
Object Management Group Meeting (Reston, Va., March 2014) Report by Claude Baudoin (cébé IT & Knowledge Management) April 20, 2014 This report contains notes from sessions the author personally attended during the OMG Technical Meeting in Reston on March 24-28: the meeting of the Business Modeling and Integration Domain Task Force, the Business Architecture SIG, the Analysis and Design Task Force, the plenary lunch presentations, and the closing plenary report sessions. A comprehensive list of all the committees, task forces and working groups of the OMG can be found at www.omg.org/homepages/. A list of all the work in progress, with links to the corresponding materials (RFPs, etc.) is at http://www.omg.org/schedule/. Contents 1. Business Modeling and Integration Domain Task Force (BMI DTF) ......................................................... 2 2. Business Architecture Innovation Summit ............................................................................................... 7 3. Analysis & Design Task Force (ADTF) ..................................................................................................... 14 4. Plenary Lunch Presentations .................................................................................................................. 18 5. Plenary Report Sessions ......................................................................................................................... 18 6. Next Meetings ........................................................................................................................................ 25 OMG Reston Meeting Report Meeting Reston OMG Copyright © 2014 Object Management Group 1 1. Business Modeling and Integration Domain Task Force (BMI DTF) Claude Baudoin (cébé IT & Knowledge Management), Donald Chapin (Business Semantics Ltd), and Fred Cummins (Agile Enterprise Design) co-chaired the meeting. Claude reviewed the agenda. The meeting started with a series of presentations and discussions about SBVR (sections 1.1—1.4). 1.1. “Whither SBVR: Insights for SBVR 2.0 from Tool Development” by Ed Barkmeyer Ed Barkmeyer said that the Semantic of Business Vocabulary and Business Rules (SBVR) is now seven years old. It has some good and some bad features. The goal was to enable the interchange between tools, but this goal has been “lost in the shuffle.” SBVR is not used in the Decision Modeling Notation (DMN) or in other recent specifications that should have used it. One of the problems, according to Ed, is that the standard is written as a tutorial to use business vocabularies, not as a standard. Another issue is that almost everyone who uses SBVR is actually using SBVR Structured English (SBVR-SE), which was just supposed to be an example, described in an Appendix, and is not a normative specification. Ed pointed out some inconsistencies in the definition of properties and relationships. SBVR uses statements like “owner has car” and “car has owner” which are not equivalent: the first statement says that there is a relationship between two entities, and the second one says that a car has the property of having an owner – but the specification confuses two meaning of “to have.” Ed made many impassioned statements, and Don Chapin disagreed with several of them. Ed criticized the Revision Task Force for “rearranging the chairs on the deck of the Titanic.” He then proceeded to propose what changes need to be made: • Better semantic formulations, specifically in OCL • Improve the conformance criteria for tools that produce and consume exchange files • Remove certain “ontological commitments,” including SBVR’s handling of a time model • Mappings between SKOS (the W3C’s Simple Knowledge Organization System) and SBVR, and an RDF representation form in order to compete directly in the SKOS market • A metamodel for Structured English. Ed presented a sample metamodel, which he has implemented in a tool. The presentation is available at www.omg.org/members/cgi-bin/doc?bmi/14-03-01.pdf. 1.2. “SBVR: What It Is, What Its Future Should Be” by Donald Chapin Don Chapin, who chairs the SBVR RTF, presented his view of the specification. He said that SBVR “has a metamodel, but is not itself a metamodel.” SBVR addresses the existence of communities that share the meanings of certain concepts, the existence of subgroups that employ different preferred terms, and the use of specialized vocabularies, such as legal terminology. OMG Reston Meeting Report Meeting Reston OMG Copyright © 2014 Object Management Group 2 Donald described the process of “deconstructing” a regulation to extract the business vocabulary and rules. He showed and example from US anti-money laundering regulations, which were analyzed by GITC, a policy group within University College Cork. The first step is to formalize the compliance rules from the source regulation; the second step is to recognize the verb concepts in the formalized source. There was some discussion of when compound verb phrases need to be split into multiple verb concepts. Noun concepts fill roles in verb concepts. There was discussion about the distinction between concepts and “things in the real world.” There is a semantic triangle in SBVR to describe the relationship between concepts, terms, and things. A term like “US dollar” is the name of a concept (the US currency), but it is also used to designate the actual things (amounts of money, or dollar bills), which are not in SBVR but live in the real world of the business. Don described the changes being made to create version 1.3: • Complete the interpretation in formal logic, consistent with ISO Common Logic (ISO/IEC 24707) • Resequence the text to improve readability In addition, SBVR needs to be promoted to business people, especially people who deal with regulations. Don’s presentation is at www.omg.org/members/cgi-bin/doc?bmi/14-03-02.pdf. 1.3. “SBVR as a Meta-Meta-Model – Should SBVR be a Language for the Specification of Modeling Language?” by Fred Cummins Fred Cummins raised the question of whether SBVR could sufficiently align with the Meta Object Facility (MOF) to express all MOF concepts. Fred presented a potential mapping, which Donald and Ed said was incorrect; however, Fred said that the point was not the correctness of the mapping, but the idea that there could be a mapping. MOF is a form of business language, so it could in theory be expressed using SBVR. Fred’s slides are available at http://www.omg.org/members/cgi-bin/doc?bmi/14-03-03.pdf. Fred’s perspective comes from experience with “reflective languages” – languages that can describe themselves, and can operate on themselves. He sees a potential RFP for a “core reflective metamodel.” Donald and Ed agreed that what Fred proposes is essentially the same thing as the Semantic Information Modeling for Federation (SIMF), for which there is an RFP that has not progressed quickly. There was also an Architecture Ecosystem Foundation RFP in 2010, which competed with UML/MOF, did not generate sufficient market interest, and was not approved. Ed also pointed out that a complicating factor is that the SIMF RFP is handled in a Platform Task Force, while BMI is a Domain Task Force. In order to have a voice in both efforts, an organization would have to be a Contributing Member, which costs more. 1.4. Roadmap Discussion on SBVR Following the presentations of these often conflicting viewpoints on SBVR, a discussion was undertaken to establish a mutually agreeable roadmap. The discussion, captured on flip charts, is transcribed in the table below. OMG Reston Meeting Report Meeting Reston OMG Copyright © 2014 Object Management Group 3 Agreement • Don has canvassed tool vendors, and they support SBVR in their databases. Areas • There is a consensus that they would support a controlled natural language (syntax, not metamodel) – so this is an opportunity for an RFP. • They would support having an RDF interchange format – hence an opportunity for an RFP or RFC. • Many of the line items about SBVR shortcomings (Ed’s presentation) are agreed by the RTF and can be addressed by it. • A mapping to ISO Common Logic could replace the one that Terry did, based on work that’s 25 years old. o There is a question whether Clause 10 needs to be redone completely or not – this led to a “battle of the experts” about common logic. Semantic • Different positions: Model of SBVR o Don Chapin: we’re going to work on it in the RTF o Fred Cummins: this is beyond the scope of the RTF, we should have an RFP o Ed Barkmeyer (NIST), Fabien Neuhaus (University of Magdeburg): we need modal logic o Jishnu Mukerji (HP): who cares about it? (answer: GITC, for one) • Elie Abi-Lahoud (UCC): Develop mappings between a subset of the vocabulary and structure rules on the one hand, and first-order logic on the other o Agreement that this can be done by splitting the table in Clause 10 into two • Clarification of the exchange for and SBVR XMI metamodel o After discussion, it seemed to be an issue of clarity rather than substance; hence it needs to be captured as an issue for the RTF. Process • Should all this be done by the RTF (Don’s position) or via an RFP (Fred’s position)? Question o Jishnu: RTFs should be limited to bug fixes; RFPs should be used to make substantial changes. o Progress in the RTF requires consensus by the submitters of the original specification, while submissions to an RFP can be competitive and allow the entire task force to interact with those submitters. o Jishnu: you need to capture everything as issues, then do triage, and then see what
Details
-
File Typepdf
-
Upload Time-
-
Content LanguagesEnglish
-
Upload UserAnonymous/Not logged-in
-
File Pages25 Page
-
File Size-