UML Roadmap Progress report

1
UML RoadmapProgress report

• Steve Cook, Microsoft
• steve.cook_at_microsoft.com

ad/2008-12-13
2
This deck reports progress of the uml-roadmap
working group

• Constituted at Orlando meeting
• Participants Adaptive, Composable Logic, CSC,
  IBM, EDS, Embarcadero, Enterprise Agility,
  Everware-CBDI, JPL/NASA, MathWorks, Mentor
  Graphics, Microsoft, Model Driven Solutions,
  NIST, NoMagic, SAP, Bran Selic, SOFTEAM, Thales,
  Tibco
• Conference calls 14 October, 28 October, 11
  November, 25 November
• Numerous emails http//www.omg.org/archives/uml-ro
  admap/

3
Business problem

• High costs for implementing UML caused by
  specification complexity and ambiguity
• High business costs and reduced effectiveness
  resulting from the inability of UML to accurately
  express domains of interest to the user
• High business costs and reduced effectiveness
  resulting from the poor interoperability between
  tools, which limits the value derived from models
  expressed in UML
• High business costs of integrating UML models and
  profiles with other OMG models and profiles

4
Homework Top 3 strategic problems

• Separation of concerns (refactoring, unintended
  inheritance, separation of syntax/semantics/notati
  on, decoupling, reuse, conflict of goals) 19
  problems
• Interchange (interoperation, compliance) 8
  problems
• Semantics (defining meaning, execution, formal
  verification) 6 problems
• Other (new features, complexity metrics,
  methodology) 5 problems

5
Areas of Relative Agreement

• Need to address known issues in a timely fashion
• Need to clean up the UML definition to remove
  inconsistency and ambiguity and improve
  interoperability
• Need to refactor UML definition layer its
  semantics to enable UML vendors to embrace
  technology evolution
• Need to keep a high degree of compatibility with
  existing UML implementations, including at least
  widely-used profiles

6
Areas of Controversy

• Scope of UML which domains should be addressable
  by extending the UML metamodel
• How to use mechanisms for UML extension
  profiles, metamodel extensions, diagram
  modifications/extensions, semantics

7
Three-track approach

• Tactical RTF
• Submit issues resulting from roadmap group
  discussions to RTF
• RTF to categorize existing issues
• in-RTF-scope (tactical)
• out-of-RTF-scope (strategic)
• Prioritize tactical issues and address them in a
  timely fashion
• Strategic Architecture working group
• Create prototype refactored metamodels for areas
  of UML (eg support for C)
• Define architectural principles for refactoring,
  semantic layering, formalization and backward
  compatibility
• Validate architectural principles against
  strategic issues
• Process
• Issue RFI to poll community on the areas of
  disagreement business case, requirements, scope,
  definition problems, process
• Hold a vendor/user workshop to discuss the future
  of UML
• After RFI responses decide on how to proceed
  based on
• Out-of-scope issues, RFI response, strategic
  prototypes/principles, workshop content