There is a general lack of composition relationships for model management – in both the RDF and OWL metamodels

The xmi for the RDF profile has been substantially simplified for ODM 1.1, but the section of the RDF profile in the document does not match the xmi. This should be corrected in the 1.2 RTF.

This section describes problems with MOF/UML due to the need to work around multiple classification issues, which was fixed by SMOF. The specification needs to be revised to (1) state that no changes are required to other OMG specs, (2) update references and add SMOF, and (3) add a discussion of the need for SMOF in the design principles section (section 7)

References in the profile section need to be revised to reflect UML 2.4.1 / 2.5 documents, and any new references need to be incorporated into chapters 3 and 19.

Multiple Domains or Ranges for Properties. Section 16.4.1.4 (Multiple Domains or Ranges for Properties) says that multiple domains or ranges for properties is equivalent to the intersection of the domains and ranges. UML properties have at most one type, and intersection can't be represented in UML without the profile (Chapter 14). How is this translated?

complementOf and disjointWith. Section 16.4.1.3 (Mapping for complementOf and disjointWith) says UML has constructions for complementOf and disjointWith in the PowerTypes pacakge. It actually has constructs for unionOf and disjointWith. Section 16.4.1.3 says no mapping is given because the OWL constructs are pairwise, but OWL unionOf and disjointWith are not pairwise, they can apply to any number of classes.

Inferring subsumption. In Section 16.5.1 (Predicate Definition Language), first sentence, UML can support subsumption reasoning also, see http://www.inf.unibz.it/~calvanese/papers-html/AIJ-2005.html

Figure 16.1 incomplete. Figure 16.1 (Key Aspects of UML Class Diagram) is missing the multiplicities on general/specific, and the subsetting between ownedEnd and memberEnd.

Range Restriction Restriction Classes. The introduction to Section 16.4.8 (Range Restriction Restriction Classes) says the translation is to a comments. But AllValuesFrom translates directly to redefinition of property types, see the profile (Chapter 14).

Constructed Classes. The introduction to Section 16.4.6 (Constructed Classes) refers to OWL "difference". I assume this is supposed to be complementOf. The introduction to the section says intersection can be mapped to subclass relationships, but this isn't true, at least not without the profile, see intersection in Chapter 14. It also says union can be translated to subclass relationships, but doesn't mention UML generalization sets and isCovering, see Section 16.3.10 (Powertypes).

Anonymous Classes. Section 16.4.4.3 (Anonymous Class to Class) can translate blank nodes to anonymous classes in UML.

Ontology Properties. Section 16.4.3.2 (Ontology Properties to Comments) should use dependencies for some of the translations. See the profile (Chapter 14).

Properties in OWL. The end of Section 16.4.9 (Properties in OWL) refers to multiple domains be ing equivalent to the domain being an intersection. This does not translate to UML, see issue on Constructed Classes

The current conformance section includes requirements that have been made invalid by removal of the RDFWeb package in the metamodel, among other things. The entire section needs a rewrite to reflect changes made for RTF 1.1.

navigableOwnedEnd. The introduction to Section 16.3.5 (Binary Association To Object Property) accounts for navigableOwnedEnd, but the introduction to Section 16.3.8 () Association Generalization) does not.

Table 16.12, classes as instances. In 16.2.3 (More Advanced Concepts), Table 16.12, class as instances appears in both this table and Table 16.11.

Table 16.10. In 16.2.3 (More Advanced Concepts), Table 16.10, the names of classes are capitalized in UML. The UML element corresponding to OWL subproperty is property subsetting. N-aries and association classes are not well-supported in OWL, so don't belong in a table of common features (see other issues on n-aries and association classes).

Classes of classes. In 16.2.3 (More Advanced Concepts), seventh paragraph, the second sentence implies classes are not instances in OWL DL, but even in DL, OWL Class is a class of classes, by definition. For example, an ontology of animals might have the class Dog, which is an instance (of OWL Class) and a class (of Fido, Rover, and other individual dogs). Ther third sentence should be moved to be the second, and start with "however"|, because it is an exception to the first sentence. After "declaration" should be replaced wtih "a common superclass".

Table 16.9 and Naries. In 16.2.2 (Class and Property - Basics), Table 16.9 replace the "Parts" header with "Properties". The Reification property isn't necessary, because AssociationClass is both a class and association, there is no separate reification of the association (this is necessary in OWL DL, however, and even in OWL Full, some extension is needed for a subclass of Property and Class to correspond to a UML Association Class). The text below the table uses the term "implements" which doesn't apply (these are platform-dependent models), and introduces the reified association, which doesn't exist in UML.

Multiplicity. Section 16.3.7 (Multiplicity), the translation can also be to OWL FunctionalProperty or InverseFunctionalProperty if the multiplicity is 1.

Names, unique names. In Section 16.5.2 (Names), the first two paragraph implies UML assumes unqiue names. M1 instance specifications in UML can have different names, but refer to the same M0 individual. They can also have the same name and refer to different M0 individuals. The third paragraph implies UML does not have name management (given the title of Section 16.5), which of course it does in namespaces.

Individuals, mapping. Section 16.4.1.1 (Mapping for Individuals), first sentence says the profile (Chapter 14) represents individuals as a singleton class. This is incorrect. The profile models individuals as instance specifications. To give property values to the individual, the profile uses a singleton class. Section 16.4.1.1 incorrectly concludes that individuals should not be mapped, which affects 16.4.1.2 (Mapping for Enumerated Classes) and Section 16.4.13 (Annotation Properties to Comments).

Boolean combination. In Section 16.5.1 (Predicate Definition Language), third sentence, UML supports the equivalent of unionOf.

Associations. In 16.2.2 (Class and Property - Basics), in the paragraph below Table 16.7, gives the wrong translation to OWL for UML associations. UML associations have properties at end, and these are often navigable. Binary associations in UML translate to two inverse properties, using these property names, not the association name. See the UML profile for OWL for the translation options for associations, and the third paragraph in 16.2.3.

Distinct associations, restrictions. In 16.2.2 (Class and Property - Basics), in the paragraph above Table 16.7, says the OWL properties "arising" (I assume due to translation) from a UML model are distinct, that OWL restrictions aren't in the translation. UML can redefine properties in subtypes of the classes where the property is introduced, which is equivalent to restriction. The method employed in the chapter is not adequate.

Distinct associations, ownedAttribute associations. In 16.2.2 (Class and Property - Basics), in the paragraph below Table 16.6, there is the sentence " Note that UML ownedAttribute M2 associations are distinct, even if ownedAttributes have the same name associated with different classes." What are "M2 owned attribute associations"? In the case of M1 properties, properties with the same name may be on different classes, but if they inherit from the same base class where a property of that name is introduced, then they are the same property from OWL's point of view. There is usually no no need to translate to unique OWL properties, just restrictions. See next issue.

Associations. In Section 16.2.1 (UML Kernel), the discussion around Tables 16.2 through 16.4 seems to be about relational implementations, rather than UML modeling in the sense that is important to OWL. My suggestion is to replace Tables 16.3 and 16.4 with the tabular forms of the metamodel, as in 16.2. The paragraph above Table 16.3, first sentence, modeling associations does not depend on the implementation of classes (the "implementation" usually refers to how the model is translated to a platform). Same comment on the second sentence, which says Table 16.2 is an implementation, when it is only a tabular form of the metamodel. The second sentence refers to the disjoint union of attributes, but there's nothing like this in UML.

The diagram (Figure 9.1) and related text is out of date given the revisions made to the ODM metamodels by the 1.1 RTF.

The ODM Metamodel takes a different approach to OWL restrictions from the Profile (and indeed from OWL): the Profile has a single stereotype Restriction whereas the Metamodel has 6 different subclasses depending on the type of restriction: HasValueRestriction, AllValuesFromRestriction, CardinalityRestriction etc. It would be more consistent if the metamodel had only a single class, though this would necessitate constraints on the properties.

The specification is sorely in need of examples showing how to represent common RDF/OWL constructs as instances of the metamodel. That’s especially the case for use of URIs and IDs; and also for anonymous classes as used in Restrictions and Intersections

Summary: Constraints in the OWL Metamodel Chapter (10) should be specified in OCL.

Description: Text based descriptions of constraints provided in chapter 11 with the OWL metamodel should be specified in OCL.

Summary: Constraints in the RDF Metamodel Chapter (10) should be specified in OCL.

Description: Text based descriptions of constraints provided in chapter 10 with the RDF metamodel should be specified in OCL.

Via the resolution to issue 16495, the ODM metamodels and profiles for RDF and OWL were revised to support internationalized URIs (IRIs). Chapter 17, which covers the mapping from Topic Maps to OWL, has not been updated to reflect this modification however. The chapter needs to be updated to be brought in line with these changes.

Currently, the ODM 1.0 specification defines several stereotypes for representing an OWL ontology in UML using the ODM stereotypes for RDF and OWL.

The ODM spec is understandably updated to support OWL2, the current recommendation from the W3C.
It is not entirely clear which OWL2 constructs are supported in the ODM profile — a cross-reference table linking the entries of the quick ref. guide would be particularly helpful: http://www.w3.org/TR/owl2-quick-reference/

Also, in practice, it would be useful to have the flexibility of showing come axioms but not others.
For example, if we have (in functional syntax):

Declaration( Class( A ))
Declaration( Class( B ))
Declaration( Class( C ))
SubCassOf( A B )
SubCassOf( B C )

Then, an OWL2 reasoner will infer the following axiom:

SubCassOf( A C )

Using the ODM profile, it should be possible to show selected subsets of an ontology.
For example:

view1:

Declaration( Class( A ))
Declaration( Class( B ))
Declaration( Class( C ))
SubCassOf( A B )
SubCassOf( B C )

view2:

Declaration( Class( A ))
Declaration( Class( B ))
Declaration( Class( C ))
SubCassOf( A B )
SubCassOf( B C )
SubCassOf( A C )

view3:

Declaration( Class( A ))
Declaration( Class( C ))
SubCassOf( A C )

This brings up the question of adding support in the ODM profile to distinguish asserted vs. inferred axioms.
Perhaps there could be a flag — e.g.,

isAsserted : Boolean = true // set it to false for an inferred axiom — or--
isInferred : Boolean = false // set it to true for an inferred axiom

Finally, additional markup may be useful — e.g., showing whether an ontology is consistent or not.

Users creating domain ontologies want their models to be user friendly and this requires phrases with spaces and other special characters. The use of “camel case” and other I.T. conventions are not friendly, however OWL has restrictions on the characters that may be used.

Potential resolution: ODM should specify the algorithm for mapping a user friendly names in the UML profile to an OWL legal name where required. The user friendly name can and should be used in the OWL label and does not require such mapping. The choice of algorithm can be to eliminate the space and enforce camel case or to substitute underscores for all illegal characters. My preference would be to introduce underscores as these are then easier to reverse map from OWL to UML and are visually similar to spaces.

Stereotypes should be shown on diagrams in an abstract syntax section under each profile, not only in text under each stereotype.

From email dated 3/12/2008 from SRI, and as discussed (and documented in the minutes from the ODM FTF2 F2F DC meeting: Section 14.2.6.5 - Simple association with properties at the end is a nice readable notation. However, the "brotherOf" property between the two classes in Figure 14.28 could be duplicated on an association between two other classes on the same diagram, but the would be unrelated in the UML model, whereas in OWL they would be a single property with multiple domains and ranges. (This comment applies also to similar graphical representation shown in other sections). So – this is true. It is managed in UML via the namespace of the relation, which may assume that you're not trying to determine all possible values with each property. The example is not a good one and could lead to inconsistent interpretation. We should get a better example. Also, we need to decide what the interpretation of the role name is, when you have mutiples (when you assume that it is or is not in the same namespace).

Specification: Ontology Definition Metamodel
FormalNumber: ptc/06-10-11
Section: 18
Summary: The mapping from RDFS and OWL to CL should be revised to reflect metamodel changes in CL due to finalization of ISO 24707.

Description: Minor changes were made to the CL language as it was finalized through the ISO process, which are not reflected in the ODM specification. These changes also need to be reflected in the mapping (embedding) description contained in chapter 18.

In the CL metamodel the associations Conjunction and Disjunction clash with class names. This is not strictly speaking an error at MOF 2 but can cause difficulty for some implementations. And these do not make good associations names. I propose: ConjoinedSentence and DisjoinedSentence (which will make them consistent with NegatedSentence).

The current ODM specification is lacking stereotype definitions for containers and collections. The entire section in the profile amounts to one sentence that references an annex. The section should be revised to provide stereotypes for these elements.

The profiles submitted with the RTF report include stereotype definitions that are not in the submitted RTF report itself. These represent work in progress that the RTF felt should be left in the submitted profiles but that are subject to change/reconciliation in the 1.2 RTF.

Via the resolution to issue 16495, the ODM metamodels and profiles for RDF and OWL were revised to support internationalized URIs (IRIs).
Chapter 16, which covers the mapping from UML to OWL, has not been updated to reflect this modification however. The chapter needs to be updated to be brought in line with these changes.

The set of implementations referenced is out of date and should be revised to discuss current implementations.