Re: some information about anchor modeling

I have written several times on this group, that the work of E. Codd's is very important for the theory of databases. But we can not say that E. Codd did something if he did not do it. So I wonder why C. Date wrote the following:

“...
ONE PRIMARY KEY PER ENTITY TYPE?
  I turn now to the second of the two issues mentioned in the introduction to this appendix: viz., that entities of a given type are supposed to be identified in exactly the same way everywhere in the database. What this means, loosely speaking, is that there’ll typically be:  

     A single “anchor” relvar for the pertinent entity type, having some particular primary key, together with  

     Zero or more subsidiary relvars giving further information about entities of that type, each having a foreign key that refers back to the primary key of that anchor relvar.  

 (Does this state of affairs remind you of the RM/T discipline discussed in Chapter 15?) ...”

First, RM / T is not discipline. There is no industry or theories, which are based on the RM / T. Second, I showed (on examles) in this thread, why is the surrogate key bad solution. Thirdly (the most important), RM / T has great theoretical drawbacks. For example, RM / T does not have anything related to the history of events, but without a history of events, the "binary decomposition" can not be done. Also, it is not clear how one transfers structures from ER model into RM model, using RM / T.(and vice verse). Do a binary decomposition is valid in the ER, using "invisible" surrogate? There are other theoretical problems in the RM / T, but it takes much more time and space.

"Anchor Modeling" has bridged some theoretical problems of the RM / T, using some of the results that I have published four years before them on this group. "Anchor Modeling" also uses the surrogate key and has some other incorrect theoretical results. Let me mention that it is allowed to delete some data in "Anchor Modeling." If someone has a delete operation and adding, then it implies that he has the update operation. And that means that there is no history.

Vladimir Odrljin Received on Wed Jan 02 2013 - 10:50:02 CET