'CapabilityConfiguration. A composite structure representing the physical and human resources (and their interactions) in an enterprise, assembled to meet a capability'

'System. An integrated set of elements, subsystems, or assemblies that accomplish a defined objective. These elements include products (hardware, software, firmware), processes, people, information, techniques, facilities, services, and other support elements (INCOSE SE Handbook V4, 2015).'

1) As defined CapabilityConfiguration in including interactions and providing a capability is closer to being able to represent a System than the System element itself
2) CapabilityConfiguration is a legacy MODAF element. In the early days of MODAF in the MOD Architectures Lab it was common place to use the CapabilityConfiguration element to represent a system. The MoD never really addressed the problem and left the two distinct elements. There seems to be no good reason to have both or, if there is, what is the differentiation [remembering that the definitition shouldn't refer to any other metamodel element as it must be atomic and self-contained. Being connected to different other metamodel elements doesn't mean that the type is distinct - this can be corrected by changing the metamodel wiring to the CapabilityConfiguration or System element.

'No new terms and definitions have been required to create this specification. All terms are available in the normative references or bibliographic citations for detailed explanation'

This is clearly untrue - there are terms, such as 'View Specification', which if distinct from Architecture Viewpoint needs to be defined to differentiate it.

As an authorative standard that claims to have semantic underpinnings this is wholly unacceptable - what is the master source to be used/referenced in an implementation? There is no traceability. From an engineering perspective any design implementation must be provided with the means to create full traceability to the UAF spec and, where appropriate, the particular source definition.

Definition: 'A type of a human being used to define the characteristics that need to be described for ActualPersons (e.g., properties such as address, telephone number, nationality, etc.).'

1) 'sub-' anything is never a type. The 'sub' simply indicates a position in a hierarchy but the type is the same.

2) An Organization is not a Human Being

3) The definition of a type should be atomic and self-contained - it cannot depend on any other element such as 'ActualPersons'

Organization is defined as 'A group of OrganizationalResources (Persons, Posts, Organizations and Responsibilities) associated for a particular purpose'. This shouldn't include other metamodel elements for the same reason i..e not OrghanizationalResource etc.

Fig 7:3
Definition - 'A Tuple denotes a relationship that exists between elements.'

This is incorrect - it defines a relationship (as it says). A tuple has to also include the start and finish nodes as well to form the smallest tuple - a triple.

Looking at the diagrams it is the legend definition that is incorrect because the green coloured elements are relationships not triples. A clue is in Fig 8:28 on p49 ActualResourceRelationship is identified in green and is clearly just a relationship.

'Further, The UAF conforms to terms defined in the ISO/IEC/IEEE 42010 standard for architecture description, where the terms .... form correspondence rules specified as constraints on UAF.'

This is incorrect. Terms / definitions cannot form correpondence rules.

It isn't obvious whether there are any correpondence rules in the UAF given that it is the definition of the individual AF that defines the rules for the content of any particular view - the UAF just provides the "kit of parts" to enable the user to create them. If the UAF does itself define its own correspondence rules this would need to specified separately and somehow deal with potential inconsistencies wrt what does conformance etc then mean.

Definition: 'A type of a current or future state of the enterprise.'

This looks as thought it's a legacy MODAF 1.2.003 definition.

1) A state is an abstract reference to a set of property values or conditions. A phase is a reference to a duration. This would be better defined as a distinct stage in the life? of an Enterprise. The problem is otherwise that there is no connection between the element name and definition.
2) 'current' or 'future' is irrelevant and incomplete - if it occurs in the past is it therefore no longer an EnterprisePhase? The definition shouldn't depend on a particular time value.
3) 'type of' shouldn't be in the definition. All the elements in a metamodel are types so it cannot form part of any differentiating feature.

toBe: Boolean[1] Indicates whether the ArchitecturalDescription represents an Architecture that exists or will exist in the future.

This is incorrect. This is a property of the Architecture not the ArchitectureDescription i.e. an ArchitectureDescription of a future Architecture not a future ArchitectureDescription of an Architecture.

This is another consequence of not differentiating properly between 'architecture' and 'architecture description'.

Definition
'Interest in an EnterprisePhase (EnterprisePhase is synonym for System in ISO 42010) relevant to one or more of its stakeholders'

1) This doesn't conform to the ISO 42010 definition of a Concern and hence invalidates claim to be compliant
2) An EnterprisePhase is not a synonym for System, and certainly not in ISO 42010

In Fig 9:211

1) the 'concern' role name placed against Concern on the Viewpoint - Concern relationship is meaningless - using the same label as the name of the adjacent node element is in effect leaving the relationship with no name whatsoever - it proviode no information about that relationship. In ISO 42010 this relationship is named 'frames' - this should be . The cardinality is incorrect - it should be 1..*.

2) the relationship between Stakeholder and Concern should be named 'has'. The cardinality is incorrect - it should be 1..*

3) There should be no relationship between ActualEnterprisePhase and Concern

'A type used to denote a model of the Architecture, described from the ResourcePerformer perspective.'

A model is not an architecture - it might, however, describe an architecture description. The definition is therefore that for a type of architecture description not architecture. The definition is wrong and needs to be changed - suggestion is to base on the ISO 42010 conceptual model - 'architecture'.

It is also an error to include in any definition any relationship - definitions of elements should be atomic / self-congtained and not depend on any other element or relationship otherwisse a) you're not defining the element - it might be a triple b) if the structure/ relationships change then the element definition no longer holds.

'The UAFP can be used to develop architectures compliant with:
• Department of Defense Architecture Framework (DoDAF) version 2.02'

This is incorrect. Architecture Descriptions might be compliant with an AF such as the DoDAF but architectures them selves are certainly not. This is a general problem - the lack of differentiation between 'architecture' (the thing being described) and 'architecture description' (the description). It probably isn't helped by loose casual everyday talk but in a document with defined semantics it is an error - it's the equivalent of confusing the 'duck' with the 'photo of the duck'.

It also looks as though this has affected some of the metamodel element definitions e.g. ResourceArchitecture

Reported in behalf of Boeing.

Continuing the progress we have been doing on getting rid of tag properties and replacing them by dependency type of relationships, there are still some tag properties to be replaced between the view, viewpoint, concern, architecture description, and stakeholder. Having tags, they cannot be properly displayed on the diagram, where dependencies can.

the term DataElement is no longer used. should update definition of ResourceExchange.conveyed from:
"ResourceCommunication, the conveyed element must be stereotyped «DataElement» or its specializations,"
to
"ResourceCommunication, the conveyed element must be stereotyped «ResourceInformation» or its specializations,"

We keep getting error while importing. It tries to load the file and fails when trying to read the first character, with a "invalid byte" message. Not sure if this has been reported before. If so would it be possible to share updated UAF.xmi? If not, would it be possible to share the name of the tool it was created with?

I learned that the OMG has deprecated UML 1.x XMI 1.x well over 10 years ago when it was replaced by UML 2.x XMI 2.x.

Also learned that the OMG UAF.xmi file is in UML XMI version 1.x. Probably generated from a tool like Sparx Enterprise. Because UML 1.x is an outdated version it is not possible to import the current UAF.xmi file posted on your site for analysis, causing import errors.

This requires providing the UAF.xmi files using the current UML 2.x XMI 2.x version.

Please provide/replace the UAF.xmi files posted on your site with UML 2.x XMI 2.x version to import the specification in to a modeling tool for analysis.

p213 Viewpoint is defined as 'An architecture viewpoint frames (to formulate or construct in a particular style or language) one or more concerns.'

1) Doesn't conform to ISO 42010 'work product establishing the conventions for the construction, interpretation and use of architecture views to frame specific system concerns'
2) This definition doesn't define what an Architecture Viewpoint is - only one of the relationships with Concern. Definitions should be atomic and not include related elements otherwise a change in model structure invalidates the definition. Also, semantically, they have to be standalone.

3) The element should be named 'Architecture Viewpoint' as the definition states. This is important as pointed out earlier to differentiate it from the casual use of 'Viewpoint' and anchor it to ISO 42010.

4)There is a general problem with diagrams where role names are used which have the same name/label as an adjacent element - see Fig 9:215.
The role names provide no additional information and in effect you have unlabelled relationships. In this particular case the relationships should take the names from the 42010 conceptual model. 'language' should be 'architecture description language'. Unclear why 'purpose' is there as distinct from framing one or more concerns i.e is the purpose different from framing the concerns, if so, how?

5) 'View' should be 'Architecture View' to differentiate it from other meanings e.g. as a collection of view definitions.

6) The cardinality between View and Viewpoint is wrong if conforming with 42010 - it should be 1:1 between an Architecture View and its governing Architecture Viewpoint.
7) To conform with ISO 42010 the relationships should take the same names i.e. Viewpoint governs View, Viewpoint frames Concern (and should show Stakeholder has Concern)

8) There is an error in ISO 42010 conceptual model replicated - the requirement for an Architecture Description to contain one or more Architecture Viewpoints - Architecture Viewpoints may be held externally as what used to be termed Library Architecture Viewpoints.

On page 4 the document states 'The UAF conforms to terms defined in the ISO/IEC/IEEE 42010 standard for architecture description'. It also states in section 4 Terms and Conditions 'No new terms and definitions have been required to create this specification. All terms are available in the normative references or bibliographic citations for detailed explanation'

Despite these claims the document frequently uses 'viewpoint' with a different meaning from that of ISO 42010 i.e. not as a specification against which an architecture view is prepared and interpreted but:-
1) 'viewpoint' = collection of architecture view definitions having a common subject area (e.g MODAF::Viewpoint)

It does this with no identification of any difference in meaning and nor any definition. It often uses the different senses of 'viewpoint' in the same sentnce and paragraph:

'The aligned and renamed viewpoints from the various frameworks provide a common generic name for each viewpoint. It should be noted that the term viewpoint is in the context of ISO 42010 where a viewpoint is the specification of a view. The UAF viewpoints are mapped to the corresponding viewpoint in the relevant contributing framework. It is the viewpoints described in the DMM that provides the basis for the Unified Architecture Framework (UAF)...'

It looks like the 'viewpoints' and 'UAF viewpoints' are collections of view definitons which is confusingly and erroneously compared with the ISO 42010 definition wjhich is not the same thing at all. It also looks like the last use of 'viewpoints' might be a collection but by this point it's impossible to tell because the authors do not define the terms that they use.

In 7.2 UAF Grid it presents what seems like MODAF::viewpoint, ISO::architecture viewpoint and UAF::view specification in the same place/graphic.

p5 Type 1 Conformance 'Type 1 Conformance: - UAF View specification conformance. A tool demonstrating view specification conformance shall implement a version of all the view specifications defined in the UAF Grid, with the exception of the view specifications in the Metadata Domain. Optionally the tool vendor can implement other donor framework viewpoints, for instance DoDAF, MODAF or NAF ..'

Here 'viewpoint' = collection not a single specification. 'view specification' appears to be indistinguishable from ISO 42010:architecture viewpoint

It is extremely confusing and far from clear and non-ambiguous which are key qualities of any good requirement.

No where in the document is 'view specification' defined despite it being used in 7.2 and as a package name in the metamodel.

In ISO 42010:2011 an 'Architecture Viewpoint' is defined as 'work product establishing the conventions for the construction, interpretation and use of architecture views to frame specific system concerns' and in the ISO 42010 conceptual model is has a 'governs' relationship with Architecture View i.e. an Architecture Viewpoint is a specification containing a set of requirements against which an Architecture View is prepared and interpreted. How then is 'View Specification' different? This is particularly the case as 'View Specification' seems not to merit a UAF metamodel element.

(typo) Page 72, View Specifications::Resources::Roadmap::Resources Roadmap: Evolution has a Stakeholder named "Implements" and I think you mean "Implementers" as is used in other areas.
(typo) Page 83, View Specifications::Projects::Taxonomy::Project Taxonomy doesn't put the Stakeholders, Concerns, Definition, and Recommended Implementation on separate lines like every other View Specification does.

Page 31, View Specifications::Operational::Connectivity::Operational Connectivity and Page 37, View Specifications::Operational::Constraints::Operational Constraints both have a Stakeholder named "Architects" whereas all other View Specifications specify one or several types of Architects. Are all other named types of Architects assumed?

In general, what are your definitions/assumptions for Business Architects versus Enterprise Architects versus IT Architects, all of which are used as Stakeholders in various View Specifications?
What are your definitions/assumptions for Solution Providers versus Implementers, both of which are used as Stakeholders?
What are your definitions/assumptions for Decision Makers versus Executives versus Owners Responsible for Operational Agents, all of which are used as Stakeholders in various View Specifications?

Page 37, View Specifications::Operational::Constraints::Operational Constraints contains a Stakeholder called Program Sponsor. What is your definition for "Program" or that overall Stakeholder?

Portfolio management involves changes to the future plans and current deployments.When to add new capabilities? When and where to change operations or change resources?. When to retire legacy resources, or re-purpose them?

Overall “score” for alternative portfolio configuration is extent to which Outcomes are achieved for the enterprise as a whole. Possibilities for change come from identified Opportunities and their associated Risks. Also these come from identified new/modified Capabilities and their associated Effects.

Portfolio is normally divided into Portfolio Segments
-Could be divided by timeframe (eg, “epochs” such as near, mid, far term)
-Could be divided by “color of money” (eg, R&D, acquisition, operations)
-Could be divided by mission area

Currently using Whole Life Configuration as the Portfolio elements in EA model. WLC is a “set of Versioned Elements”. But would be more convenient to have a different model element for “Portfolio”.
-Alternative Portfolios would be examined for most cost-effective combinations
-Each Portfolio option is a provisional “set of Versioned Elements”
-Then the selected combination gets promoted to a new Whole Life Configuration (becoming the new “baseline” for program planning and scheduling)

Portfolio (kind of) Project does not have the necessary semantics with the features and relationships enumerated below.

WLC element only gives part of the picture of what is in the Portfolio.
-Acts as a “collection” of things with an associated Gantt chart for deployment timelines
-Does not tell us who “owns” that collection of things in the Management Accountability sense
-A Portfolio can consist of multiple Whole Life Configurations

Portfolio needs to have the following features:
-Things it is delivering (eg, in its Whole Life Configurations)
-Who runs it (ie, the Actual Organization)
-Particular programs within it (eg, actual Projects or project elements)
-Alignment with budget elements, assigned Opportunities and Risks, etc
-Assigned responsibility for execution of particular Enduring Tasks (and associated Capabilities)
-Accountability towards achieving Effects (eg, MOE target levels) for a given time period
-Can be assigned responsibility for certain Missions, Enterprise Goals and Enterprise Phases
-Can be jointly managed with another Enterprise for shared Operations and Resources

Might be that Portfolio could be a kind of Whole Life Enterprise (ie subtype) that has the features and relationships noted above

SysML allows the use of CallOperationAction as an alternate means of mapping Behaviors (via Operation->Method metaproperty).
UAF 1.1 includes OperationalActivityAction, ProjectActivityAction, ServiceFunctionAction, FunctionAction or SecurityProcessAction which all extend CallBehaviorAction.
UAF 1.1 also includes OperationalMethod, ServiceMethod, ResourceMethod, which extend Operation, and even has metaconstraints requiring appropriately UAF stereotyped Activity used as the Method.

We recommend allowing CallOperationAction to also be stereotyped by the UAF extensions to CallBehaviorAction listed above.
We also recommend creating ProjectMethod and SecurityMethod as extentions to Operation to be consistent with other Domains.
This would facilitate it's use in Operational::Connectivity, Operational::Processes and Security Processes, Project Processes, Personnel Processes, Resource Processes and Resources Connectivity Diagrams and Tables.
This is especially useful to visualize and take advantage of CallOperationAction target pin and method resolution per UML 2.5 13.2.3.5 Behavioral Features and Methods. It is also useful for fUML 1.2.1 Polymorphic Operation Dispatching ability to execute method resolution during simulation and analysis. Thus a tighter integration of UAF with UML method resolution semantics and executability can be achieved both visually and from a simulation and analysis perspective.

ISODateTime is currently mapped to literal string. In my opinion it does not make sense. This mapping does not allow to reuse dates in the easy way. Plus we also have the same approach with milestones implemented very differently. I do believe that both approaches need to be aligned. Dates should be easily reusable all over model.

Actual Resource Relationship is described as: An abstract element that details the ActualOrganizationalResources that are able to carry out an ActualResponsibility.

This appears to be incorrect. First, this element is not abstract. Second, it should not be restricted to Organizational Resources since it should apply to any Actual Resource. Third, it does not help understand responsibilities of organizational resources.

Description of this element should be changed to read something like: An element that details the ActualResources that are involved in exchanging resources through a realized ResourceExchange.

In the UAF DMM document (https://www.omg.org/spec/UAF/1.1/DMM/PDF), starting right after "7.2 Domain Interrelationships", there seems to be an error causing the Table of Contents (TOC) to incorrectly "Domain Metamodel Diagram Legend" as Chapter 8 insead of section 7.3. This offset continued through TOC the document such that Chapters 8 and 9 are mislabeled.

In the traceability appendix, the undefined "Services Sequences" view is referenced in the DoDAF and MODAF tables, mapping to the SvcV-10c and SOV-4c, respectively. I assume this should be "Services Interaction Scenarios".

The view traceability tables in the body of Appendix A are not numbered in a logical order. They're shown as Table 2:1,1:2, 2:2, 2:3, 2:5. The numbers in the List of Tables (p. v) appear to be correct.

The elements in the Actual Resources views are instances of the "definitional" elements in other domains. For example, Actual Resource (classified by Resource Performer element types in Resources domain); Actual Post, Actual Person, Actual Organization (<all classified by elements from Personnel domain); Actual Resource for a Resource Mitigation (from Security domain). The new Resource Service element will be classified by a Service in the Services domain.

Notice that not all the elements shown in Actual Resources views come from the Resources domain views (ie, they also come from Personnel, Security, and Service domains). So, the name "Actual Resources" is somewhat misleading in this respect. Therefore, it would better if this domain is renamed to more accurately reflect its intended content. Suggest the name be changed to "Actualization" (Az) domain.

Identify and update the DMM to reflect new concepts and relationships to existing concepts. Additionally, identify all impact to existing diagrams.

Add support for Value Streams in UAF. Please consider Value Streams implementation as described in Business Architecture Metamodel. Provide support for strategic value streams, value stream stages, before/after relationship between different value stream stages, value networks, operational value streams, and outcome (value) produced by the value stream.

Reported in behalf of Boeing