Journal Home >> DoDAF V2.02 FAQs
- Is DoDAF V2.0 mandatory?
- Do I have to create all those DoDAF-described Models?
- What must architecture tools do to comply with DoDAF V2.0?
- Where can I see exemplars of each DoDAF-described model?
- Is DoDAF useful outside DoD?
- When developing the viewpoints, do we have to list or show processes controlled by software, such as "receive crew data", "transmit crew data", "render crew data" or "process crew data", "provide transmit data"?
- How is server/workstation processing indicated in the architecture?
- What does DoDAF 2 consider an "External Performer" and how does the DM2 handle it?
- The DM2 appears very abstract. Is it just for guidance?
- The mathematics of the DM2 are difficult to learn. Is this really necessary?
- What is IDEAS?
- Is there a scientific basis for the DM2 and IDEAS?
- How do DM2 and IDEAS relate to OWL?
- Are there tools for DM2?
- IDEAS applies mathematics that are not normally taught in IT curriculums and so some learning is required. How do you learn IDEAS?
- Why was UML used for the DM2 Logical Data Model instead of an ontology tool?
- Why did this mathematics suddenly emerge as applicable?
- Why is AS&I spearheading the introduction of ontologic mathematics in DM2?
- What’s the difference between ontology and taxonomy?
- I hear the word “ontology” used a lot nowadays. Is it just another IT fad?
- Formal methods in computer science have been around for quite a while. They usually were too intractable and inaccessible. Why are we adopting them now?
- Some of the IDEAS and DM2 mathematics seem to be esoteric – addressing issues below the 90% or good enough level. Is this degree of precision really necessary?
- Who’s developing DM2 or IDEAS analysis tools?
- Why are there so many DM2 open action items?
- Is the DM2 done?
- How can the DoDAF / DM2 Working Group be effective if anybody is allowed to join and participate? Doesn’t it just become chaotic?
- Why didn’t DoD just adopt a commercial standard for EA data exchange?
- Is there a way to represent metrics with DM2 and what kinds?
- How are temporal models handled in DM2?
- How are Services modeled in DM2?
- How can I implement the DM2?
- I don’t see how to model requirements vs. solutions in DM2. Is it possible?
Is DoDAF V2.0 mandatory?
While DoDAF is indeed prescribed for use in the development of architectural descriptions within the Department, DoDAF V2.0 currently serves as guidance. It is expected, however, that a growing number of commands and components will adopt V2.0. For such organizations, architectural descriptions they may have developed in accordance with prior versions of DoDAF should brought into compliance with V2.0 upon their next major release. In addition, architectural data should be stored in a data system – PowerPoint, Visio, Excel, etc. can only be used to present architectural information. For components within which the use of V2.0 is not mandated, it can still serve as an architecture best practices reference.
Do I have to create all those DoDAF-described Models?
No. DoDAF V2.0 does not prescribe any models – instead, it concentrates on data as the essential ingredients of any architecture development. It seeks to make architectural descriptions “Fit-for-Purpose”, based on decision-maker needs. Process owners may therefore prescribe a specific set of DoDAF-described Models to answer a particular purpose. For example, regulations and instructions issued by both DoD and the Chairman of the Joint Chiefs of Staff (CJCS) contain particular requirements with respect to Presentation Views. In general, whatever combination of views – both DoDAF-described and user-tailored – legitimately answers a need, aligns well with the intended use of the architecture as a whole, AND can be justified per common-sense professional practice in architecture, is acceptable. Consult the regulations and instructions issued by your component for specific model and view requirements.
Go to top of page ↑
What must architecture tools do to comply with DoDAF V2.0?
It is only necessary to implement the Physical Exchange Specification (PES). Note that while architecture teams may evaluate tool sets and recommend specific tools based upon their capabilities in a given architecture environment, DoD does not plan to formally endorse, adopt, certify, or mandate any tool.
Where can I see exemplars of each DoDAF-described model?
See the DoDAF Journal – a compendium of DoDAF V2.0 background information, implementation guidance, news, and other content useful to the DoDAF architect and decision-makers alike.
Is DoDAF useful outside DoD?
Yes! Given the unprecedented, growing, and mutual dependence between DoD, Intelligence Community (IC), and Coalition architectures, we both encourage and expect the early adoption of DoDAF V2.0 principles outside the Department. As a vital partner and contributor to our nation’s defense, the IC should continue to represent a significant portion of the DoDAF user base.
Go to top of page ↑
When developing the viewpoints, do we have to list or show processes controlled by software, such as "receive crew data", "transmit crew data", "render crew data" or "process crew data", "provide transmit data"?
The determination of whether to list the processes controlled by the software is reliant on the purpose of the architecture. If the purpose requires the processes controlled by the software, then it will be required. However, if it is not required at that level, it may be sufficient to indicate that "Provide Target Location" is the resource flow, rather than having multiple resource flows "Transmit Target Location data", "Receive Target Location Data", "Acknowledge Target Location Data", and "Receive Acknowledgement of Target Location Data".
How is server/workstation processing indicated in the architecture?
From a systems or services viewpoint, the server/workstation processes are Activities, but the related Performer is a System or Service (e.g., "CursorOnTarget Service")."
What does DoDAF 2 consider an "External Performer" and how does the DM2 handle it?
In DM2, a Performer can be categorized as internal or external, based on specific need, although such categorization may not be standard across all organizations. External Performers do not need to be modeled, as DM2 does not require documentation of Activities other than acknowledgement that an unknown producing or consuming Activity does, in fact, exist (see UPDM SAR DM2 markup examples). However, although an OV-2 diagram need not show implied, external Activities, the DM2 PES XML must show them, even if only as placeholders for subsequent completion such as during OV-5 development. It is this precision that addresses the “over-specification” problem of earlier DoDAF OVs.
Go to top of page ↑
The DM2 appears very abstract. Is it just for guidance?
No, the DM2 can be used for simple to very detailed and complex architectural descriptions by combining its elements appropriately. It has few elements making it seem abstract because it is not language, but mathematically, based. The DM2 Physical Exchange Specification (PES) is the prescribed data exchange format and semantics for DoDAF 2.0 conformance.
The mathematics of the DM2 are difficult to learn. Is this really necessary?
The predecessor of the DM2, the CADM, was language-based. It was a state-of-the-art Entity-Relationship model at the time. E-R models have been very successful and useful throughout the business and government communities. However, the nature of Enterprise Architecture entails integration and analysis of multiple independently-developed architectural descriptions. The CADM and E-R models that were name and definition based did not work for this purpose. Hence, the DM2 has brought to bear additional science to help achieve these DoD EA goals.
What is IDEAS?
IDEAS is the International Defence Enterprise Architecture Specification. It is an international project of the US , UK , AU, SW, and CA for the past 5 years to develop a way to exchange EA data in support of Coalition operations. Early on in the project it was recognized that we needed more precise and unambiguous ways to label data so the science of formal ontologies was brought to bear. The IDEAS ontology is first-order, extensional, and 4-dimensional, employing the mathematics of set theory and 4-D mereotopology.
Is there a scientific basis for the DM2 and IDEAS?
Yes, the mathematics and science underpinning DM2 and IDEAS have been in development for many years, particularly with the development of set theory in the 19th century. What is new is the application of that science to IT data representations, specifications, and models.
Go to top of page ↑
How do DM2 and IDEAS relate to OWL?
DM2 and IDEAS data can be represented in RDFS and OWL. Pure OWL makes some commitments that are incompatible with IDEAS so a fallback to the less-committal RDFS was necessary in those areas. An RDFS/OWL generator for and early version of IDEAS was developed and will be updated and included in the ModelFutures’ IDEAS Profile soon.
Are there tools for DM2?
A script for generating a DM2 database is available on CD at this conference and on the DM2 collaboration site – www.silverbulletinc.com/dm2. A bare-bones front-end (RDF triples generator) to the database will be available soon. Pilots and early adopters are building tools to generate and parse DM2. The UPDM Team is working on UPDM 2.0 which will be compatible with DM2. The AS&I team is working with other EA and M&S tool vendors to achieve DM2 compatibility. An RDFS/OWL generator is planned that will allow analysis by RDFS/OWL tools that comport with IDEAS ontology constructs.
IDEAS applies mathematics that are not normally taught in IT curriculums and so some learning is required. How do you learn IDEAS?
The DM2 collaboration site at www.silverbulletinc.com/dm2 has many resources, including DM2 description documents (CDM, LDM, PES / IDEAS), an IDEAS bibliography, 1,000’s of reference documents, and a free electronic version of IDEAS Group member Dr. Chris Partridge’s book, “Business Objects – Engineering for Reuse.” In addition, on-site outreach tutorials can be requested through Mike.Wayson@osd.mil, membership in the DM2 Working Group is open to all by registering at www.silverbulletinc.com/dm2. This science is emerging as the future for knowledge representation for applications where integration of multiple heterogeneous data sources or automated algorithmic analysis or processing is required and so IDEAS learning is professional development that will be applicable to and open up future career paths.
Why was UML used for the DM2 Logical Data Model instead of an ontology tool?
A UML tool was used for the DM2 LDM. However, it is not a UML class model because the ModelFutures’ IDEAS Profile turns the UML tool into an ontology tool. Existing ontology tools make commitments and lack features necessary for IDEAS. Consequently, the ModelFutures’ IDEAS Profile was developed that allows the UML tool be used for ontology development in a simple yet thorough way.
Go to top of page ↑
Why did this mathematics suddenly emerge as applicable?
As IT has developed, greater indirection has been permissible. In similar vein to Codd’s introduction of relational databases was enabled by higher performance IT, so too higher performance IT is now enabling the adoption of ontologic mathematics.
Why is AS&I spearheading the introduction of ontologic mathematics in DM2?
Enterprise Architecture is ambitious in supporting transformational processes in DoD. We know those transformations must be accomplished so we will have an agile and efficient defense. That makes it incumbent on AS&I to apply whatever science is needed to support the DoD’s core processes.
What’s the difference between ontology and taxonomy?
Technically, a taxonomy is a “type” structure, much like naïve set theory but with provisions to prevent paradoxes. So a taxonomy may represent categorizations of real world things (e.g., a simple set), subsets and super sets, categorizations of sets, and so on. An ontology goes beyond this an includes other types of relationships between concepts such as whole-part, overlaps, temporal whole-parts, etc.
I hear the word “ontology” used a lot nowadays. Is it just another IT fad?
Yes and no. As in all progressions in IT, there tends to be a bit of overselling. No doubt some ontology projects will fail to live up to expectations. There are many challenges in developing automation, whether for data integration or analysis, as there always have been. However, the newly adopted tools of ontology science – e.g., applying set theory, mereotopology, and 4-dimensionalism – will be long-lasting contributions.
Go to top of page ↑
Formal methods in computer science have been around for quite a while. They usually were too intractable and inaccessible. Why are we adopting them now?
The key to simultaneously achieving user understandability and rigorous formality in the DM2 is the layering: the Conceptual Data Model (CDM), Logical Data Model (LDM), Physical Exchange Specification (PES), and IDEAS Foundation. The formality in DM2 is largely hidden in the IDEAS Foundation layer which most users will never need to look at or understand.
Some of the IDEAS and DM2 mathematics seem to be esoteric – addressing issues below the 90% or good enough level. Is this degree of precision really necessary?
90% level disambiguation and semantic precision works well for human-readable and interpreted data, as we fill-in missing information and bring to bear interpretive knowledge or for rehearsed automated processing -- when the programmers or database administrators can iterate and trial-and-error towards proper processing of the exchanged data. Automated processing by non-rehearsed algorithms, e.g., by integration or analysis algorithms by heretofore new data sources, can be very sensitive to flaws in datasets such as imprecisions, ambiguities, or unstated incompletions.
Who’s developing DM2 or IDEAS analysis tools?
This is just starting. We anticipate at least these two categories: M&S tools and entailment tools. M&S tools will be able to ingest DM2 datasets and, because of the precision and disambiguity afforded by DM2, be able to “run” or “execute” the architectural models to measure performance and/or effectiveness of proposed architectures. Entailment tools, some of which exist today and can operate on RDF, RDFS, and OWL datasets, will be able to carry out the logical implications of DM2 datasets whereupon contradictions and inconsistencies can be detected. For instance, an interoperability assessment tool might entail that two systems need to interact in some way (e.g., exchange data) but that is contradicted by all the means available to do so.
Go to top of page ↑
Why are there so many DM2 open action items?
The application of ontology science to Enterprise Architecture descriptions is new. There are still many things the community does not yet know how to represent mathematically. We could always fall back on language-based representations but we know that will result in improperly integrated or analyzed data. In other words, the DM2 Working Group is the forum for clearly defining and disambiguating EA concepts in the DoD. Membership in the DM2 Working Group is open to all by registering at www.silverbulletinc.com/dm2. The DoDAF / DM2 Action Item tracker is updated weekly and is available for download there.
Is the DM2 done?
Yes! Version 2.0 was baselined May 2009 and version 2.01, with 68 fixes and improvements, was baselined Feb 2010. However, as the DoD EA community seeks to represent additional things, as DM2 pilots and early adopters develop, and as concepts evolve – e.g., Services, Capabilities – the DM2 will respond to the community’s needs. This done by a formal Configuration Management (CM) process by the DM2 Working Group, a subordinate body to the Federated Architecture Council (FAC). Membership in the DM2 WG is open to all by registering at www.silverbulletinc.com/dm2.
How can the DoDAF / DM2 Working Group be effective if anybody is allowed to join and participate? Doesn’t it just become chaotic?
Remarkably, the WG is very effective even with 100’s of members. The reason for this is the process and business rules established and documented in the DoDAF / DM2 Configuration Management (CM) Plan which can be obtained at the DM2 Collaboration Site, www.silverbulletinc.com/dm2. Although the process and rules are subject to modification, once agreed-to, they provide a principled basis for discussion, debate, and analysis of potential issues.
Why didn’t DoD just adopt a commercial standard for EA data exchange?
Existing commercial data exchange formats do not meet the representation requirements for DoDAF architectural descriptions or are tool or tool-type specific. For instance, the XMI UML model interchange standard is specific to UML tools and consequently has many elements that are not applicable to non-UML tools. The DM2 Conceptual and Logical Data Models are the DoD expression of required data semantics for EA descriptions; the DM2 Physical Exchange Specification (PES) is a simple tool- and methodology-neutral format for EA data exchange between databases, repositories, EA development tools, EA analysis tools, authoritative data sources, EA reporting tools, and M&S tools.
Go to top of page ↑
Is there a way to represent metrics with DM2 and what kinds?
The DM2 represents metrics as in IDEAS as what may be thought of a “measure sets.” The DM2 defined several broad categories of metrics (measures) and then allows users to define as many additional types of measures as needed. Measures can be associated with any DoDAF concepts, e.g., Systems, Resource Flows, Capabilities, Desired Effects. Measures can be at a technical performance level (e.g., MOPs) up to operational effectiveness levels (e.g., MOEs).
How are temporal models handled in DM2?
DM2 is founded on IDEAS which is 4-dimensional. So all real-world things are modeled as per their spatial and temporal extent. In other words, everything in DM2 is temporal. DM2 and IDEAS have elements for temporal boundaries and before-after and temporal-whole-part relationships to model any form of temporal behavior. For more on 4-dimensionalism, see, e.g., http://en.wikipedia.org/wiki/Four-dimensionalism or a very popular book, Sider, Theodore; “Four Dimensionalism: An Ontology of Persistence and Time”; Oxford University Press, Oxford ; 2001.
How are Services modeled in DM2?
The DoD defines services as a mechanism to access capabilities (or resources.) The DM2 models Services as types of Performers that have a Service Port that is described by a Service Description. The Service has relationships to the Resources that are accessed. The Service Description is a type of Architectural Description and so it can have all the structure of an Architectural Description, including functionality, behavior, rules, information schemas, etc. New service concepts emerging from OASIS, OMG, and other organizations are being considered by the DM2 Working Group for incorporation in later DM2 baseline versions.
How can I implement the DM2?
All DM2 implementers should join the DM2 Working Group so they will be up to date on all developments, actions in-progress, and gain access to DM2 resources. In addition, the DoDAF / DM2 AS&I has resources to assist DM2 pilots and early adopters. Contact Mike.Wayson@osd.mil to request assistance in your pilot or implementation project.
I don’t see how to model requirements vs. solutions in DM2. Is it possible?
Yes! DM2 supports multiple levels of reification, indeed, as many as are needed by your project. Each level of reification is an architectural description that becomes rules that constrain the next level. Conversely, each artifact at a level can trace its pedigree to a higher level using DM2’s Pedigree model. With DoDAF 2, you are no longer restricted to just the OV “requirements” level and SV/TV “solution” level, but can have as many levels of reification as are needed, with each level having whatever mix of operational, capability, systems, services, or technical description as is appropriate for your project.