Re: The original version
Date: Mon, 3 Jan 2011 18:09:30 -0800 (PST)
Message-ID: <2636a874-693c-4c4e-bcfe-2d3a1b64c6c7_at_l8g2000yqh.googlegroups.com>
The first version of Anchor Modeling (from 2009) has the following
definition of an anchor:
Def 2 (Anchor). An anchor A(C) is a table with one column. The domain
of C is ID. The primary key for A is C.
In my paper from 2008 I paid special attention to defining a concept, an extension, an abstract object, the interpretation of real objects and their identification and connection to knowledge. As far as I know, my paper is the first instance in which these ideas are put together in such a way. My definition of a concept describes the construction of a plurality. The problem of identification is addressed in Section 5, but is also analyzed throughout the entire paper.
Among other terms, I defined a concept of an identifier. Two years after my paper from 2008 and after my public critique of the definition of the anchor as a table, the authors of Anchor Modeling published a new definition of the anchor in the second version of their paper (from October 2010). It is the following:
An anchor represents a set of entities, such as a set of actors or
events.
Definition 4 (Anchor). An anchor A is a string. An extension of an
anchor is a subset of I.
An example of an anchor is AC_Actor with an example extension {#4711,
#4712, #4713}.
So, the above definition of an anchor is written in 2010. My
definition of the concept of an identifier is written in 2008.
In this definition of an anchor there is some obvious misunderstanding
of basic things. An anchor cannot represent a set, as the authors
stated. A Venn diagram, for example, can represent a set. In the above
definition, an anchor is defined as a name. Names denote (not
represent) something and names have other significant roles in
semantics.
Also, we cannot say “a set of actors” because sets don’t contain
physical objects, rather abstract objects belong to a set. A set has
members and these members can denote actors.
It is bad design when one just puts numbers into a database and uses
these numbers as identifiers or as “surrogate keys”. This technique is
especially bad for databases that maintain history, because the
following two cases are possible:
1. one identifier can identify two distinct entities;
2. two distinct identifiers can identify one entity;
In my paper, I described how the process of identification should be done. In Anchor Modeling this part is not done at all.
As I have already written in this thread, the structures, which are named “Attributes” in Anchor Modeling, are special cases of binary structures from my paper. I explained this in my message from May 31, 2010 in this thread. I also wrote that these special cases are not always correct.
Vladimir Odrljin Received on Tue Jan 04 2011 - 03:09:30 CET