ISO/IEC
JTC1/SC32/WG2-EM0112-012
Date: 2001-12-14
REPLACES: -
The following is the text of
the Japanese comment, included here for ease of reference:
Begin Text:
In
this section (clause 4.8), according to the explanation of 'property', a
property is a characteristic of an object class. However, in the metamodel
diagram(Figure 5), 'Property' is just a (meta)attribute of the
'Data_Element_Concept', not a (meta)attribute of 'Object_Class'. So, if the
diagram is right, Object_Class cannot have a 'property' directly. Therefore, it
seems that there are some inconsistency between the explanation and the
diagram.
(It
might be possible to say that "an Object_Class may have a relation to 'a
property' through a Data_Element_Concept" because Data_Element_Concept
also have an attribute 'data_element_concept_object', ,and through this
attribute, an Object_Class and a 'property' may have a relationship. However,
even so, it seems unnatural to say that an Object_Class HAVE a 'property'.)
End Text.
The description of "Property" in clause 4.8.2 is
taken straight from the definition in clause 3. Note that this is the definition of 'Property', not of
"data_element_concept_property".
An Object Class has a set of Properties, and a particular
Property may be associated with more than one Object Class. The Data Element
Concept corresponds to a particular Property associated with a particular
Object Class, essentially resolving the many-to-many relationship. The
(meta)attributes on Data Element Concept are
"data_element_concept_property" (which identifies the Property) and
"data_element_concept_object_class" (which identifies the Object
Class).
A further complication is that some members want to be able
to register Data Element Concepts without defining the associated Object
Classes and Properties, and thus these attribute capsules have been made
optional to support this stated requirement.
1. Delete the
last sentence of clause 4.8.1:
" An object
class may reside in this structure but may not necessarily have a property."
This will be replaced by new text in clause
4.8.3.
2. Delete the second sentence of the third
paragraph of clause 4.8.2:
" A property may reside in this structure
but may not necessarily be associated with an object class or may not have a representational form (i.e., value domain) stored."
This will be replaced by new text in clause
4.8.3.
3. Delete the fourth sentence of the third
paragraph of clause 4.8.3:
"In other words, having a classification scheme of properties is optional."
This is redundant with the first
sentence.
4. Insert a new paragraph at the beginning of
clause 4.8.3:
"An object class may have many properties, and a property
may be associated with many object classes. This many-to-many relationship is resolved
through the use of data element concepts. For organizations that administer object classes and properties,
a data element concept represents a property
in the context of a particular object class.
5. Move the last two sentences of
the existing first paragraph to the end of the new first paragraph, with slight
modification.
"The union of a property and
an object class through the data
element concept provides significance beyond either that of the property or the object class alone. A data element concept thus has a definition
independent from the definition of
the object class or the property."
6. Change the remainder of the existing first
paragraph of clause 4.8.3 as follows:
"In addition, a data element
concept is a concept that
can be represented in the form of a data element, described
independently of any particular representation. For organizations that do not wish to
administer object classes and properties, the attributes data element
concept object class, and data element concept property are allowed to be
optional. Thus, a data element concept
may have zero or one object class and
zero or one property. "
[EDITOR'S NOTE:
Can we find an alternative term for "concept" in the sentence:
" a data element concept is a concept"? In the model, a "data element
concept" is not "Concept".
We are using the term "concept" in different ways.]
The remainder of clause 4.8.3 should remain
unchanged.