Stated: "DoD's doctrine, organization, training material, leadership & education, personnel, and facilities (DOTMLPF)"

Error: "training material"

Corrective Action: Entire acronym, properly stated should read: DOTmLPF-P: Doctrine, Organization, Training, materiel, Leadership and Education, Personnel, Facilities and Policy.

Definitive Source: JCIDS Manual, 31 August 2018

Applicable Proponent: J7/JIB, on behalf of the Joint Staff Director for Joint Force Development (DJ-7)

Actual Resource - Traceability is described as a View Specification on pages 101-102, but that square in the Grid is black.

Within UAF Appendix B - Sample Problem, v1.2, Figure 9-8: Personnel States for SAR Rescue Team represents a number of states with multiple do actions in a list. This may be a tool limitation, but State Machines in Cameo EA and Rhapsody are limited to one do action per state.

The definition for "SubOrganization" does not address its meaning/use; instead it uses the same definition as for "Person".
(see also related reported issue)

Requested to provide a list of deprecated elements for new version in order to note gaps and make necessary refactorizations.

The Domain Metamodel Elements (Section 9.1.1) included Architecture Management::Sequences elements, but the Architecture Management::Sequences diagrams were not included in Domain Metamodel Diagrams (Section 8.1.1) and Figure 7:1- UAF Grid.

The abstract metaclass "Process" in the UAF metamodel is represented by the abstract stereotype "Activity" in the UAF Profile.
This is confusing, The "Activity" stereotype should be renamed to "Process" to achieve consistency.

The UAF specification defines Security Control as a kind of SysML Requirement. However, the specification also references NIST SP 800-53 in the definition of Security Control. NIST SP 800-53 clearly separates Security Controls from Requirement. E.g., in section 2.1 Requirements and Controls of NIST SP 800-53r5 it says "...It is important to understand the relationship between requirements and controls." It goes on to say "...the term requirement is generally used to refer to information security and privacy obligations imposed on organizations." It defines controls as "...descriptions of the safeguards and protection capabilities appropriate
for achieving the particular security and privacy objectives of the organization...". In other words, Security Controls are a kind of capability.

Current implementation of information elements (both Operational and Resource) does not allow for decomposition, which would allow for multiple layers of fidelity to be represented in a model. Most information elements in a real system are "decomposable" in the sense that they can be represented by parts (in the case of a document, multiple data elements may be represented in one Operational element; in the case of a Resource information element, headers and multiple payloads may be represented within a higher level "information" element. The data model is particularly important within the EA reprsenetation of a system.
Generalization is also needed within the information elements; flow properties on intefaces should accept elements of a less generalized implementation.

The EA Guide and the formal UAF documents suggest that UAF can be used to capture non-security risks at various levels of abstraction of the AD. The view specifications and UAFML limit the mitigation of a risk to a security control. It appears that an operational or resource mitigation should be elements that are able to mitigate a risk. The current metamodel only allows risks to affect those elements. Some rework in this area would improve the ability to capture non-security risks for ADs. Please see attached pdf for additional information.

Operational Performer is what does an activity in your model: Person, Organization, Service, and only if-and-only-if fully automated - a System. There's no linkage in UAF between these elements, and will thus cause duplication and/or configuration item bloat when you assert that Operational Performer named CAOC/CAOC Director sends ATO to Wing/Squadron/Mission Planner, when CAOC, Wing, Squadron are all Organization (types), and CAOC Director and Mission Planner are Person (Roles).

View specification for <<Standards>> elements do not allow for directed composition into other <<Standards>>. Specification for <<Standards>> elements do not specify a part property.
This is needed because many ConformsTo relationships conform to only a specific part of a standard, rather than the entire standard. Traceability is enhanced by allowing for a UAF Element to ConformTo a standard at a lower level than the parent.

if one looks at the other *ExchangeItem descriptions...

there is a *Parameter that has a type as an *ExchangeItem which is missing.. (it is in the description of the ServiceParameter)...

also, it is missing the derived relationships between ServiceExchangeItem and ServiceFunction

Signals should be ExchangeItems... since Exchanges are ItemFlows from SysML...
Signals should be included

The description says "A dependency relationship denoting that a Risk is drawn out by an Opportunity." I have tried three different online translators that presents different meaning of "is drawn out by". Can the description be made more crisp and less ambigous for a non-native English speaking audience?

DataElement is obsolete and should change name to ResourceInformation.

ClassificationAttributes all Strings are missing multiplicities

think this is wrong description for BillingItem

Properties indicating the assurance of a piece of information

endDate : ISO8601DateTime[0..1]
End time for this ActualProjectMilestone

should have multiplicity of [0]

missing definition of Cost for BillingItem

The change of name from the UAFP to the UAFML is technically incorrect. The UML or SysML are modelling languages because they can be used to model many real world things and relationships - the UAFML is fixed and it only represents a particular UML (and SysML) representation of the UAF DMM. It is a UML profile as stated in the text - 'It was created by mapping the UAF concepts and relationships to corresponding stereotypes in the UAFML Profile.' So now we have a UAFML profile (aka UAFPP ...)

There is and can only ever one example/individual use of the UAFML. It cannot be used for any other purpose. It is not, therefore, a modelling language and defining it as such does a disservice to the UML et al.

The statement - 'Due to the complexity of managing the multiple viewpoints with overlapping concerns and metamodels' ... needs to be clarified. Concerns never overlap - viewpoints may share concerns. I think what is being described is that donor AFs may have shared concerns - if so the sentence needs to be qualified because within a UAF you woudn't expect shared concerns because it is then poor viewpoint design and leads to inconsistency if the user can't easily select the same viewpoint to use (too much overlap).

It's all the more confusing because Fig-8:2 Architecture Views states that a zero or more Concerns are (framed?) by a single Viewpoint so it's not even possible for viewpoints to share a concern.

Centre- rather than left- or fully-justified text

42010 uses Architecture Description whereas in UAF we have what seems to be the same concept but is 'ArchitecturalDescription'.

The model in Figure 3.21 is missing relationship names between ArchitecturalDescription and View, ArchitecturalDescription and Architecture, ArchitecturalDescription and Viewpoint. These should in accordance with 42010 be 'has part, or is composed of', 'expresses' and 'has part or is composed of'.

It should be noted that architecture descriptions using the UAF do not include architecture viewpoints - the viewpoints used are separate. In old 42010 terminology the UAF uses what is called 'library' viewpoints to save having to include in each and every AD. The relationship is something like 'references' or a form of trace.

It is poor practice not to label relationships and repeating the target stereotype name as a role name on a connector adds no information.

The figure also misses the association with ArchitectureFramework that is mentioned in Associations.

The architectureFramework association doesn't, as stated, indicate the type of architecture framework - it identifies the particular architecture framework.

3.1 lists Normative References, 3.1.1 lists normative OMG references. Non-OMG normative references should still be part of 3.1 and hence should be 3.1.2 not 3.2

There is also a problem in the wholesale application of normative references to any part of the UAF. The DMM is supposed to be implementation-agnostic and therefore ADL references should not be normative for that. Similarly the only place where MODEM and DODAF etc are referenced are in the Traceability document (perhaps the UAFML?) so they are not normative for the DMM itself - particularly as there are elements in the DMM that are not in the MODEM, or DODAF or MODAF [this is also a problem within the Traceability document - it isn't complete].

It would make sense to qualify the scope of the normative references by splitting into sub-sections e.g. general (common to all), DMM, UAFML, Traceability - if you're going to place all of the references in the one place. Alternatively add a References section to each sub-document and only state the normative references that apply to the document being read.

The statement is made 'The following normative documents contain provisions which, through reference in this text, constitute provisions of this specification.'

This isn't true.

UML profile for BPMN isn't normative in the DMM - it might be in the UAFML which defines a UML profile for the UAF but ADLs such as the UML, SysML, BPMN, IEPPV, BMM cannot be normative for the DMM which is supposed to apply to non-UML implementations. The UML spec is only informative for the DMM in as much as it conceivably (impossible in practice) might provide advice on how to read and interpret the notation used.

Any wiki, for example the IDEAS one, cannot be normative. For a start there is no defined baseline date or issue let alone the fact that it doesn't state any requirements. At best this particular one is an informative reference.

The OMG identifier on the front cover is formal/22-07-03 which turns out to be a SysML specification according to the issue tracker.

However using the URL https://www.omg.org/cgi-bin/doc?formal/22-07-03 seems to want to return the UAF specification.

If the OMG identifier for UAF 1.2 DMM is formal/22-07-03 then section 1 and Table 1-1 contain the wrong OMG identifiers (dtc/.... rather than formal/...)

There is no stereotype defined for ResourceInteractionScenario and the other variants of InteractionScenarios (OperationalInteractionScenario and OperationalInteractionScenario).

I expected stereotypes for these concepts in the same way that there are stereotypes defined for the different varieties of StateDescriptions.

ResourceInteractionScenario is defined within chapter Domain MetaModel::Personnel::Sequences in the specification
The correct chapter shall be Domain MetaModel::Resources::Sequences.

On figure 8:79 Standards Taxonomy we appear to have something like UAFElement conformsTo Standard. It is unclear because 'conformsTo' appears to be a role (which is incorrect if so), the multiplicity should be 0..* not *.

What provides the 'conformsTo' relationship required?

Relationships on the diagram aren't named. The multiplicities aren't correct - they should never be * (1 is always permissible). 'ratifiedBy' role is naming a relationship - as the target element is Organization the role name ought to be 'ratifyingOrganization' i.e. characterising the target node.

There appears to be no conformsTo UAF element. The relationship on the diagram isn't named. The other fly in the ointment is that someone has copied SysML for the Requirements diagram which does show a 'Satisfy' relationship. So 'Satisfy' vs 'conformsTo'? This is one of the problems in not separating implementation from the DMM (the DMM shouldn't simply repeat the names of SysML, UML etc elements since there's no reason that they represent the same thing and you need the ability to diconnect as this is supposed to be ADL agnostic.

How is this view specification supposed to be implemented when relationships aren't labelled and there appear to be elements missing?

The use of a '*' multiplicity is incorrect - should be 1 or more Concerns. Similarly an ArchitecturalDescription (name error - should be ArchitectureDescription) should be composed of 1 or more Architecture Views (not *). A Viewpoint may frame only 1 Concern (not *)

The use of a direction indicator is undefined and looks to be incorrect. In the direction from View to Viewpoint what is the relationship? In ISO 42010 there is a 'governs' relationship.

The use solely of role names is incorrect - roles only label the target (source) node - they do not define a relationship. How is anyone supposed to implement a set or relationships where the relationships are not labelled? Figure 8:2 should use the relationship names defined in 42010 i.e. standardise.

Using a role name that simply repeats the name of the source / target element is pointless - it adds no information whatsoever. The other problem with solely using role names is that if the label is offset it isn't clear whether it is a role or a relationship name and therefore whether it applies to the node or relationship.

Inconsistent name wrt ISO 42010 and with its own definition - ArchitecturalDescription vs Architecture Description.

The multiplicities aren't correct. For example we have on Architecture to ArchitecturalDescription we have * on both ends which means that an Architecture has no existance independently of its ArchitectureDescriptions (plural as there have to be many of them). Should be 0..* on ArchitecturalDescription. Similarly with View it ought to be 1..* View and 0..* Viewpoint (otherwise an ArchitecturalDescription is required to include multiple Viewpoints which doesn't allow for a central set of 'library' viewpoints in ISO 42010 terminology).

The relationships with View, Viewpoint, Architecture need to be named.

'expresses' looks like a role but cannot be. If it is a role it needs to qualify the target element e.g. 'describedArchitecture' not the relationship.

Why does ArchitectureMetadata have a multiplicity of 1 on ArchitecturalDescription - is each piece of metadata unique to every ArchitecturalDescription?

What is an ArchitecturalReference? This looks like a relationship. What then is the point of adding a role name to a relationship? Isn't this just 'UAFElement traces to ArchitecturalDescription?

The concerns identified are - 'Concerns: captures activity-based behavior and flows.'

Figure 8:92 only defines a class hierarchy mapping UAF elements to BPMN elements. Ignoring that this is ADL-specific and shouldn't be in the DMM, this doesn't deine how process flows, activities are described and linked together. If such a viewpoint is felt necessary the UAF elements (not the BPMN elements) should show the triples that are permitted to be used in a conforming view.

As defined this, the NIEM etc view specifications are a) not defined and b) impossible to implement in a non-UML ADL.

Within 8.1.15 there are a number of ADL-Specific View Specifications. The DMM is supposed to be (but isn't) ADL-agnostic. It is impossible for any non-UML implementer to use or implement this content. Given that ADL-specific implementation is elsewhere e.g. the UAFML for UML,SysML then if there is a need to define this should be in a separate document and not the DMM.

Within the document there appears to be no usable link or definition of Satisy [another implementation artefact] and no definition of a Requirement element or other relationships linked to Requirement (Trace, Verify, Refine).

Figure 8:86 multiplicities don't look to be consistent - 0..1 on Trace but 1 on Satisfy. Trace looks to be wrong - think someone trying to describe Requirement traces to Requirement, UAFELement traces to UAFElement and UAFElement traces to Requirement but there is no identification of path and its ambigous or wrong.

Why is Requirement not a UAF Element? If it's not why is it even in this specification? Again I suspect this is an implementation artefact - someone thinking of a SysML::Requirement (which is an implementation of the DMM)

8.1.2 View Specifications::Summary & Overview::Summary & Overview

General problem - there is no such thing as 'view specification' - it doesn't exist in the DMM itself (text should only include defined concepts), it seems to be being used as a synonym for ISO 42010::'architecture viewpoint' (if this is the case then 'architecture viewpoint' is the correct and consistent term to use.

Concerns: ''quick overview of an architecture description and summary of analysis. In the initial phases of architecture development, it serves as a planning guide. Upon completion of an architecture, it provides a summary of findings, and any conducted analysis.'

Problems -
1) this is an overview and its application not a list of concerns held be stakeholders
2) Given the common confusion in the UAF between 'architecture' and 'architecture description' - is this 'phases of architecture development' or 'phases of architecture description development'? Since an architecture description may be used to provide comment on architecture this is ambiguous in the UAF contetx where these terms aren't differentiated.
3) 'It isn't completion of architecture - it's not even completion pf 'architecture description' - what the author seems to be stating is completion of the 'architecture task' that produce the architecture description. The description os wrong and possibly more triples need to be defined in the DMM

Defintion
'provides executive-level summary information in a consistent form that allows quick reference and comparison among architectures. The Summary and Overview includes assumptions, constraints, and limitations that may affect high-level decision processes involving the architecture.'

Problems;
4) This is a description not a definition

Figure 8:11 - Summary & Overview

The green elements are relationships - not tuples as defined in the legend. The expectation is that view content consists of triples (node-connector-node).
5) General problem - There is nothing that defines how these views are to be interpreted (ISO 42010 requires this and it would aid consistent development and provide rules for verification of each view).

Anything that doesn't form the basis for a triple shouldn't be in the viewpoint definition i.e.

6) ArchitecturalDescription, ArchitectureMetadata, Metadata as these don't appear to form any triple

7) View, Viewpoint. There are no relationship elements provided to form a triple with ArchitecturalDescription

8) Stakeholder, Concern, OrganizationalResource, ActualOrganizationalResource have no relationship with ArchitecturalDescription

9) There is no relationship element between ArchitecturalReference and Architecture so it is impossible to establish a link between the Architecture, its impacts etc and the ArchitecturalDescription so impossible to describe this.

Triples should be able to be read as sentences and have clear semantics:

'ArchitecturalDescription Architectural Reference Architectural Description'
10) What does this mean? Is this a trace between ArchitecturalDescriptions i.e. 'ArchitecturalDescription traces to / references / etc ArchitecturalDescription'. A currently defined this makes no sense. I suspect that the problem is that elements are added but the triples so-formed are not being read to make sure that they are intelligible i.e. object-focussed rather than relationship-focussed approach.

11) 'Opportunity Phases ActualStrategicPhase' - what does this mean? If the sentence is unclear it either won't be used or be used inconsistently as each individual attempts to try and understand it (not conducive to shared understanding)

[Note - owing to inconsistency in form behaviour between specification and OMG Document number the original was misallocated to SysML as SYSML17-647 which can be deleted]

The definitions of ActualEnterprisePhase, WholeLifeEnterprise and EnterpriseMission are:-

'ActualEnterprisePhase - A time period within which a set of Capabilities are deployed
WholeLifeEnterprise - A WholeLifeEnterprise is a purposeful endeavor of any size involving people, organizations and supporting systems. It is made up of TemporalParts and StructuralParts

EnterpriseMission - Mission captures at a high level what you will do to realize your vision.'

Figure 9:164 defines:-

1) WholeLifeEnterprise (endeavour) is a ActualEnterprisePhase (time period) - this is incorrect
2) EnterpriseMission (what you do) is a ActualEnterprisePhase (time period) - this is incorrect
3) The definition of ActualEnterprisePhase is not atomic - it includes Capabilities. Type definitions should be atomic, not depend on anything else (otherwise you're defining all or part of a triple and if the external element or relationship named changes the definition is invalid)
4) The definition of WholeLifeEnterprise is similarly non-atomic - it should not include reference to any relationship e.g. temporal or structural for example.

UAFElement is defined: 'Abstract super type for all of the UAF elements. It provides a way for all of the UAF elements to have a common set of properties.'

This doesn't define the concept. It simply states what purpose the element serves for a developer. The name tells us nothing. In the metamodel the names are supposed to indicate what they represent.

Do all UAFElements appear in an architecture view and therefore architecture description? if so it might really be 'Architecture Description Element'. This term first appeared in ISO 42010:2011 and in 42010:2022 is an 'identified or named part of an architecture description'. Alternatively it might be 'An individual architecture description object that is used to describe or represent an item of real-world architecture. An architecture description element can appear in an architecture description.' <https://trakmetamodel.sourceforge.io/vocab/TRAK_metamodel.html#Architecture_Description_Element>

This is a consequence of the lack of differentiation within the UAF (and donor AFs) between 'Architecture' and 'Architecture Description'

Fig 9:122 Metadata.
Definition: 'A comment that can be applied to any element in the architecture. The attributes associated with this element details the relationship between the element and its related dublinCoreElement, metaDataScheme, category and name. This allows the element to be referenced using the Semantic Web.'

it is clear that this refers to 'architecture description' not 'architecture' e.g. the application of DCMI elements

It should therefore be 'applied to any element in the architecture description'

If Metadata is already defined as applying to an architecture description, how can ArchitectureMetadata specialise this and also be applied to the architecture description description - this is what Metadata does and as part of an AD which describes an architecture there is no need for 'ArchitectureMetadata'?

Figure 8-14 only shows Capability, CapabilityRole elements.

The definition of the view specification is: 'Definition: shows the relationship between EnterprisePhases and the Capabilities that are intended to be developed during the enterprise phases, and the organizations involved in the enterprise.'

It is impossible to construct a view using Figure 8-14 that satisfies this because:-

a) No EnterprisePhase (should this be ActualEnterprisePhase?) nor element to describe relationship(s) with Capability
b) What has CapabilityRole to do with this view specification? As defined this seems unsolicited and isn't connected to the definition
c) Even with Capability, CapabilityRole elements provided the view specification doesn't include any relationship elements ('tuple') in green to connect Capability to anything else

In establishing a mapping there are 3 possible outcomes (gap analysis):

1) Target element is not linked to any UAF element
2) Target element is linked to a UAF element
3) UAF element is not linked to any target element

The Traceability document only considers 2). This is inadequate. For example a statement is made concerning linkage to MODEM but this is only true up to the point where MODEM was released - UAF elements added which are not present in a 'donor AF' such as Requirement will not be linked to MODEM. Without a full list of all of the UAF elements its difficult to do a completeness check and to spot potential errors. It's important for an implementer to understand what might be in the UAF that isn't in a donor AF - particularly where the donor AF is no longer being maintained and over time therefore there will be more new-to-UAF elements.

Figure 9:134 – UAFElement shows a single attribute - URI

Don't UAFElements have a name, identifier (other than URI), description etc? According to this they don't.

URI is incorrectly defined - 'Captures Unique identifier for the element.'

assuming that URI = Unifiform Resource Identifier - which a url-form of identifier (W3C define this so the definition ought to be standards-based rather than local). It isn't just an identifier. 'captures' shouldn't be part of the definition - that's how the attribute is used.

Note: this repeats the misallocated https://issues.omg.org/issues/SYSML17-646. Issues SYSML17-645, SYSML17-647, SYSML17-648, SYSML17-649 should also be allocated to the UAF not SysML (problem where drop-down list of Specification doesn't correspond to the OMG document number]

Missing - Satisfy, Refine, Verify, Trace, Requirement. These are UAF elements (unless it's possible to conform without using them). These should be listed in Table 4.1

1) Satisfy, Refine, Verify show an Association to both UAFElement and Requirement and appear to implement a many:many relationship. The Associatioons do not, however, define a direction for the relationships. As this is supposed to be a normative specification are the relationships implemnted as bi-directional, uni-directional? This is a problem in partially defining a (relationship) node 'from node' and 'to node' rather than node - relationship - node with direction on the relationship itself.

2) There are two identical Associations shown between Requirement and Trace - same multiplcities, not labelled. It is impossible to distinguish them apart and all they do is define that there is some unnamed relationship in some (or both) directions. I suspect that what was meant was a UAFElement traces to Requirement and possibly Requirement traces to UAFElement but this isn't what the diagram says. Had a node - connector (with direction) - node style been used the differentiation problem wouldn't arise and errors would be more easily spotted.

ServiceArchitecture is defined as 'An element used to denote a model of the Architecture, described from the Services perspective.'

1) This doesn't define what service architecture is
2) ... denote a model ... almost everything in the UAF metamodel is an AD Element i.e. something that describes some real world concept and forms part of an AD (a model). The definition is supposed to define the real world thing that it is used to represent not itself.
3) Architecture is essentially the set of temporal, spatial, functional etc relationships that something has [internally and with anything external]. ServiceArchitecture is not a specialisation of Service ie it is not true to state that 'ServiceArchitecture is a Service'. The class structure is not therefore correct. I suspect what has happened is that this is a bottom-up derivation which enables behaviours and or properties to be inherited that are useful in a UML implementation. That doesn't make it correct however. Occam's razor should apply in these cases.

The UAF defines an identifier for each of the UAF diagrams. These are listed, for example, in the Traceability doc Table 2.1.

Instead of using anique identifier each of the 'view specifications' has what looks like a package structure string i.e. 'View Specifications::Resources::Structure::Resources Structure' should have the identifier - 'Rs-sr' so the identifying string is - 'Rs-sr - Resources Structure'

Using a package containment as an identifier isn't correct. If you move a 'view specification' to another package the package string changes. The identifier, however, doesn't and hence this should be what is iused to identify a view specification. Any text string taken from the package containement structure is a beneficial addition but it doesn't identify the view specification.

The identifier used for the traceability document in 7 UAF Grid of 'dtc/21-12-10' is incorrect.

1.1 item 5 states 'The approach defined in this Guide is just one way to approach architectures when using UAF' which should be 'The approach defined in this Guide is just one way to approach architecture description when using UAF'

The use of 'architecture' to sometimes refer to 'architecture' and sometimes refer to its description ('architecture description') is a perniscious problem in the UAF. It might be acceptable in casual conversation but technically the terms from 42010 should be used correctly. This is particularly the case since the DMM states that 'The UAF conforms to terms defined in the ISO/IEC/IEEE 42010 standard for architecture description' - it doesn't. See also 'view specification' which is an undefined and unnecessary concept.

1.2 we have 'UAF extends the scope of UPDM and generalizes it to make it applicable to commercial as well as military architectures' which should also be 'architecture description' not 'architectures'

Metadata attributes - DCMI

1) category. Refers to DCMI abstract which isn't a category
2) dublinCoreTag refers to 'A metadata category that is a DublinCore tag.' This doesn't define anything. Any DCMI tag? Not all CMI tags are categories. Why not simply list those DCMI tags that you think are essential / useful rather than what looks to be an arbitrary and inconsistent reference?
3) Metadata is defined as 'a conment that can be applied to any element in the architecture'. This is incorrect - it can be applied to any AD element. The DCMI tags only apply to artefacts - video, text, sound, document etc so they don't apply to every AD element (UAFElement) as many/most of these do not represent artefacts. DCMI tags only apply to documents, standards etc.

[The original ticket was misallocated to SysML as SYSML17-649 which can be closed / deleted]

Requirement is not a UAFElement - p133, p103

Figure 8:86 permits UAFElement Trace Requirement and UAFElement Refine Requirement

It isn't therefore possible to present 'Requirement Refine Requirement' nor 'Requirement Trace Requirement' triples in a View Specifications::Motivation:Requirements view.

Type 1 Conformance exempts - 'with the exception of the view specifications in the Architecture Management Viewpoint'.

An (architecture) 'viewpoint' is a specification against which a view is prepared and interpreted etc iaw ISO/IEC/IEEE 42010. It is an error to use 'viewpoint' to mean a collection of architecture viewpoints

1.1 states:- 'UAF Modeling Language (UAFML) (this document formal/22-07-05) provides the modeling language specification for implementing the UAF DMM using UML/SysML'

and 'The UAFML defines a set of stereotypes and model elements and relationships to satisfy the requirements of the UPDM 3.0 RFP and the UAF DMM.'

1) There is no evidence provided wrt how each DMM AD element is implemented i.e. for each DMM AD element what UML or SysML stereotype has been chosen to implement the DMM AD element. A traceability table is needed for the UAFML specifically that a) shows matched (traced elements) and b) shows unmatched SysML/UML stereotypes which might be added for other reasons, for example to add wanted tool behaviours, but which have nothing to do with DMM AD element conformance. Such a traceability table must be required in any case to verify against the DMM and trap potential implementation errors.

2) As the UAFML is an implementation I would expect class inheritcance hierarchies separate from view content. It's almost impossible to use this specification where there are tiny fragments of what must be a unifying model (somewhere)

3) The statement is made 'The UAFML defines a set of stereotypes and model elements and relationships ....'. Given that a UML profile defines a set of UML/SysML stereotypes what then are 'model elements and relationships' - don't the UAFML stereotypes define node and relationship elements? If not what are these other elements, why do they exist? This might be explained or justified by the traceability table in 1) but it's an odd statement that suggests that something is missing. A relationship is a model element i.e. not just nodes.

View Specifications::Operational::Sequences::Operational Sequences View Specification:

'Concerns: express a time ordered examination of the operational exchanges as a result of a particular operational scenario.

Definition: provides a time-ordered examination of the operational exchanges between participating nodes (OperationalPerformer roles) as a result of a particular operational scenario'

1) Figure 8:29 includes no triples to define the order of exchanges of operational messages. Even the incorrectly included UML Lifeline does not define an order - it is inferred by visual presentation reading from, say, top to bottom or annotating with a sequence number. There are no formal semantics that define the order of a sequence. In any case the UML shouldn't be in this document - it is a specific implementation. The missing part in this view specification is something that semantically describes order or=f exchanges (which the UML etc is then able to implement) i.e. something like' OperationalExchange follows OperationalExchange'

2) The concern isn't 'time-ordered examination' - it is something like 'in what order ....'

3) The definition should be phrased in terms of real world things not the metamodel representing the real world. It might include the application of the viewpoint to real world situations. Simply describing the metamodel provides no utility.

Another consequence of the failure within the UAF to distinguish between 'Architecture' and 'Architecture Description' concepts.

The title of the viewpoint (not 'View Specification') is ' - View Specifications::Architecture Management::States::Architecture Status

'Stakeholders: Enterprise Architects, people who want to understand the architecture governance, Technical Managers. Concerns: architecture status.

Definition: captures version number and approval workflow of the architecture. Recommended Implementation: SysML State Machine Diagram, state table, text.

ArchitecturalDescription

status : String [*]

...

Figure 8:5 - Architecture Status'

Figure 8.5 shows the ArchitectureDescription element - NOT any ...Architecture element. It is clear that the viewpoint describes the state of the architecture description not architecture.

Hence NOT 'architecture governance' - should be 'architecture description governance', 'approval workflow of the architecture description', The name of the viewpoint should be 'Architecture Description State' since 'St' = state and the subject is 'Architecture Description' not 'Architecture'

Figure 8:29 Operational Sequences shows 3 UML Metaclasses - Message, Interaction and Lifeline.

The DMM is supposed to be agnostic and therefore should include no UML, SysML, BPMN et al ADL elements. The UML/SysML implementation is supposed to be in formal/22-07-05 UAF Modelling Language

Table 2:1 - UAF 1.2 to DoDAF 2.02 Mapping at the bottom. The Operational Structure viewpoint identifier is shown as 'Op-St' which duplicates the Operational States Viewpoint identifier. it should be 'Op-Sr'

8.1.1 View Specification - Am-Pr Architecture Development Method is defined:

'Stakeholders: Enterprise Architects, Model Managers, Modelers, Enterprise Systems Engineers. Concerns: development sequence of models and views and how they are related to each other. Definition: defines workflow or process steps used in managing the architecture development. Recommended Implementation: SysML Activity Diagram, text.'

Figure 8.4 shows a single element - ArchitecturalDescription. No relationship elements. No architecture views,

1) The subject description of views, their sequence etc is not 'architecture' it's 'architecture description'. The view specification name is therefore not correct - the subject doesn't appear within the view specification name - it should be 'Architecture Description Method'
2) There are no elements / triples provided to describe the sequence of models, views and how they relate to each other hence Fig 8.4 does not provide the means to address the concerns. Presumably there should be at least the means to describe a trace. To describe a sequence of views you need something like 'View follows View'
3) Figure 8.4 only shows a single attribute for ArchitecturalDescription - see Figure 9:129. It ought also to include purpose, assumptionandConstraint etc (all of the ISO/IEC/IEEE 42010 required properties)
4) It suggests using a SysML Activity Diagram but provides no triples to describe Activities or their sequence.
5) If models and their sequence are expected you need to provide the means to describe them.
6) This view specification does so little that there seems to be little point in its existance - it would be easy to incorporate the concerns into another Am view specification.

Figure 8:10 - Architecture Traceability.

The (Am-Tr) Architecture Traceability view specification has a number of problems:-
1) Figure 8-10 only allows one relationship (in green) - ...Architecture implements ...Phase - what is this supposed to mean - there is no direction ? How does this address the understanding of 'the impact of change'?
2) Under stakeholders - 'people who want to understand impact of change across the architecture supporting assets,' is a concern not a role
3) Under concerns - 'reuse of architectures' - given the widespread conflation of 'architecture' and 'architecture description' does 'architecture' here mean 'architecture'?
4) ArchitecturalDescription is show yet there is no allowed relationship with it. As it does not contribute any tripes that are allowed it should not be shown in the Figure
5) Under definition we have 'shows references to ... external sources, e.g. documents'. There are no relationships or artefact elements shown to enable these references to be described
6) Typo on role on Architecture - 'realizingArchitedcture'

Section 7 UAF Grid. Fig 7-1 and note e) refer to a 'Architecture Extensions' view specification:
'The Architecture Extensions view specification provides a means to extend the framework to other domains'

1) There is no 'Architecture Extensions' view specification in 8.1 View Specifications
2) The name 'Architecture Extensions' is incorrect - the subject is not the architecture but the architecture description (it describes means to extend the architecture framework not the architecture'. The name should more accurately be 'Architecture Description Extensions' or 'Architecture Framework Extensions'.

The tables Table 2:1, 1:2, 2:2, 2:3 mapping the various AF viewpoints to the UAF view specifications show a Am-Tx Architecture Extensions view specification that doesn't exist. Until the view specification has been defined these rows should be deleted.

Table 1:2 has a table number that doesn't fit with the previous 2:1 table and the following Table 2:2 identifier. The identifiers for the tables 1:2, 2:2 and 2:3 need to be corrected.

9.1.1 Domain Metamodel::Architecture Management - Definition / Figure 9:118

The 'Definition' metamodel element title has somehow got merged into the preceding Figire 9:118 caption - 'Figure 9:118 – ArchitectureMetadataDefinition'

9.1.6 Domain Metamodel::Personnel - Person, Figure 9:234

Person is defined as '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) This doesn't define what the 'Person' concept is - it only defines how it is used for an implementation that seeks to be able to represent a real world person. Every Class/Type defines properties applied to an individual - this is not an identifying feature.

2) In most modelling languages such the UML etc we have a means to represent a type (Class) e.g. UML Class and an individual (instance of a UML Class). The individual / instance automatically takes on the propertties of the type / Class.

If, for example, I instantiate a UML Class named 'Car' that has a property, 'manufacturer' the resulting element is an individual which we understand to represent a real world thing, Car. I can then allocated the manufacter's name to the 'manufacturer' property. None of this requires an ActualCar - all of the Class/Type properties are defined as part of the Class/Type - there is no new type or stereotype needed to define Class properties.

For example if I have a model element of Type/Class = Standard and then give it a DCMIIdentifier = 'ISO/IEC/IEEE 42010' and DCMITitle = 'Interational Standard - Systems and Software - Architecture Description' everyone understands that this is a representation of the real world artefact - ISO/IEC/IEEE 42010. There is only the one Class/Type which defines the properties that make the Class unique from all other types. Class definitions have to be unique - if they depend on another Class then they're not. There is no ActualStandard required.

ActualPerson is not a type of Person - as defined it only exists to hold a set of properties that can be applied to an individual or instance of the Person Class/Type.

This separation is completely inconsistent with every other UAF metamodel element - we don't have ActualArchitecturalDescription, ActualArchitectureViewpoint, ActualStandard et al. This looks to be an old error inherited from early MODAF days where someone was thinking of abstractly representing instantiation but made the mistake of embedding this in what was a UML implementation (the M3).

The 'Actualxxx' construct isn't needed in the UAF. Nor is it needed by any implementation of the UAF in any other ADL. On p288 we have ActualPerson = 'An individual human being.' - it isn't - it's supposed to be a type of Person. This makes no sense since the metamodel presents types.

It adds more elements, relationships, complicates implementation, adds inconsistencies and increases the cost of maintenance for no good reason.

From a practical means to define properties of individuals there is no technical need for ActualPerson, Actual-anything.

The DMM looks to be a bottom-up merging of donor AF metamodels warts and all rather than a top-down what-do-I need-to-address-each-viewpoint-concern. This has not only not preserved errors but does not produce the smallest metamodel needed (very much like constructing programme task logic from the start/left rather than starting with the end deliverable and working to the left).

9.1.5 Domain Metamodel::Services - ServiceRole , Figure 9:215 - ServiceRole

'A behavioral feature of a Service whose behaviour is specified in a ServiceFunction'

a Role is something that is played by typically an Actor. i.e. the relationship is something like Service plays ServiceRole (if the metamodel names are accurate).

Instead what is shown is:-
1) a role name of 'type' on Service - Service is a type of ServiceRole?
2) whole:part relationship - ServiceRole is a part of Service - it shouldn't be a whole:part if a role is played by the element
3) How does anyone in a non-UML or even UML ADL implement this view specification - the only 3 green relationship elements provided are ServiceMessage from/to ServiceRole, ServiceConnector from/to ServiceRole, the other end of PerformsInContext isn't shown. There are many possibilities (OperationalRole, OperationalActivityAction, FunctionAction, ResourceRole, ServiceFunctionAction, ServiceRole) - Figure 9:107 - which one or ones are permitted for this view specification? This is a problem in not presenting whole triples i.e. there should be no case where both the from and to node elements aren't also shown.
4) role names on ServiceMessage should be on ServiceRole not the ServiceMessage connector element
5) The definition of ServiceRole is not atomic and hence incorrect - in depends on the existence of a relationship with a ServiceFunction element hence is not independent and defines a triple.

9.1.10 Domain Metamodel::Standards::Structure , Fig 9:236 ProtocolLayer

ProtocolLayer is defined as 'Usage of a Protocol in the context of another Protocol. Creates a whole-part relationship.'

1) ProtocolLayer (as identified by 'another' and 'usage') is not therefore distinct from Protocol and shouldn't therefore be a metamodel element - it should be represented by a whole:part relationship on Protocol.

2) ProtocolLayer is not a type of Protocol - which I what I think the incorrect role name on the Association from ProtocolLayer to Protocol is stating. whole:part + type = ?

3) Any metamodel element including 'usage' is likely to be invalid. Usage is a particular UML implementation and often a fudge. No metamodel element is a usage of any other. An elementis its own thing - if usage is involved it is not a distinct concept.

1.1 Overview of the Unified Architecture Framework - Modelling Using the UAF. states:-
'Modeling Using UAF. The UAF Modeling Language 1 (UAFML) is an implementation of the DMM that specifies how the UAF views can be modeled using the SysML notation and semantics. Even though the UAFML is based on SysML, there are some significant differences that should be noted. SysML is great for doing the following activities:
(a) modeling systems and for doing systems engineering,
(b) defining and tracing between levels of abstraction within a system,
(c) defining the logical and physical attributes for a system and the mapping of requirements and functions to these attributes. The UAF Modeling Language provides all this, plus more:
a) Capability and Enterprise Concepts: defines the “why” and “what” and “when” before the “how”
b) Services Concepts: definition of enterprise services (producing and consuming) and traceability to capabilities, operations and implementing resources
c) Human Factors: How people and systems interact, and their expected knowledge & skills
d) Security: Identifying risk, its mitigation, and integrating security into the architecture
e) Standards: definition of and compliance with standards in the architecture
f) Project Deliveries: phased milestone approach to capability deployment
g) System Configuration Over Time: deployment and changes in roadmaps and timelines
h) Tie-in to Non-System Elements in the Architecture: Easy way to link the entire Architecture to Requirements
i) Built-in Traceability Between Multiple Views: Between Layers and Across Layers'

The problems with this are:-
1) The title is modelling using the UAF but most of the subject matter seems to be the UAFML
2) This is a user guide - the user will never use the UAFML directly. The UAFML, more correctly the UAF (UML) Profile XML (), contains only UML and SysML implementations of the DMM elements. It does not contain any UAF architecture views nor does it define what triples appear in each UAF architecture view. This is implemented by the respective tool vendor. If the user did load the profile they would see a large set of UML/SDysML elements and they'd have to figure out which to use to produce a conforming UAF architecture view. The UAFML does not 'specifies how the UAF views can be modeled' - it simply defines which UML and SysML elements are used to implement the DMM elements needed.
3) The UAFML isn't a modelling language - it's a single use of a modelling language (UML and SysML). It cannot be used as a modelling language by the user. It is a UML profile so new terms shouldn't be created for well defined and understood concepts. Are the OMG now renaming UML profile to 'Modelling Language'? This just adds inconsistency and shows a lack of the importance of standardisation. The language being used is still the UMNL or the SysML not the UAFML.
4) d), e), h) incorrectly use 'Architecture' which should be 'Architecture Description'
5) h) is meaningless - what is a 'tie-in to Non System Elements'? The subject seems to be traceability / conformance to requirements which can be described. And what does 'entire architecture' mean? Do you mean 'any Architecture Description Element can be linked to one or more Requirements'? Why 'Non-System'? Don't you allow a System element to be linked to Requirements?
6) f) is titled 'Project Deliveries' but then describes 'capability deployment' - projects always deliver tangible things, which may include Systems. The sentence should be modify to describe what the UAF can describe wrt project delivery [how this affects capability is the subject of a))
7) c) Human Factors - 'expected competences and skills' should be 'competences' (the text needs to use the terms in the DMM so cross-comparison is possible)
8 e) h) Standards - 'Requirement' isn't a UAFElement or indeed a DMM element (DMM Figure 9:134) - it doesn't exist within the UAF
9) Built-in Traceability. What does this mean? What DMM elements are automatically linked to what other DMM elements? The UAF Grid doesn't define layers so what is a 'layer'? This could be a liability if the user has no control.

Puzzled why the UML profile for the UAF, the UAFP, is now incorrectly termed the UAF Modelling Language as this introduces yet another unwanted term for a UML Profile that is in use for the SysML and the UML yet somehow confusing or not good enough for the UAF requiring it to invent its own bespoke term for the same concept.

Looking at closed issue UAF-29 - https://issues.omg.org/issues/UAF12-29

The argument then made:
'Misunderstanding about whether the UAF Profile can be used in modeling an enterprise architecture. Not uncommon for managers to think that they must use SysML to model their EA since they don't realize that the UAFP is already designed with the semantics for modeling enterprise constructs such as capability, enterprise phase, processes, personnel, operations, services, portfolios, etc. This misunderstanding is largely due to fact that UAFP is called a "profile" and many don't understand what is meant by profile. '

1) The UAFP is as is explained at length in the EA User Guide - an implementatoion of SysML so, yes, SysML is used for the UAF views. Are you stating that the UAF views can be produced in a UML modelling tool without the UAFP and without the SysML (non SysML users Don't Care About the this as it is irrelevant)
2) The UAFP is a UML profile. If it isn't please explian why - there does seem to be an XML file that is a UML profile
3) No users should be concerned with UML profiles - if they are it is because the OMG unnecesarily include this in user-facing documentation. The solution is to remove references to it from the EA User Guide

The disposition on closing then states:

'The decision has been taken by the group to rename the profile part of the specification to the modelling language following the naming convention of OMG, e.g. UML, SysML, SoaML, RAAML, etc. The change will improve clarity of the purpose of the document as the term "Profile" is not so well understood in non UML modellers community. Plus it is more than just profile. It also brings notation. '

4) The UAFML is not distinct from the SysML - is every model using the SYSML now its own modelling language? Of course not.
5) '"Profile" is not so well understood in non UML modellers ' - why does any non-UML modeller care about a profile - it's completely irrelevant because it's an implementation mechanism for a UML modelling tool. Non-UML modellers use the DMM which is supposed to be - but isn't quite yet - UML free / agnostic. This is invalid.
6) The SysML and the UML are UML profiles. They also have 'notation' - whatever 'notation' means since this doesn't form part of any specification of the UAF DMM.
7) The DMM correctly identifies the UAFML as a profile : 'The Unified Architecture Framework Modeling Language (UAFML) is the standard implementation of the UAF DMM. It was created by mapping the UAF concepts and relationships to corresponding stereotypes in the UAFML Profile.'
8) The user cannot use the UAFML to create views - the XML profile does not content elements to represent the UAF architecture views nor define what is allowed in each `UAF architecture view - this is done in the UML modelling tool. The users cannot therefore use the UAFML without this extra hidden 'magic' (technical debt). They can, however, open a UML modelling tool and use the tool to create a UAF view of a particular flavour (with no mention or reference to the UAFML in sight).

It is this casual creation of new terms to refer to existing terms that is one of the root causes of inconsistency within the UAF. The point of standardisation is sticking to the terms not constantly rolling your own.

1.1 Overview of the Unified Architecture Framework p2 states:-

'The UAF Grid (Figure 1:2) has rows that represent typical stakeholder domains (or viewpoints as they are called in UAF)'

Figure 1:2 shows a left hand column which the callout labels as 'Viewpoints'. The cells identifiy 'View Specifications'

This is clearly wrong and inconsistent with ISO/IEC/IEEE 42010. It looks as though the authors have not appreciated that 'viewpoint' for the leftmost column derives from the MODAF use of the term to collect together a group of architecture view definitions with a common theme. In the old ISO/IEC/IEEE 42010:2011 text this was referred to as an (architecture) perspective. In misusing 'viewpoint' this then left a problem of how to refer to the things that specify architecture view content (ISO42010:Architecture Viewpoint) and therefore another undefined term was added - 'view specification' which seems to mean the same thing as ISO42010:Architecture Viewpoint.

The use of 'view specification' is further muddied by the DMM Figure 7-1 p13 which states that 'view specifications (cells) correspond to viewpoints'. This could be read a view specifications is a synonym for viewpoint.

It's a mess. The easiest solution is to consistently use the ISO 42010 definitions and the names of the ISO 42010 concepts without alteration

1) Eliminate any mention of 'view specification' where this refers to an artefact that specifies architecture view content. The correct term is 'Architecture Viewpoint' not 'Viewpoint' (as this is otherwise easily confused with other uses of the term).
2) The use of Viewpoint in reference to the left hand column of the UAF Grid is incorrect - that is an inconsistent and incorrect and doesn't (as claimed) conform to ISO/IEC/IEEE 42010. The suggestion is 'Architecture Perspective'. 'Domain' often refers to other things. Whatever term is used it needs to be prefixed with 'Architecture' to separate it from the casual use of the term.
3) Correct the names of the metamodel elements in the DMM. If you can't or won't, produce a mapping table that demonstrates the mapping and the equivalence to the standard.

4) A viewpoint is not a 'stakeholder domain'. This seems to be using 'viewpoint' in another sense. It is also not a synonm for 'domain'. The cells identifiy 'architecture viewpoints' - specifications against which architecture views are created and interpreted.

A.4.3 Architecture Description Organization and Relationships p 119. 120

The problem in not using the terms defined by ISO 42010 is that inconsistency results:

'As the enterprise architecture description is built in those steps, it....'
'Some architectural descriptions may represent generalizations or elements intended to convey guidance and patterns for reuse'

'Architectural Description – a work product used to express the Architecture of some System Of Interest.It provides executive-level summary information about the architecture description in a consistent form to allow quick reference and comparison between architecture descriptions – It includes assumptions, constraints, and limitations that may affect high-level decisions relating to an architecture-based work program. (Note: architectural description is a UAF model element and “architecture description” is simultaneously a concept used in ISO 42010 that is defined as a collection of architecture views)'

The Note is meaningless - what does ' simultaneously a concept used in ISO 42010 that is defined as a collection of architecture views' and how somehow does 'architectural description' differ?

The definition deceives the reader into thinking that this is ISO 42010 definition. The first part has indeed been taken from ISO 42010 - reference needed and then has been added to starting with 'It includes...' which is not part of the ISO 42010 definition of 'architecture description'

and in the DMM 9.1.2 p 130 we have 'Architectural Description' defined as 'An Architecture Description is a work product used to express the Architecture of some System Of Interest.'

If it's an (ISO 42010) Architecture Description use the term Architecture Description.

'to federate multiple architectural descriptions in a structured manner or as a set of associations or usages within an organization’s architecture description repository ...' - 'architecture description repository' (correct) storing 'architectural descriptions' (incorrect)

'so that architecture description elements may be directly translated into model ...' - 'architecture description element' (correct) - not 'architectural description elements'. Interstingly the UAF DMM has no such thing as an Architecture Description Element.

Later on we have:

View – an “information item..' - this is incorrect the ISO 42010 term is 'architecture view' since 'view' has many other meanings.

In 3.3 Establishing the Purpose and Scope of the Architecting Effort we have :-

'The enterprise architecture description is used by stakeholders to improve communication and cooperation among affected parties and enable them to work together in a more integrated, coherent fashion. This will, in turn, help the enterprise more effectively achieve its goals. This can be facilitated by creating a “reference architecture” that guides development of the rest of the enterprise architecture in Steps 3-7, as well as using an architecture framework that defines the views to be used.'

The subject is an enterprise architecture descriptiuon.

"Reference architecture" is incorrect - it should be "reference architecture description" that guides development of ... the 'enterprise architecture description' (not enterprise architecture)

1.6 Cross Cutting Viewpoints and Figure 1:5 - Cross-Cutting Viewpoints in UAF

These are not 'Viewpoints' in the sense of either ISO/IEC/IEEE 42010 or even the UAF's own DMM - 9.1.2 Summary & Overview - Viewpoint p134 - Viewpoint.

This inconsistency needs to be removed. Terms should not be incosistent with international standards. Any term used in the text should also be consistent with the underlying metamodel.

4.1.3 Architectural Description Structure p26 states:

'Each architectural description is either decomposed into, or synthesizes, views and viewpoints, which may be laid out in dashboards and other fit-for-purpose perspectives to aid in understanding an architectural plan. Each planned viewpoint should address specific concerns held by relevant stakeholders'

1) Simply 'decomposes into or synthesizes' to 'formed from' - the aim should be to have as short a document and as short as sentences as possible.
2) 'decomposes into ... viewpoints' is incorrect for the UAF. Users are not going to include the UAF viewpoints in each of their architecture descriptions. Even in ISO/IEC/IEEE 42010 where it got less clear in the 2022 edition owing to removal of multiplicities the inclusion of architecture viewpoints is a 'may' not 'shall'
3) What's a 'planned viewpoint' ? Architecture Viewpoints don't address concerns - iaw ISO/IEC/IEEE 42010 they frame concerns. It is not 'should' - Architecture Viewpoints always frame at least one Concern - mandatory i.e. don't misquote the standard. In following the guide the user should reasonably expect to conform to the standard. This process step seems to be the inverse of what the standard says - you first establish what the architecture description task stakeholders concerns are, then you select the architecture viewpoints that most closely frame those concerns.

Later on we have:

'Concern – interest in a Strategic Phase (Strategic Phase is synonym for System in ISO 42010) relevant to one or more of its stakeholders. (Note: a concern may be a “matter of relevance or importance to a stakeholder regarding an entity of interest” [ISO 42010] that will be addressed in an architecture)'

4) A Concern is not an interest in a Strategic Phase. In ISO/IEC/IEEE 42010 it's 'matter of relevance or importance to a stakeholder'. In the DMM p131 it's 'A matter of relevance or importance to a stakeholder regarding an entity of interest.' which is more restrictive that the standard - you can only raise a concern about the entity of interest. It's definitely NOT 'interest in a Strategic Phase'. Suggest that you adopt the ISO definition consistently and without additions.
5) Strategic Phase is NOT a synonym for System in ISO 42010 - a duration or time period is not equivalent to a System
6) a Concern is addressed in an 'architecture description' NOT 'architecture' via the content of the architecture views produced.

4.1 Step 1 Define Architecture Drivers and Challenges p22

1) The subject isn't 'architecture' - it's 'architecture description' - reference materials, the AD effort. The name should therefore be 'Define Architecture Drivers and Challenges'
2) It incorrectly states 'The Summary and Overview architecture view defines the overall architecture'. It doesn't. It defines the architecture description. If you look at p25 in the DMM - View Specifications::Summary & Overview::Summary & Overview - states 'Concerns: quick overview of an architecture description and summary of analysis.' i.e. 'architecture description' not 'architecture'

p23 we have a definition of 'architecture':
'Architecture – fundamental concepts and properties related to an entity in its environment and governing principles for the realization and evolution of this entity and its related life cycle processes [ISO/IEC/IEEE 42020:2019] (Note: This architecture entity can be an enterprise or system or some other kind of thing. These fundamental concepts and properties can be about key entities and relationships, along with associated behaviors, which are characterized in an architectural description.)'

3) The source for the definition should be ISO/IEC/IEEE 42010 not ISO/IEC/IEEE 42020
4) There is an additional note which isn't part of the original source - this incorrectly states 'This architecture entity can be an enterprise or system or some other kind of thing'. This cannot be true. In ISO/IEC/IEEE 42010:2022 we have in the conceptual model 'Entity of Interest has Architecture' therefore Architecture cannot be the same as Entity of Interest or System.

In the DMM 9.1.2 Domain MetaModel::Summary & Overview p 131 we have a definition for the Architecture metamodel element:

'An abstract type that represents a generic architecture. Subtypes are OperationalArchitecture, Service Architecture, and ResourceArchitecture.'

5) Ignoring the fact that the DMM doesn't define the Architecture concept - it only defines its use in representing a real world thing (as all AD elements do) - we have 2 different definitions of Architecture. The DMM one is incorrect. The suggestion is to use the ISO/IEC/IEEE 42010 one unaltered - the UAF incorrectly claims that it conforms to ISO/IEC/IEEE 42010 so why wouldn't it use the definitions and relationships from the standard?

The metamodel element Definition is defined 'A comment containing a description of an element in the architecture.' and Figure 9:119 shows that there is some unnamed relationship between UAFElement and Definition (impossible to extract any meaning from this).

1) The definition of Definition is wrong - 'a comment containing a description' - is this Comment, Description or, as the element name says,a definition?
2) Every UAFElement presumably has a 'description' - if Definition is an attribute of every UAFElement why not show it as such on the UAFElement (as has been done with the 'author' element of Definition - inconsistent representation of attributes). If this isn't an attribute there needs to be something in the text of the document to explain how to interpret. All relationships should be named - any unnamed relationship, excepting specialisation, is an error as it leaves the interpretation to guesswork . Role names are a bonus.
3) The 'author' element name needs to be qualified to 'definition author' or 'comment author' to separate it from the DCMI creator tag (= author). It isn't clear in the UAF whether there is an 'element author' - the person who created the element. It would make more sense for the element creator to be captured since they presumably define the element on creation - simply capturing the definition author and not the element author makes little sense.

Figure 1:4 - Architecture Views and Models are Used in Other Architecture Processes shows a box labelled 'Architecture Repository'.

This is incorrect. Using the ISO 42010 conceptual model the (real world) 'architecture' and 'architecture description' (an artefact) are separate concepts. You do not store the real world thing on a hard drive - you store its description.

I would recommend checking and issuing the following as guidance to the UAF authors to avoid keep making this error - [1] ‘The Treachery of Images, 1929 by Rene Magritte’, Rene Magritte - Biography, Paintings, and Quotes. [Online]. Available: https://www.renemagritte.org/the-treachery-of-images.jsp

It might seem trivial but it makes reading hard work particularly where you legilimately use an 'architecture description' to make comment on an 'architecture'. Because this error is so widesperead it's impossible for the reader to understand Figure 1.4 because anywhere 'architecture' appears it could really mean 'architecture description'

If someone told you that they were storing a duck on a hard drive you'd question their sanity. If they instead correctly said they were storing a photo of a duck on a hard drive it'd be a perfectly understandable and reasonable thing to say. This persisant and endemic confusion of the two terms is that wrong.

The abstract states:
'This document describes a workflow for creating Enterprise Architecture (EA) views in accordance with the Unified Architecture Framework (UAF) Modeling Language (UAFML).'

The traceability is incorrect. No process conforms to a UML profile - this makes no sense.

A UML profile, the UAFML, is not the root / master / defining specification for the UAF - it is the agnostic UAF DMM specification. Unless you are stating that the EA guide is only applicable to one specific implementation and not any implementation of the UAF. If this is true the statement should be qualified to state that it only applies to the particular UML + SysML implementation provided by the UML profile (UAFML).

Presumably since SysML has its own normative UML profile this profile should be renamed the SysML Modelling Language (SysMLML) ....

7.2 Viewpoint Interrealtionships p 16

The errors in this section are:-
1) title - consistency 'Viewpoint' here is MODAF::Viewpoint NOT ISO42010::Architecture Viewpoint (or indeed the casual meaning of 'viewpoint'). None of this is clear. This is the UAF not the MoDAF - there is no need to use MODAF::Viewpoint and it's confuysing when the text uses 'viewpoint' to mean different things without defining terms. The section doesn't identify relationships between (MODAF) Viewpoints either.

2) '...relationship between the Viewpoints, Aspects and View Specifications,' should be 'Architecture Perspectives, Aspects and Architecture Viewpoints'. There is no such thing as 'view specification' - if you mean the mechanism to specify an architecture view content etc the standard term is 'architecture viewpoint'.

'two-dimensional nature it is not adequate to explain the abstract interrelationships that exist between the viewpoints. The following diagram is an indication of the how the viewpoints are interrelated'

3) Figure 7:2 has no axis into the page - it has just as much '2-dimensional nature' (whatever this means) as the grid.

4) Figure 7:2 neither shows nor explains any relationships between 'MODAF::viewpoints' in the UAF. It describes a set of levels with no explanation or rationale in terms of the order. In any case this ordering seems to be more associated with an order of describing things i.e. process and therefore shouldn't be in the DMM at all. The DMM isn't a process document - it defines metamodel elements, the triples formed and is supposed then to define the which architecture views (via ISO 42010::architecture viewpoints) allow the triples to be shown. Why, for example, is 'Projects' at the bottom? Why is 'Standards' on its own on the right hand side?

5) '...the Viewpoint is a cross cutting concern' - is not true a (MODAF) Viewpoint is not a concern. No (MODAF) viewpoint is a concern. Do you mean than any element in the 'Architecture Management' (MODAF) Viewpoint may appear in any other 'viewpoint'? Probably not. I suspect it means that architecture views within the Architecture Management Perspective may be developed throughout the process of creating an AD. This isn't clear and it's not the correct document.

6) 'the Viewpoint exists in a layer of abstraction between the Viewpoints above and below it' - (MODAF) viewpoints are not abstractions of each other. At best you can say that the metamodel elements in a particular 'viewpoint' represent concepts that are more abstract than ... but it is incorrect to say that a Viewpoint has any abstraction - it hasn't.

The idea that this diagram describes any relationships is simply not true. The easiest solution is to delete section 7.2 - it isn't used anywhere within the DMM. All you need is 'the grid'.

7 UAF Grid states:

'Due to the complexity of managing the multiple viewpoints with overlapping concerns and metamodels, the standard viewpoints are refactored as described in the donor frameworks into a more manageable format. This decision led to the development of the UAF grid which is described below.'

1) This isn't accurate or really correct. All AFs have multiple overlapping viewpoints - that's the basis of architectures description and the basis for ISO/IEC/IEEE 42010. The particular problem that the UAF had - not anyone else, including each donor AF, was a mechanism to classify and compare donor AF view subject area/purpose to support the creation of a common encompassing view set of 'view specifications' (in UAF speak). It has nothing therefore to do with overlapping viewpoints - it's a problmen caused by trying to represent multiple AFs using a single implementation. This is a UAF-management problem not a user- one.

The text then states 'The intent of the grid is not to be complete, but to capture the information that is present in the frameworks that contributes to the UAF, consequently, some gaps are evident.'

2) This isn't accurate either. The UAF defines 'view specifications' that are unique to the UAF and which are not present in any of the donor AFs - the grid therefore shows content that has nothing whatsoever to do with the donor AFs.

3) The grid shows 'Simulation' and 'Parametric Execution/Simulation'. The subject matter of the grid is the set of 'view specifications' The fact that someone may choose to do some form of simulation is irrelevant to the DMM which is supposed to be a specification of the metamodel and view content. It mustn't include process. The grid is supposed to show architecture view coverage - these cells do not correcpond to view specifications and do not belong in the table or document. It misleads the reader into thinking that there are more view specifications than there are.

4) This section keeps referring to donor AFs. It is almost impossible to figure out how they map to UAF 'view specifications'. The reader has to open the traceability document (which uses view specification identifiers not present in the DMM, find the table for the particular AF (or an appendix) and then switch back to the DMM. There should be another grid which simply places the donor AF view identifiers in each cell (and the DMM needs to eliminate 'view specification' and use the UAF identifiers for each definition.

The UAF is supposed to be a technical standard. The effectiveness of standardisation requires the use of the correct standard terms consistently. ISO/IEC/EEEE 42010 provides the correct terms needed for architecture description - both their names and the meanings - but the UAF mixes and matches terms, changes their names and intoroduces its own non-standard new ones. The use of 'viewpoint' is inconsistent and incorrect and there is nothing provided to the reader so that they can understand the different meanings of the same term. This wouldn't be a problem if the UAF used the term 'architecture viewpoint' (not 'viewpoint') as defined by ISO/IEC/IEEE 42010 for the meaning defined by the ISO.

The abstract states:
'The nine steps of the workflow are laid out in alignment with the stakeholder viewpoints in UAF for producing the requisite architecture views in each of those viewpoints'

1) 'stakeholder viewpoints' should be 'architecture viewpoints' (viewpoints and views do not record something when "viewed from a particular direction"). The problem here, though is that these 'stakeholder viewpoints' are incorrectly termed 'view specifications' both in the UAF grid and the DMM package structure.

2) 'architecture views in each of those viewpoints' - an architecture viewpoint is a specification for an architecture view not a containing mechanism - this use of 'viewpoint' is actually MODAF::viewpoint - a different thing and undefined

3) wherever 'view' appears it should correctly be 'architecture view' in accordance with ISO/IEC/IEEE 42010 so that that the meaning is uambiguously tied to the ISO concept not the various casual uses of the term.

It states:
'UAF provides a complete set of stakeholder viewpoints as the basis for defining the variety of necessary architecture views of an Enterprise and these views are specified in the UAFML.'

1) 'stakeholder viewpoints' should be 'architecture viewpoints' in accordance with ISO/IEC/IEEE 42010
2) a UML profile such as the UAFP does not specify architecture views nor view content. There is no standardised mechainsm within the UML to do so. The normative UML profile for the UAF - dtc/21-12-14 - contains no stereotypes to represent any UAF architecture view nor does it contain any definition of UAF architecture view content. [ The statement that 'these views are specified in the UAFML' is flat out untrue ]. Do the UAF authors not understand the architecture of the UML profile for the UAF?

7.3 Domain Metamodel Diagram Legend - Figure 7:3 Legend of color codes for element types defined in UAF

This shows 'External Type' defined as 'An External Type is an element that exists outside of the core DMM but is referenceable by elements in the DMM'

1) Consistency - What is a type? The term 'element' or metamodel element' is used

2) There should be nothing external. The DMM is agnostic. It is also not a UML profile. it is not a SysML profile. The purpose of the DMM is to define an abstract metamodel (complete) and the ISO 42010::architecture viewpoint content using these metamodel elements.

Incorrect section title - '1.4 Layered Progression of Architecture Definition'

As it states: 'The viewpoints allow for a logical and systematic flow of architecting activities'. This has nothing to do with 'architecture definition' - it's the evolution of 'architecture description'

The accurate title is therefore '1.4 Layered Progression of Architecture Description'.

EA guide p 136 - Appendix B - Glossary - Viewpoint

This states that the ISO 42010 term 'Viewpoint is '“conventions for the creation, interpretation and use of an architecture view to frame one or more concerns” [ISO 42010] that governs the creation of views'

It then states that the UAF term 'viewpoint' is 'frames (to formulate or construct in a particular style or language) one or more Concerns. A Concern can be framed by more than one Viewpoint.'

1) the ISO 42010 term is 'Architecture Viewpoint' not 'Viewpoint' - this distinguises it from incorrect casual uses of 'viewpoint'.
2) If anyone sees 'Viewpoint' how will they know whether it's the ISO 42010 'viewpoint', the first meaning of UAF 'viewpoint' defined above or indeed the MODAF viewpoint which is used within the documents bit not defined anywhere?
3) If you modify the ISO 42010:2022 definition by appending anything it isn't then the ISO 42010 definition and shouldn't be labelled as such because it misleads. If you need to add notes use a footnote so that there is clear physical separation of what is a local annotation. Adding a description of relationships etc shouldn't be part of the definition of any concept - it should be atomic - it is an element in the 42010 conceptual model and describing the conceptual model shouldn't be in the table itself. Or create an additional 'Comments' column and move the non standard text there.
4) Until all references to (MODAF) viewpoint are removed the UAF cannot claim and a 'viewpoint' = ISO 42010:: architecture viewpoint because in many places in the text it isn't

Simnilarly for 'View' above - there is not such thing in ISO 42010 as 'View' - the correct term is 'Architecture View' because there are many casual and incorrect meanings of 'View' (as the UAF element is named)

This is a self-inflicted error - if the UAF stuck to the name and definition defined by ISO 42010 it wouldn't arise - there wouldn't be 2 (actually) 3 uses of 'viewpoint' with different meanings. That's the point of standardisation - only 1 term, 'architecture viewpoint', used throughout with the one consistent meaning.

2.2 Core Principles - Compliance Levels p 5 states:

'Compliance levels: UAFML has a single compliance level based upon a combination of the reuse of UML and SysML elements. It is expected that the views that are created as result of this profile have frames that reflect the underlying SysML diagram type that is used as the basis for the view. It is also expected that the graphical notation used to display elements within those views correspond to the standard SysML graphical notation of the SysML/UML metaclass that the stereotype extends.'

1) What does 'the UAFML has a single compliance level' actually mean?

2) This makes no sense. The UAFML must conform to the DMM - the DMM defines the concepts, their meaning etc with which the UML + SysML implementation in this profile is required to conform. The DMM essentially defines the set of input requirements which any implementation, including the OMG's own UML + SysML profile, must meet

3) Where is the traceability mapping to the DMM? This is missing and is required as part of the evidence of conformance to the DMM.

4) This conformance is local to this UML profile. Given that there is a normative profile file what does conformance mean? The OMG define both so either the UML profile that the OMG also produces either conforms or it doesn't . If the OMG are stating that in order to conform some extra requirements have to be met which are not defined by the UML profile these requirements must be identified to enable a fully-traceable conformance assessment i.e. to UML profile + to other atomic requirements. At the moment this is incomplete - where is the evidence that the OMG with their UML profile for the UAF + 'something else' conform to the DMM?

5) There is no reference to and this might be inconsistent with the conformance types stated in the UAF DMM section 2 (Type 1 to Type 3). The UAF DMM defines conformance levels for an implementation which have nothing to do with this local conformance and vice versa.

EA Guide p17 - 3.1 Architecture Management Concepts (more correctly, 'Architecture Description Management Concepts' as the subject is 'Architecture Description' not 'Architecture'

States:
'illustrated in the conceptual schema shown in Figure 3:1. These key concepts are highlighted in italics within the Narrative and some of the less obvious concepts are listed with the associated ISO-42010 2 meaning or the UAF meaning of that concept'

The DMM 1.3 states - 'Further, The UAF conforms to terms defined in the ISO/IEC/IEEE 42010 standard for architecture description'

1) If any definitions differ in the UAF vs ISO/IEC/IEEE 42010 then the UAF does not conform to the standard and any such claims need to be removed.
2) Any references should be at the beginning of the document as other UAF documents not spread throughout the document on footnotes

p117 of A.4.1 Standards and Practices in the EA guide:-

states - 'ISO/IEC/IEEE 42010:2021 – Software, systems and enterprise – Architecture description: provides core terms, definitions and relationships for architecture descriptions'

1) The correct version is ISO/IEC/IEEE 42010:2022 - see https://www.iso.org/standard/74393.html
2) The DMM refers to the 2011 version under Normative references. Not incorrect but likely inconsistent with any claims of conformance by the UAF

in UAFML Operational::Structure concern is
identifies the operational exchange requirements between OperationalPerformers

where as in UAF DMM Operational::Structure concern is:
identifies the operational exchange requirements between nodes

so need to replace nodes with OperationalPerformers

View Specifications::Architecture Management::Motivation

Am-Mv Architecture Principles 'View Specification' p19, Figure 8:1

States: 'Concerns: alignment of architecture with architecture heuristics, guidelines, and principles.
Definition: identifies relevant architectural principles and other guidelines to be used in architecture development and evaluation.'

1) What does 'alignment of..' 'with' mean? Do you mean identification of the reference sources? Do you mean identification of input requirements? Do you mean conformance with requirements (process, technical etc)? The current is unspecific and will therefore yield inconsistent results. Should be 'alignment of architecture description' ('architecture description' not 'architecture')

2) Fig 8:1 only provides a 'Driver' element and possibly typed by DriverKind (this isn't clear because of the notation. Is this a normative value - if so this needs to be made clear. At the moment it looks like an error of ommission. ''Architecture Principle' should be 'Architecture Description Principle'.

3) Is the notation supposed to indicate that Driver has an attribute 'DriverKind' which has a set of enumerated values? Even in the UML you'd show an attribute on Driver and link this to the enumeration so this notation doesn't look to be complete. It could be interpreted as a form of specialisation of Driver

4) There are no means to describe the 'alignment of architecture (description) with architecture heuristics, guidelines, and principles' - no relationships provided, no target elements to link Driver to etc. As shown a compliant architecture view wouldn't even constitute architecture desctiption since it'd be pemissible to provide a single Driver element not linked to anything - an orphan or soliton - which isn't describing anything.

5) The multiplicity on DriverKind is incorrect - shows 0..1 which allows Driver to take an empty undefined value. This not only fails to address the architecture viewpoint concerns but is indeterminate. Should be '1'.

6) The enumeration shown for DriverKind and on Fig 9:138 has no default value specified.

9.1.1 Domain Metamodel::Architecture Management - Architectural Reference. p128 , Fig 9:125

ArchitecturalReference defined as 'A tuple that specifies that one architectural description refers to another.'

1) The name of the element should be 'ArchitectureDescriptionReference' as it isn't 'architecture' that's being referred to - it's an ArchitectureDescription element
2) ArchitecturalReference isn't a tuple. The shortest possible tuple, a triple, is either node - relationship - node or node - relationship - (same) node if reflexicve. ArchitecturalReference is a relationship.
3) The definition of a metamodel element must define the concept itself. This must be atomic. The 'one architectural description refers to another' defines how it may be connected - it doesn't define what 'ArchitecturalReference' is. If the element were defined you'd probably correctly decided that it isn't a concept for a metamodel because there is no atomic definition other than 'undefined' because it's semantically equivalent to the other "universal stuff" releationship, ArbitraryConnector. This sort of error would be much easier to trap if there was rigour in the definition of individual metamodel elements.
4) A 'tuple' does not specify anything. This connector element is purely a description or assertion that there is some unknown relationship between 2 ArchitecturalDescription elements
5) If a reference is required why isn't a trace relationship used? Then we'd have one less arbitrary metamodel element.

p40, 41 - 8.1.4 View Specifications::Operational::Processes - View Specification Op-Pr Operational Processes BPMN Semantics

The DMM is supposed to be an agnostic specification. This clearly shows the bottom-up derivation of the DMM by chopping bits out of the UML profile.

1) The title is invalid - no architecture viewpoint should be named after a particular ADL implementation - the name ought to be 'Operational Processes' if that reflects the purpose/ concern not the means of realising this.

'Concerns: captures activity-based behavior and flows using BPMN notation.'
2) This is invalid - the ADL or method is never part of a valid concern. The stakeholders presumably only care about the processes not how they are represented.

'Definition: describes the BPMN processes that are normally conducted in the course of achieving business goals that support a capability. It describes operational activities, their Inputs/Outputs, operational activity actions and flows between them using BPMN notation.'

3) This is invalid. It should describe processes. It shouldn't contain any reference to any particular ADL, UML or SysML element names et al

Recommended Implementation: BPMN Process Diagram.

4) It shouldn't make reference to any technology or ADL. In any case why can't processes be described using a UML ACtivity Diagram? What's so special that this has to be BPMN?

5) Figure 8:27 shows many BPMN elements. Delete them - they have no valid place in the DMM

p99 8.1.10 View Specifications::Standards::Traceability

The concern for the 'UAF::view specification' is:-
'Concerns: standards that need to be taken in account to ensure the interoperability of the implementation of architectural elements.'

1) Why is a view needed to describe the standards affecting the interoperability of an architecture description?
2) 'implementation of architectural elements' is meaningless - do mean the standards that affect the system of interest? Standards don't just affect implementation

'Definition: shows the applicability of standards to specific elements in the architecture'

3) Difficult to make sense of given the endemic conflation of 'architecture' to mean both 'architecture' and 'architecture description'. Standards apply to the system of interest. 'elements' is too much like 'architecture description element' used to form an AD.

Figure 8:82 Standards Traceability

This shows Standard, Protocol elements and the generic UAFElement.

4) There are no relationship elements provided to establish what looks like a 'conforms to' relationship between a UAFElement and a Standard or Protocol

5) Either the diagram doesn't provide the elements to address the concerns or the concerns are incorrect. The diagram describes conformance not a trace ('standards that need to be taken in account'). Which is it? The concerns and the triples shown in the diagram need to be consistent with each other.

p38 8.1.4 View Specifications:Operational - Operational Connectivity, Figure 8:25, p165 9.1.4 Domain Metamodel::Operational - OperationalActivityEdge Figure 9:202

The OperationalActivityEdge element is defined as :-

'A tuple that shows the flow of Resources (objects/information) between OperationalActivityActions'

Figure 8:25 shows that this element realises the OperationalExchange element (indicated by the realizedByActivityEdge role name on OperationalActivityEdge and the realizes role name on OperationalExchange).

An ActivityEdge is a particular UML method to provider a connector to join two Activity node elements on a UML Activity Diagram. This is a particular UML implementation. It should be in the UML profile for the UAF but not in an agnostic metamodel or architecture viewpoints in the DMM. How someone chooses to implement an OperationalExchange is up to them - it isn't the subject of the DMM.

p158 9.1.4 Domain Metamodel::Operational - Operational Architecture, Figure 9:187

OperationalArchitecture is defined as:
'A type used to denote a model of the Architecture, described from the Operational perspective.'

1) We incorrectly have a model of the architecture being a type of architecture (OperationalArchitecture specialises Architecture).

2) The description of the 'operational architecture' is the set of operational node elements and relationships (forming triples), perhaps augmented by some standards - it is incorrect then to have a concrete single element representing the same thing.

It isn't really any different as a type from Architecture - it seems to be purely a mechanism to show relationships with different operational elements. It isn't a distinct concept in what is supposed to be an abstract metamodel.

So what then is the relationship between 'OperationalArchitecture' and the description of an operational architecture (if the element really is 'architecture' and not 'architecture description'? It ought to conform to ISO 42010 ie something like 'expresses' or in the reverse sense 'is described by'

p71 8.1.7 View Specification - Rs-Tx Resources Taxonomy, Fig 8:55

states
'Concerns: resource types.
Definition: shows the taxonomy of types of resources.'

1) Architecture viewpoint has no identifier (a package containment isn't a unique identifier) - should be Rs-Tx according to the Traceability document.

2) Figure 8:55 shows ResourceExchange - this is a connector. It isn't used in the architecture viewpoint - a taxonomy uses 'is a' (specialises / generalises). This shouldn't be in the diagram

3) The Measurement element isn't a Resource and also isn't involved with taxonomy - it doesn't address the concerns of the architecture viewpoint and should be removed.

4) The realisation of an OperationalPerformer by a ResourcePerformer doesn't involve taxonomy and doesn't address any of the concerns. This should be removed.

Fig 7.1 in the DMM and the UAF Grid in Appendix C are different...

for example... one has Am-St and the other one does not... there are other differences

Ar-Tr is missing on both Graphics of the UAF Grid

What is called a Method in UAF is really an Operation... this is confusing as the semantics for Method and Operations are set in
SysML via UML

Missing metamodel for StrategicInformation

Sm is not defined in Table 7.1

p245 9.1.12 Domain Metamodel::Parameters - PropertySet & Fig 9:363

PropertySet is defined as 'An abstract type grouping architectural elements that can own Measurements.'

1) Incorrect. The definition is not atomic - refers to 'owns' which implies a relationship that isn't present with PropertySet in the metamodel. 'Architectural Element' is actually an AD element (cf ArchitecturalReference). Things don't own measurements - they have characteristics that are quantified by measurements. The definition should only define the concept not the use or application of the element - this would then highlight the problem wrt what the concept actually represents. It appears to be an implementation artefact that has nothing to do with the fundamental metamodel concepts.

2) Figure 9:363 shows that many metamodel elements specialise PropertySet, most of which are semantically incorrect. We have, for example, 'System is a PropertySet', 'Viewpoint is a PropertySet', 'StrategicPhase is a PropertySet', 'OperationalInterface is a PropertySet'. The full class hierarchy needs to be shown including all types of Asset, Resource and MotiovationalElement otherwise the sense (or error) of what is being asserted cannot be verified. The only assertion that is probably correct is 'MeasurementSet is a PropertySet'

3) There is a reflexive PropertySetGeneralization on PropertySet - presumably to assert a specialisation of successive PropertySets. However, because it is on PropertySet - the parent - and because the child elements inherit this relationship with any PropertySet element it therefore asserts that 'System PropertySetGeneralization Viewpoint', 'EnterpriseGoal PropertySetGeneralization Service' i.e. the child PropertySet elements can be connected together using PropertySetGeneralization which is semantically meaningless.

4) There are several elements - possibly many - where a triple including PropertySetGeneralization can never appear in any architecture view (doesn't appear in the definition of allowed content of any 'view specification'). For example the View, Viewpoint elements. You might expect these particular triples to appear in the Summary & Overview 'view specification' if valid - but there is nothing shown in 8.1.2 and doesn't address the concerns. If triples do not address concerns and do not appear in any 'view specification' they should be deleted from the metamodel - they serve no defined purpose.

5) There is no defined relationship between Measurement and PropertySet that can be used in any 'view specification' - this would require a relationship element to be shown and there is only the reflexive PropertySetGeneralization

For Protocol it extends Class which is redundant and probably unhelpful...

since Protocol is a specialization of Block (which extends Class) this is not helpful, because one needs to know that this stereotype
would be used instead of <<Block>> (since one can not have both a generalization and specialization of a stereotype on the same element)...

Block here is the more fundamental than the metaclass Class... but neither should be specified redundantly in the model...
it is just not helpful... what would be helpful is defining the fundamental thing in the definition of the UAF elements

CompetenceToConduct in the metamodel is in traceability
and in the profile is in Process

should they not be in the same place?

CompetenceToConduct in the metamodel is in traceability
and in the profile is in Process

should they not be in the same place?

used "Function" in some places and "Activity" in others... e.g. OperationActivity and Function in Resources

Domain MetaModel::Architecture Management::Sequences::InteractionMessage has no metamodel

No Description for CapabilityRoleDependency

no metamodel for StrategicExchangeItem

Even thought the UML Standard does not explicitly state this, extends are inherited by specializations of Stereotypes...

1. The UML standard show examples of specializations of stereotypes that do not extend metaclasses indicating that extends are inherited see Fig. 12.21
2. ExtendEnds are forms of Association Ends which are inherited normally (and there is nothing to the contrary in the standard that they are treated any differently)
3. SysML also uses this... e.g. see Fig. 8.8 of SysML 1.7 where BoundReference does not show an extends (but gets it from EndPathMultiplicity)...

so given UAF... I think most everything that extends Element is wrong... because for example, you do not want to be able to put the stereotype
<<OperationalActivity>> on anything but a UML/SysML Activity... but Activity is MeasureableElement which is a UAFElement which extends Element, so therefore <<OperationalActivity>> can be put technically on anything (e.g. a Comment)... so this is wrong and needs to be fixed

p74 8.1.7 View Specifications::Resources::Structure

1) identifier - Rs-Sr, missing'

Definition: defines the physical resources, e.g. capability configuration(s)/system(s) and interactions necessary to implement a specific set of OperationalPerformer(s).'

2) CapabilityConfiguration and System are not physical resources (as Figure 8:56 shows)

3) ResourceArchitecture specialised elements - System, SecurityEnclave allowed but missing from Fig 8:56. All elements of a taxonomy tree should be shown. If not then the parent should not be shown and particular individual child elements shown.

'Concerns: reference the resource structure, connectors and interfaces in a specific context.'
4) Meaningless - what does 'reference the ...' mean? 'In a specific context'? Back inAbitraryConnector territroy. How can you possibly determine that these elements represent the allowed view content when there is no clear set of concerns which triples formed from the elements should address?

5) Is nothing carried by a ResourceExchange? If so elements are needed on Fig 8:56 to describe this.

6) Protocol is not involved in any triple. There appears to be a triple involving ProtocolImplementation but not Protocol. If there is a triple involving Protocol expected then additional elements are required. If there is no triple involving Protocol delete it.

7) Similarly there is no triple involving Measurement - again define one including a relationship element or remove.

8) The 'ResourcePerformer IsCapableToPerform Function' triple has no connection to the structural and interface concerns. It makes no sense to try and describe functional, interfaces and structure on the one architecture view. Function should be the subject of e behaviour-oriented 'view specification' e.g. Resources Process - Rs-Pr. The IsCapableToPerform and Function elements should be remoed from Fig 8:56

p72 8.1.7 View Specifications::Resources::Structure Rs-Sr Fig 8:56

'Concerns: reference the resource structure, connectors and interfaces in a specific context.'

1) Fig 8:56 requires 1 or more Measurement for each ResourcePerformer i.e. have to show Measurement(s)?
2) No triples provided to describe applicable measurements
3) No concerns require Meaurements to be shown - so the elements shown are surplus to requirements for the architecture view
4) The semantics of Asset is a PropertySet are incorrect. An Asset property or characteristic might be quantified by one or more Measurements. This does not mean, however, that a set of measurements is an asset - it's a meaningless and incorrect assertion. How can the UAF claim any semantic rigour?

Decisions within a process often conform to standards and require measurements to determine the quality of the decision. However, because the DecisionNode element is inherited without modification from SysML/UML, the <<ConformsTo>> relationship and <<MeasurableElement>> are not available on decision nodes. Adding a UAF DecisionNode as a MeasurableElement would improve traceability in the model.

So UAF should remind people using Exchanges that in SysML the standard says...

"A Connector or an Association, or an inherited Association shall exist between the source and the target of the InformationFlow"

not that I like this or think it is correct... but for an ItemFlow this is the restriction given in the standard...

exchanges across ActivityEdges and Messages need to have this...

all the *ActivityAction's extend CallBehaviorAction... but there are other Actions one might want to have in the Activity ... e.g. we have people that just want an OpaqueAction...
so would like OperationalActivityAction, ProjectActivityAction, FunctionAction, SecurityProcessAction, ServiceFunctionAction should all extend Action (not CallBehaviorAction)

since PropertySet types things... one can be tighter about what it extends... it should extend Classifier rather than Element

notice on fig 3.107 ServiceParameter relates to ServiceExchangeItem

but on serviceEchangeItem fig 3.112 it is missing the ServiceParameter

should Service relate to ServiceExchangeItem
similar to how OperationalActivity
relates to OperationalExchangeItem?

The description in this section is a copy of View Specifications::Resources::Traceability and View Specifications::Personnel::Traceability. The desciption os also not in accordance with the example problem in section 14.3 of the Sample Problem (dtc/2021-12-12). Also, Ar-Tr is missing from the UAF Grid. In summary, it is difficult to understand the view specification Actual Resources::Traceability. However, we do understand the need for a view specification according to Figure 4:82.

The Standard Operational Activity definition is not clear ("standard operating procedure" is fairly ambiguous). Would be important to clarify where it should be used instead of an Operational Activity. It is also listed as a Doctrine of the Capability Configuration, which should be checked against the doctrine elements being added to UAF 1.3 to support mission engineering.

The word ‘realizingArchitecture’ between the [Abstract]Architecture and [Tuple]Implements is misspelled in Figure 8:11 - Summary & Overview. The letter d is redundant.
In addition, thanks to the OMG organization for developing the documentation about UAF, which made me learn a lot.

Figure 9:321 – ActualProjectMilestone
Constraints
[1] unnamed1
startTime=endTime

non-named constraint

also ActualState has attributes startDate and endDate... not startTime and endTime

need to have a separate MilestoneSequence from MilestoneDependency

and should add a ProjectDepdency as well

some places are Project some place are Projects

e.g.
View Specifications::Projects::Roadmap::Project Roadmap

ActualService in Metamodel is abstract... it is not in profile... need to fix

endDate is multiplicity 0

VersionSuccession does not tell which is the successor in the documentation does not explain which is the successor to whom... the client or the supplier... I am presuming that the client is dependent on the supplier so the client is the newer version and the supplier is the older version... needs to be explicit in the standard

ComparesTo is ambiguous ... which is the desired, which is the achieved... presuming desired is the supplier and achieved is the client but should be made explicit in the standard

Forcast needs to specify explicitly in the documentation that the supplier end is the older subject and the client end is the newer subject

Forcast needs to specify explicitly in the documentation that the supplier end is the older subject and the client end is the newer subject

architectureFramework as a String should be in the Attributes... and shown as such in the diagram and the text below needs to be altered

UAF is missing a CourseOfAction element... would suggest looking at ArchiMate's version of CourseOfAction as an example

ProjectSequence is ambiguous ... needs to specify before and after with supplier and client

Figure 9:336 – ActualService is italics yet it is not abstract... need to change

Viewpoints are defined as a convention for creation, interpretation, and use of an architecture to frame one or more stakeholder concerns.
Views are defined as the information item governed by the Viewpoint that communicates some aspect of an architecture.
Stakeholder Concerns relate to a Viewpoint, and to a View only through its "governed by" relationship to the Viewpoint.
Views may only be governed by one Viewpoint, but Viewpoints can govern many Views.

The problem is that the traceability for a Stakeholder Concern only comes through the Viewpoint (not the View itself) and thus every View that is governed by the Viewpoint is said to relate to every Stakeholder Concern framed by the Viewpoint. This may not be true.

For instance, I will use Resource Connectivity diagrams as an example for a Viewpoint. I have multiple stakeholders that may want to see a Resources Connectivity diagram. One might want to understand the internal connectors of the system. One might want to understand the flow properties of the system. One might want to understand the external connections to the system. If I make three views (one that shows internal connections, one that shows flow properties, and one that shows external connections), then all three Views are governed by the Resource Connectivity Viewpoint, and every stakeholder Concern can be addressed by Resource Connectivity, but not every View does address every stakeholder Concern. Further, I can't use the model relationships to filter my Views by Stakeholder Concern because the traceability provides every view that is governed by the Viewpoint.

I can solve this by further specifying my Viewpoints. I can specialize Resource Connectivity Viewpoint and require external connections, or internal connections, as related to the Stakeholder Concern. This requires a very high degree of specificity in my viewpoints to ensure that a query across one Stakeholder Concern does not return Views intended for other Stakeholder Concerns.

It may be better to trace Stakeholder Concerns to the Views themselves (having an "answers" or some other relationship) to allow for limitation on this query.

in the standard there is no statement for Sequence about whether the client is assumed to be before or after the supplier... one could
reason it either way... clients should always be dependent on suppliers... and because suppliers should have to finish before
clients can proceed it is assumed that the suppliers are the before and the clients are the after... but this should be made explicit

Many Standards prescribe particular processes (such as ISO 15288, DO-356A, etc). Having a Standards Activity with Standards Process Flows would allow us to model the Standards as processes with inputs and outputs, then show how Operational processes ConformTo the Standards Process.

Op-If is mapped with DIV-1, and Rs-If is mapped with DIV-2.

According to Appendix A the correct mappings should be: Op-If <-> DIV-2, Rs-If <-> DIV-3

MilestoneDependency should be renamed MilestoneSequence to
be consistent with the definition and consistent with ProjectSequence

Dependency is too overloaded

Fig 3.38 is missing MotivatedBy

"Enterprise Visions" & "Enterprise Goals" are supposed to be valid for the whole life of the enterprise whereas "Enterprise Objectives" are time time-targeted.

Proposal:
1. Remove "Enterprise Vision" as a sub-type of Phaseable Element. Enterprise Vision is "owned" by a "WholeLifeEnterprise" as a kind of "DesiredResult" (Ends versus Means).

2. Potentially remove "Enterprise Goal" as sub-type of Phaseable Element. "Enterprise Goal" is "owned" by a "WholeLifeEnterprise" as a kind of "DesiredResult" .
This was the original intent in MODAF (see quote below from MODAF Handbook")
"Within the StV-1 view, certain Enterprise Goals are identified. These goals relate to the entire enterprise lifetime, they are not specific to an
Enterprise Phase"

3. Add "Enterprise Objective" to enterprise phases. "Enterprise Objective" is owned by enterprise phases as a kind of "DesiredResult" .

PS: The current "Phases" relationship mixes relationhip to "Ends" (Vision, Goals, Objectives) and relatonship to Means (Capabilities).
The best practice is to separate Ends and Means.

The names of Ends on MotivatedBy to not match the EA Guide and the principles establishes by the BMM on which it is grounded.
The EA Guides says that "Enterprise Goals" are "motivated by" Drvers.

Proposal:
Rename "motivatingEnterpriseGoal" in "MotivatedEnterpriseGoal"
Rename "motivatedDriver" in "motivatingDriver".

It is not clear why "Opportunity" and "Challenges" are part of "Motivatedby" as it seems to be a separate topic.

Multiple conflicting [1.1] multiplicities on Implements

Proposed solution:
(1) "Implements" is made abstract
(2) "Implements" relates UAFElement (or better "Class of UAFElement").
(3) Appropriate concrete sub-types are created for Resource, Function, OperationalAgent, etc..

The concept of "Concept Role" is missing in the DMN while it is available in the UML profile specification.
As a consequence, ConceptItems (Location and Asset) are owned by HighlevelOperationalConcept which should not be the case.

The Achiever abstraction bundles in the same hierarchy "production agents" (actual resources) and "transformation agents" (actual strategic phases, actual projects).
While both produce effects: the former (actual resources) are about outcomes delivered by the enterprise, while the latter are about change of the enterprise itself.
Shouldn't we distinguish these two kinds of "changes".
In page 39, indeed, the EA guide states that measurements of effect are different for 'transformation effect" versus "ways and means effects":
"A Measure of Effect (MOE) parameter describes qualitative or quantitative states or levels that directly relate to outcomes. Outcomes are sought by stakeholders, whose interests and perspectives are addressed by the actual measurements of effects. These outcome measurements are different from measurements of associated ways and means."

Proposed resolution:
1. Rename "Achiever" in initiative
2. Remove "Actual Resource" as a sub-type of Achiever

Desirer bounds "Class Entities" (Capability, Resource Performer, ..) to individual state (Actual State).
This breaks the BORO principle that classes should not be time bounded.
There is no way to relate Capabilities to Effect (class).

Proposed resolution:
1. Desirer is made deprecated and kept for compatibility
2. A new relationship is added between capability and effect