RM. It is true that the entities and
relationships improve semantics. But, let me say it is about 10% of this job.
Note that Codd was the first who applied Frege's theory to database theory, and
thus first achieved semantics. He had developed completely RM. In my opinion,
the following conclusion is correct:
(i) Chen applied the existing results from theory to creation of E/RM. He
accepted Godel's definition that the world is discrete, that is, that the
world consists of individuals, relationships among individuals, and
attributes. (Note that some other scientists are coming up with the same
or similar conclusions).
(ii) Further Chen uses the RM , which Codd was already made.
(iii Chen did not do conceptual part. That part which is fundamental to
science. Chen did it at the level of a few naive and very vague and
confusing observations.
For example, in section 2.1 Chen wrote the following:
"(1) Information concerning entities and relationships which exists in
our minds."
Chen also claims the following: "The entities, relationships, and values
at level 1 are conceptual objects in our minds. "
In section 2.3 Chen wrote the following: “Basically, an entity key is a
group of attributes such that mapping from the entity set to the
corresponding group of value sets is one-to-one. If we cannot find such
one-to-one mapping on available data, or if simplicity in identifying
entities is desired, we may define an artificial attribute and a value set
so that mapping is possible.”
I must say that the entity with the artificial attribute does not exist in
the real world.
Therefore a concept for the real entity with the artificial attribute does
not exist.
=============================================
In the text that follows I will show that Anchor Modelling trying to fix the big mistakes of other db theories, using my results.
5.
As I showed in my last two posts, conceptual modelling is not well defined in
many parts of this model.
Let me mention just some of them;
(i) There is no good definition of the concept.
(ii) There is no solution that will associate to the concept the appropriate
mathematical structures.
(iii) There is no theory that defines the mapping from the conceptual model to
other data models.
(iv) There is no solution for the decomposition into the atomic structures
(v) There is no general theory that connects all parts.
Anchor Modeling is published two papers. In their first paper (which won first
prize at the ER2009 - Conceptual Modeling Conference) their two main structures
are defined on the same data model that uses E / RM. The data model consists of
tables. For example, the main structures are defined as follows:
Def 2. An anchor A(C) is a table with one column. The domain of C is ID. The
primary key for A is C.
Def 5. A historized attribute Hatt(C, D, T) for an anchor A(C) is a table with
three columns. The domain of C is ID, of a non-null data data type, and of T a
non-null time type. Hatt.C is non-null foreign key with respect to A.C (Hatt.C,
Hatt.T) is a primary key for Satt
In the improved version of Anchor Modeling, (which was published in DKE, which chief editor is P. Chen) the new definition of Anchor is introduced.
Def 4. An anchor A is a string. An extension of an anchor is a subset Of I. Def 7. A historized attribute BH is a string. A historized attribute BH has an anchor A for domain, a data type D for range, and a time type T as time range. An extension of a historized attribute BH is a relation over I x D x T.
What is wrong here, it is that this data model with "extensions" and "sets" is not defined. It is not known whether the authors used an existing theory or is it their invention. The authors did not provide any reference. There are no definitions of basic terms. Note that this matter is about fundamentals. In my opinion this is not science and this is not the way how scientific paper should be presented.
Note that my model has extensions and concepts. Anchor model has two data
models??? First one uses tables, defined in the first paper, while the new model
uses "extensions" and "sets" defined in the second paper. Obviously, the latter
is published in order to fix the existing conceptual model.
Note also that this version of Anchor Modeling has improvements based on my
critique from my thread "The original version". I think that conceptual model
(which is on the level of thoughts) can not be built on the tables.
6.
In OO data model there is one big unsolved problem. OO uses surrogates; all
entities stored in OO model are on the db level. Note that in RM, entities are
at a relvar level. As entities can change their attributes during the time, it
is possible, for example, the following situation:
Five entities can get all the same values for all their attributes. For example,
entitet1 has all the values of all its attributes, the same as the values of the
corresponding attributes from entity2. The same holds for attributes from enity1
and entity3, entity1 and entiy4, entity1 and entity5.
Anchor Modeling use also surrogates, but the surrogates are not on db level, because the anchors are on the entity level. This means that anchor modelling has the same collapse but on the entity level.
In RM this case is forbidden. In RM it is not possible that two relations have the same attribute’s values. So RM is limited and obviously RM can not support this kind of problems (entities with same attributes).
The authors of Anchor Modeling have tried to fix all problems using my solution. They tried to apply the history of events in solving problems. However they didn’t understand many things in my solution.
Note that my solution enables that the OOM and RM applications have always complete solutions. I want to say, that just using my solution, OOM and RM for the first time can solve this case. As I already wrote, my data model belongs to General database theory.
7.
I started to write about this plagiarism on May 26, 2010 in my thread "The
original version".
On 8 June, 2010, in my thread “The original version” I wrote on this user group the following: “An identifier of a state of an entity allows decomposition of the concept of the state into binary concepts. The same hold true for relations. The identifier of the state of an entity provides straightforward mapping between the binary schemas as well as inverse mapping. “
In short, I tried to explain that the transition from E / RM to RM or OOM must be done using some mapping. I also wrote that you can not use tables for this mapping. Then I explained that my "identifier of state" determines completely "decomposition into atomic structures" and all these mappings.
On September 2010, the authors of Anchor Modeling submitted improved version of
Anchor Modeling. In this version the authors of Anchor Modeling are plagiarized
the most important part of my work and it is the identifier of a relationship.
Using my solution they have done "decomposition of the atomic structure" and the
mapping between two data models.
In addition, they solved problems for themselves which previously could not be
solved, the authors of AnchorModeling enables E / RM, for the first time, that
can do the mapping into RM, OOM, XML and other data models.
The technology that is applied to the publication of of the improved version of Anchor Modeling is very interesting. This paper is published in Jurnal DKE whose editor in chief is P. Chen. But the two probablly most important results in the theory of databases are not published in this paper but the following was given as a reference [29] and [30]. These references are in fact private website - with the address: www.anchormodeling.com
Note that my identifiers of states and maintaining of states were published in 2005 on my website. All this was presented on this user group and thoroughly discussed in 2005.
Five years before of publishing improved version of Anchor Modeling.
The following three things are key in the mapping between the two data models:
(i) Mapping between schemas
(ii) Mapping data between data models
(iii) Keeping of the meaning of the corresponding data from the corresponding
two data models
(i) and (ii) are completely determined by my “identifiers of states”.
Relationships between meaning, truth and facts are presented in my paper
“Semantic databases and semantic machines”, see section from 1.3 to 3.2.1
8.
In the Entity-Relationship Model – Toward a Unified View of Data, Chen wrote the
following: In a join meeting of the RDF and Schema Working Groups over one year
ago, they issued the Cambridge Communiqué that states: “… RDF can be viewed as a
member of the Entity-Relationship model family…”
The mentioned Communique can be seen at http://www.w3.org/TR/schema-arch Note that the decomposition into the atomic structures, which was done in my papers, enables that the atomic structures can be separately deployed on the www. The atomic structures also enable that current data, what are globally distributed and separated, can be easily and formally integrate into a global db. Obviously atomic structures, “history of events” and General database theory enable tremendous, possibilities on the www.
So my position is that databases are a key part of the solution to the global communication with the information, www is just a technical resource.
It seems to me that w3c philosophy is wrong; they think that it is opposite, that web resources are leading part of the solution. Note also that temporal and spatial knowledge in my data model has superior solution. My solution has complete history of events. These events can be related to many objects and subjects.
Vladimir Odrljin
I intend to end this thread, soon. Maybe I will add two or three posts, but it will be accelerated. So if you have some questions or comments if you notice some my mistakes, then let me know. After that I will be very busy. Received on Wed Nov 27 2013 - 19:12:52 CET