Re: some information about anchor modeling

Dana ponedjeljak, 11. veljače 2013. 08:41:14 UTC+1, korisnik Derek Asirvadem napisao je:

Hi Derek,

In the last few posts I wrote about the concepts and problems that were caused by Russell’s Paradox, related to Frege's definition of the concept. These problems affect the Set theory, which is the foundation of mathematics. Conceptual model is essential in my model, in which I use Frege's definition of the concept (see my paper from 2008 section 4.2.1. at http://www.dbdesign11.com ). But I have added a number of other things on this Frege's definition of the concept. I would like, in the next post point out the importance of these supplements on to conceptual modeling. In this post I want to point out the low level at which there is a current conceptual modeling. Here, first of all I think on the low theoretical level. Today, E / RM is taken as a paradigm for the conceptual modeling of databases. Under the leadership of people from E / RM, every year is organized the International Conferences on conceptual modeling, with honorary president P. Chen. At these conferences begins to form one large group. This group forms all those that have to do with the concepts and semantics - this primarily refers to E / RM, semantic db, semantic web, OO databases, and Ontologies.

1.
It seems that in 1976 Codd requested banning of publishing of Chen’s paper. At ER 2009 The International Conference on Conceptual Modeling, P. Chen had presented the paper, "Thirty Years of ER Conferences: Milestones, Achievements, and Future Directions." In the paper, according to Chen's words, "Codd wrote a long letter to the editor of ACM Transaction on Database Systems criticizing the author's paper ..." In this paper, Chen writes that Codd had begun to accept E / RM in his paper RM / T. Chen wrote in this paper: "Furthermore, in the 90's, the Codd and Date Consulting Group invited author to serve as a keynote speaker (together with Codd) several times in their database Symposia in London." (Here, "author" denotes P. Chen) Note also that on this ER 2009 conference, the paper "Anchor Modeling - An Agile Modeling Technique using the Sixth Normal Form for structurally and Temporally Evolving Data" is awarded as the best.

2.
It is true that P. Chen introduced data model that uses tables. However tables have two main disadvantages. They have no semantics, from them; you can not determine the meaning. Tables have weak mathematical tool. But the biggest weakness of the tables is that they did not agree with Frege's theory. In my opinion Chen did not understand that tables do not form predicates and propositions. Codd was done this part correctly. Codd's model has relations, while Chen's model has tables. In my opinion, both of them, Codd and Chen did not understand what connects concepts with predicates, and what separates them. Codd did not mention concepts in his RM. It is usual to treat E / RM as the leading conceptual model, although the concept is not defined in the E / RM. It is true that with the introduction of entities and relationships, is slightly improved semantic approach to db theory. But it was done very naive. In Chen's paper and the model, the concept has not been defined at all. If Chen called his model, conceptual, then an elementary thing to do is the definition of the concept. In contrast to Chen's work with the concepts, I have recently, on this user group presented a technique done by very good mathematicians. These techniques are related to the concepts and extensions, sets, membership relations, logical axioms etc. See my post that is related to Frege's Theorem. These techniques are superior to Chen’s conceptual model, both theoretical as well as practical solutions. This approach to concepts determines the bases of mathematics, i.e. it determines Set Theory.

We can notice that the entities, relationships and attributes were introduced long before the E / RM. For example, Kurt Gödel wrote in 1944: “By the theory of simple types I mean the doctrine which says that the objects of thought (or, in another interpretation, the symbolic expressions) are divided into types, namely: individuals, properties of individuals, relations between individuals, properties of such relations, etc. (with a similar hierarchy for extensions), and that sentences of the form: " a has the property φ ", " b bears the relation R to c ", etc. are meaningless, if a, b, c, R, φ are not of types fitting together. Mixed types (such as classes containing individuals and classes as elements) and therefore also transfinite types (such as the class of all classes of finite types) are excluded. That the theory of simple types suffices for avoiding also the epistemological paradoxes is shown by a closer analysis of these.”

Godel used term individuals instead of entities. Note that Godel used term "types" for entities, attributes and relationships. "Types" are just part of today's database theory. As you can see from this quotes, the main structures of the E / RM were completely determined yet 1944th Therefore we can not say that Chan introduced these constructs, rather we can say that Chen applied existing theory about attributes, entities and relationships to the theory of databases. Note that there are many philosophers, mathematicians and logicians who had been working on entities, attributes, and relationships, before than Chen is applied them to databases.

3.
P. Chen in his work on the E / RM, defines the entity as follows: An entity is a "thing" which can be distinctly identified. (See section 2.2 at Chen's paper).

Note that the entity and the "thing" are synonyms. Note that we can determine the difference between two entities, only by applying Leibniz's Law, not by using the mentioned Chen's "definition". Note that Chen determines entities by using intrinsic properties. On this user group, I presented example about 2000 Honda cars that have all its attributes the same. It means that (Chen's) intrinsic properties are not a solution. The surrogate key also can not help here; we must use the VIN number which is:

1. The real attribute of the entity Car;
2. The primary key (not a surrogate);
3. This is about db design. In my paper “Semantic databases and semantic machines” I introduced “General Law” (See section 5.6). This law enables very different picture about entities from Chen’s picture about entities.

4.
In Abstract of Chen's paper is writen the following: "The entity-relationship model can be used as a basis for unification of different views of data: the network model, the relational model, and the entity set model." However, this is not true, because Chen did not defined mapping from E / RM into RM, OO model, etc. Mapping between the two data models is a very complex matter and it is not universally resolved. In addition to the mapping between the two models, there is also the concept of translation of one model in another. This implies that it is not enough just to do the mapping between the two models. It is also necessary that the appropriate things from the two models must have the same meaning. So the translation should preserve meaning. Please note that this problem recently emerged as important because the translation has become important. For example, Google has translators for many spoken languages. In db theory we use formal languages. R. Montague, probably Tarski's the best student, is done a lot of things in the formalization of spoken languages. However, for translating, Frege's Principle of Compositionality is the most used. This crucial principle says that the meaning of a whole expression is composed of the meaning of its component parts. This matter is very complex and has many cases. For example, changing the order of words in a sentence changes its meaning. So, as a conclusion it can be said that the mapping and translation from E / RM to another data model has not been done, at all.

5.
Chen's E / RM is not correct. For example, in E / RM attributes of entities are not defined. Anchor modeling use normal forms. If we want to do normalization, then it is necessary to do mapping and translation from E / RM in RM. In this text I have already explained that Chen did not specify the mapping and translation, at all. So, there is the following question: who guarantees that a beginner (who believes to the claims of P. Chen) does not create an entity whose attributes, in fact, belong to two different entities?

6.
As I already mentioned, both Chen and Codd did not understand the essence and role of concepts. They did not understand what it is that connects a concept and the corresponding predicate. They also did not understand what it is that separates the concept from predicate. To explain this, we need a good understanding of Frege's work. I will simplify these things, so that it is understandable for those who did not familiar with Frege's work. Frege here has a big picture about very important things. He divides this matter on two parts, thoughts and language. The concepts belong to thought level, while predicates belong to the language level. A concept and the corresponding predicate is one thing which has different constructs at different levels.
(i) The construction of Frege’s concepts can be seen in my post on 26 Septembar, 2013 about Frege’s Theorem at:
 http://plato.stanford.edu/entries/frege-theorem/

(ii) The construction of Frege’s predicates you can see at my post on 4 September, 2013 at: https://groups.google.com/forum/?hl=en#!msg/comp.databases.theory/IfFnvnKoP4w/KkqT0DFeEzQJ
(In fact, we can say that predicates are grammatical constructs, they are about sentences and names)
In this manner I introduced “assignment” as a grammatical construct which bind names. For example we can bind the name of a variable to the name of a value. In languages ( programming and db languages) the assignment is only atomic command. In my db solution it is only possible to assign a new value or to “close” existing value (i.e. data).
(See my paper “Semantic databases and semantic machines”, section 7.3.
 

Here, the constructions of concepts and predicates are presented very shortly. To know complete Frege,s theory one must spent a lot of time. Note that Frege introduced these constructs as reality. Some scientists call it the third realm, i.e. the realm of semantics. According to Frege these objects are a realm and they enable semantics to us. Before Frege, there were two realm accepted in science: the realm of the external world and the realm of the purely mental.

Codd and Chen did not even define the concept. Apparently they did not realize the important relationships in conceptual modeling.

Vladimir Odrljin Received on Fri Nov 08 2013 - 18:02:53 CET