1. © Copyright 2012 TopQuadrant Inc. 1 Investigating Semantics for Systems Engineering David Price, TopQuadrant dprice@topquadrant.com Allison Feeney, NIST Analysis Validating Maintaining Deploying Evolving Acquiring Modeling Architecting Integrating

2. © Copyright 2012 TopQuadrant Inc. 2 Project Motivation Make significant progress in MBSE areas of collaboration, knowledge management, data integration and timely decision making Major, rather than incremental, improvements in current practices might be enabled by applying Semantic Web technology This NIST-sponsored project is an investigation into that idea … while also supporting the reuse of and alignment with existing applications and standards such as SysML.

3. State of the Project  The current state of the project is a work in-progress –Have identified some key elements, but they are not yet pulled together into coherent whole  Very interested in sharing ideas to make the most of this investigation –Use cases that make sense –Info exchange with others looking at semantics for SE

4. © Copyright 2012 TopQuadrant Inc. 4 Big picture - A more complete solution for industry ERP PLM CRM Data Warehouse Unstructured Data (e.g. documents) Data on the Web Semantic SOA Adapto r WS Query Builder Results Processor Semantic Cache Query Dispatcher Query Orchestrator Rules Engine Triple Store Adapto r Web Services Controller WS Mapper Acquirer IT artifacts enable implementation, using industry best practice tools Enterprise vocabularies define business terms – think dictionary, not model Languages and models add formality to informal business terms, but core not tied to specific technology

5. © Copyright 2012 TopQuadrant Inc. 5 What technology is interesting? ERP PLM CRM Data Warehouse Unstructured Data (e.g. documents) Data on the Web Semantic SOA Adapto r WS Query Builder Results Processor Semantic Cache Query Dispatcher Query Orchestrator Rules Engine Triple Store Adapto r Web Services Controller WS Mapper Acquirer IT artifacts – SysML/UML metamodel & models (XMI), data models (e.g. AP233), XML Schemas, etc. Transforms – W3C SPARQL RDF Query Language & W3C submission SPARQL Inference Notation (SPIN) Enterprise vocabularies – W3C Simple Knowledge Organization System (SKOS) Language and model core – W3C Web Ontology Language (OWL) … use or relate to ontologies to put models on better foundation and enable cross- discipline communication

6. © Copyright 2012 TopQuadrant Inc. 6 IT Framework Investigation Scope

7. © Copyright 2012 TopQuadrant Inc. 7 Coordination Points  INCOSE –International Workshop Jan 21-24 2012 Jacksonville, FL –International Symposium, Rome, July 9 – 12, 2012 –Webinar?  OMG –March 19-23, 2012 Reston, VA –Future meetings?  ISO SC4 –IDIOM Framework, AP233 SE and AP239 PLCS  OASIS PLCS TC  NIST

8. © Copyright 2012 TopQuadrant Inc. 8 Project Deliverables (1)  INCOSE SE handbook as SKOS Concepts with definitions and relationships  OWL strong ontology SE example(s)  OWL as SE data interchange/integration format –AP233 and sibling standard PLCS which already use OWL for ‘reference data’ –‘proxy ontologies’ for XML Schema and AP233 EXPRESS  Links –between SKOS and OWL artifacts, E.g. class in OWL is defined by definition in SKOS –between SKOS and SysML, E.g. SysML Block traces to SKOS concept  Transforms between all using SPARQL/SPIN

9. © Copyright 2012 TopQuadrant Inc. 9 Project Deliverables (2)  Papers –How to use SKOS, OWL, SysML, XMI, AP233 EXPRESS together using IDIOM IT framework –SE strong ontology example and approach (creation and use, handling potential overlaps with other strong or proxy ontologies  Demonstrations to OMG and INCOSE –Use the IT artifacts in a small application highlighting the results of the investigation

10. 10 © Copyright 2012 TopQuadrant Inc. Investigations into SKOS for Systems Engineering Dictionaries

11. © Copyright 2012 TopQuadrant Inc. 11 Simple Knowledge Organization System  W3C standard built on RDF –A means of organizing and communicating terminology about a domain as Controlled Vocabulary, Taxonomy, Thesaurus, etc. –Allows different groups to organize terminology, while allowing them to link, using URIs, to one another “The St. Louis Office notion of ‘Customer’ is broader than the New York Office notion of ‘Customer’” You can’t even have this discussion, if you don’t recognize that there are actually two contexts for the word “Customer”!

12. © Copyright 2012 TopQuadrant Inc. 12 SKOS Intro  Takes concept-centric approach –Concept, not term, is the primitive object –Concept has preferred and alternative labels One preferred label per natural language –Concepts grouped into Concept schemes –Concept hierarchy defined using broader and narrower (transitive) relationships –Concept mappings exact match, close match, etc. –Definitions, notes and examples

13. © Copyright 2012 TopQuadrant Inc. 13 What do we get from that?  Clearly defined business concepts  Everything ‘in the Web’ or at least ‘on our Intranet’ so sharing, modularity and reuse are possible  Links to the future –The Internet of Things is coming - be prepared –Linked Data - Governments and others are publishing data we can use (e.g. detailed location and geography data)  Flexible solutions, designed for extensibility, that are cheaper to evolve –Blending, merging or aggregating new concepts and data so much simpler

14. © Copyright 2012 TopQuadrant Inc. 14 Initial SKOS Investigations  Starting simple - INCOSE Handbook Terms and Definitions –Incorporates some ISO/IEC 15288:2008 terms  ISO owns 15288 terms, INCOSE owns Handbook terms –These, however, need to be related  First step = pure SKOS –Populate two SKOS concept schemes, one for 15288 and one for INCOSE Handbook –Define World Wide Web-unique identifier for a term (called URI) –Add relationships as well as definitions –Publish as if an INCOSE artifact  Second step = add OWL –many approaches to SKOS use add a bit of OWL to improve organization of concepts

15. © Copyright 2012 TopQuadrant Inc. 15 A Web View in a SKOS Tool

16. © Copyright 2012 TopQuadrant Inc. 16 Zoom - Term and Definition

17. © Copyright 2012 TopQuadrant Inc. 17 Zoom (2) - Term Relations

18. © Copyright 2012 TopQuadrant Inc. 18 Follow Link to related ‘System’

19. © Copyright 2012 TopQuadrant Inc. 19 A Graph View in a SKOS Tool

20. 20 © Copyright 2012 TopQuadrant Inc. Investigations into Semantics for for Systems Engineering Data Models

21. © Copyright 2012 TopQuadrant Inc. 21 Data Model Plan  Investigate how semantic technology can be used to relate to data (exchange) models and to exchange and integrate data  ‘proxy ontologies’ is using OWL for data integration/exchange equivalent to – ISO AP233 –OASIS PLCS –other XML Schema-based standards  Relating proxy ontologies and strong ontologies –How to feed data provided in a proxy for AP233 into a strong ontology

22. © Copyright 2012 TopQuadrant Inc. 22 Constraints are an issue  Data exchange depends on constraints  Business rules are kinds of constraints  Traditional Description Logic reasoners over OWL cannot handle some simple constraints, due to Open World Assumption  Example: A Requirement has exactly one requirementId –Requirement with 2 requirementId values fails validation –Requirement with 0 requirementId values does not fail validation

23. © Copyright 2012 TopQuadrant Inc. 23 Use SPARQL for constraints  SPARQL is W3C standard RDF query language  SPARQL Inference Notation (SPIN) is W3C Member Submission making the use of SPARQL for constraints and transforms simpler # must be at least 18 years old ASK WHERE { ?this hasAge ?age. FILTER (?age < 18). } AdultPerson spin:constraint

24. 24 © Copyright 2012 TopQuadrant Inc. Investigations into Strong Ontologies for Systems Engineering

25. © Copyright 2012 TopQuadrant Inc. 25 Not a New Topic … 2006 paper

26. © Copyright 2012 TopQuadrant Inc. 26 Henson Graves Built Significantly On Earlier Work

27. 27 © Copyright 2012 TopQuadrant Inc. Systems (OWL DL) SysML User Requirements Company Systems Repository S-satisfies-R S-subClass-R (OWL DL) DL Reasoner Rule Engine Requirements (OWL DL) Simple Use Case

28. © Copyright 2012 TopQuadrant Inc. 28 Inference Layer RDF-izer Layer Models Transforms Reasoner Rule Engine Asserted Model Data Inferred Model Data Strong Ontologies based on Infers data The magic happens here based on More Technical View

29. © Copyright 2012 TopQuadrant Inc. 29 Applicable Technology SPARQL Inference Notation (SPIN) Description Logic Engine Transforms Rule Engine Reasoner SysML, Spreadsheets, Relational Databases, etc. turned into RDF data RDF-izer Layer Models Strong Ontologies Model Data OWL ontology and individuals

30. © Copyright 2012 TopQuadrant Inc. 30 RDF-izer Layer Relational XML RDF Spreadsheet UML RSS

31. Scenario  System component and configuration model with property ranges  Requirements based on properties  Specific systems available for search  Reasoner finds specific systems that satisfy requirements  Rule engine can augment cases where Description Logic reasoner is insufficient –DL does have limitations

32. 32 © Copyright 2012 TopQuadrant Inc. Sample Bike model in SysML

33. 33 © Copyright 2012 TopQuadrant Inc. Same model as OWL … right now the transform has been done manually.

34. 34 © Copyright 2012 TopQuadrant Inc. We can RDF-ize XMI files (because they are XML)

35. 35 © Copyright 2012 TopQuadrant Inc. Which means we can now SPARQL over SysML XMI files … so the SysML-to-OWL transform is doable using SPIN/SPARQL.

36. 36 © Copyright 2012 TopQuadrant Inc. An individual requirement statement about an individual property

37. 37 © Copyright 2012 TopQuadrant Inc. A requirement collection of other requirements statements about individual properties. Of course, combining collections into collection is fine as well

38. 38 © Copyright 2012 TopQuadrant Inc. A bike that satisfies the requirement becomes a subclass when a DL reasoner runs over the data (blue satisfies black).

39. © Copyright 2012 TopQuadrant Inc. 39 Strong Ontology Plan  Work with INCOSE and OMG SE DSIG to define use case(s) where reasoning and inference can be useful  Develop strong ontology examples  Investigate SysML models feeding into and getting feedback from strong ontologies  Publish paper  Demonstrations to OMG, INCOSE (Webinar?)  Publicize semantics benefit for Systems Engineering

40. © Copyright 2012 TopQuadrant Inc. 40 If these topics are of interest or relate to an existing effort, it would be great to share experiences Contact points: Allison Feeney at NIST David Price at TopQuadrant Thanks for your time