1 Extended Metadata Registry (XMDR) November 2004 Bruce Bargmeyer +1 (510) ISO/IEC JTC 1/SC 32/WG 2
2 Topics F XMDR project direction u What, when, who, how F XMDR relationship to WG 2 projects F A view of semantics based computing F Some specific semantics challenges F Some technology choices F A preview of issues to be raised for Parts 2 and 3 F Frank Olken will describe some of the content and ontology issues and approaches
3 XMDR Project Direction F Extend the capabilities of metadata registries to register complex metadata structures (concept structures, terminologies) u Ontologies, Graphs, Taxonomies, Thesauri, … This presentation will use the term “concept structures” as synonymous with “complex metadata structures” F Extend the capabilities of metadata registries to record correlations and interrelations between data (e.g., data elements & domains) and other concept structures. F Extend the capabilities of metadata registries to record correlations and interrelations between the various concept structures themselves.
4 XMDR & SC 32/WG 2 F Propose draft text for Part 2, Version 3 u Preview November 2004 WG 2 meeting u Proposals April 2005 WG 2 meeting F Propose issues for Part 3, Version 3 F Test & demo extended capabilities in a reference implementation u Tests & demo starting March 2005 u Register concept structures nominated by participants
5 XMDR Project Direction F Part of an Interagency/International Cooperation on Ecoinformatics u Sponsors & participants: EPA, EEA, USGS, DOD, NCI, Mayo Clinic F Extend semantics management capabilities for ISO/IEC u Produce design for next generation of operational ISO/IEC registries F Test & demo extended capabilities in a reference implementation F Research, develop, evaluate, adapt, extend, and demonstrate techniques and technologies for semantics based computing u Facilitate early adoption of these technologies F Establish best practices for semantic web and semantic based computing Forging Semantics Based Computing
6 Project Direction - Ecoinformatics Information science and information technology for the environment F Sound information as the basis for environmental policy, decisions, and action F Information technology that supports and enables development of sound information F Facilitate interaction with environmental information u Human - Computer u Computer - Computer
7 People Involved F XMDR Project at LBNL (Ecoinformatics +) u LBNL: Bruce Bargmeyer, Frank Olken, Kevin Keck, John McCarthy (ret. consulting) u DOD: Nancy Lawler & Sam Chance u EPA: Larry Fitzwater, Howard Tsai, Linda Spencer, William Sonntag u USGS: Gail Hodge (IIA, for USGS) (Lisa Zolly, USGS, is joining L8) u Mayo Clinic: Harold Solbrig u NCI: Sherri De Coronado, Denise Warzel F SC 32/WG 2, INCITS-L8 F Ashton Computing & Management: Judith Newton F Farance Inc. (consulting)
8 XMDR Liaison Activities F OMG Ontology Development Metamodel (ODM) F W3C Semantic Web Best Practices and Deployment Working Group F Ecoterm F EU Joint Research Center (EDEN-IW) F National Science Foundation (Ecoinformatics) F Interagency/International Cooperation on Ecoinformatics
9 Project Direction F What the project is not: u An attempt to make metadata registries be a development and maintenance facility for every type of concept structure u An attempt to standardize the complete range of terminology servers
10 XMDR Relationship to WG 2 Projects - Introduction Real World Modeling Tools Model Artifacts and Exchange Applications Methodologies/tools: EDR, NIAM, O-O, RDF, UML, Ontology CDIF, (UML: MDL, XMI), OWL, CL & SCL (KIF) SQL (relational), Object, Semantic Web (might be agent based, grids, etc. Metamodel constructs: CDIF Core, MOF, XML-Schema, RDFS, ODM, OWL, Common Logic (CL)
Semantics Management Real World Modeling Tools Model Artifacts and Exchange Applications Methodologies: EDR, NIAM, O-O, RDF, UML, Ontology CDIF, (UML- MDL, XMI), OWL, CL & SCL (KIF) SQL, Object, RDF, Semantic Web Metamodel constructs: CDIF Core, MOF, XML-Schema, RDFS, ODM, OWL, CL MDR Semantics Management: Data elements, Domains, Concepts, Terms …
12 ISO ANSI Industry Gov’t Country Country changes CZ Czechoslovakia* CZ Czech Republic** LO Slovakia** Change MDR – Keeping Track of the Real World
13 SC 32 Standards & Projects Real World Modeling Tools Model Artifacts and Exchange Applications Methodologies: EDR, NIAM, O-O, RDF, UML, Ontology CDIF, (UML- MDL, XMI), OWL, CL & SCL (KIF) SQL, Object, RDF, Semantic Web WG 3 - SQL Metamodel constructs: CDIF Core, MOF, XML-Schema, RDFS, ODM, OWL, CL WG 2 - MMF (19763), CL (24707) & MOF PAS submission WG 2 - MMF (19763), CL (24707) & XMI PAS submission WG 2 – WG MDR Semantics Management: Data elements, Domains, Concepts, Terms … WG
14 XMDR Focus Methodologies: EDR, NIAM, O-O, RDF, UML, Ontology CDIF, (UML- MDL, XMI), OWL, CL & SCL (KIF) SQL, Object, RDF, Semantic Web WG 3 - SQL Metamodel constructs: CDIF Core, MOF, XML-Schema, RDFS, ODM, OWL, CL WG 2 - MMF (19763), CL (24707) & MOF PAS submission WG 2 - MMF (19763), CL (24707) & XMI PAS submission XMDR project Real World Modeling Tools Model Artifacts and Exchange Applications WG 2 – Parts 2 & 3 WG MDR Semantics Management: Data elements, Domains, Concepts, Terms … WG
15 A Current Example Real World Modeling tool Model artifact/ exchange Application OWI Knowledge Server Application Ontology Works Inc. (OWI) IODE data modeling tool Domain ontology expressed in Simple Common Logic (based on Draft ISO/IEC MDR Semantics Management: Data elements, Domains, Concepts, Terms …
16 Another Current Example Real World Modeling tool Model artifact/ exchange Application OWI Knowledge Server Application, possibly built on Objectivity as the persistent object Store (DBMS) Protégé ontology tool Domain ontology expressed in As an OWL ontology MDR Semantics Management: Data elements, Domains, Concepts, Terms …
17 Semantics Based Computing F What is it? F Evolution of semantics management F Evolution of technologies that utilize semantics
18 Semantics based computing F Computation based on the meaning of data rather than on the manipulation of syntactic structures.
19 Semantic Mapping
20 Metadata Registries Semantics Management Evolution Initial “data standards”, evolved to stronger semantics management F Common data across information systems (data standards) F Database (schema) integration F Data use - metadata F Warehouse support – schema and metadata F XML support (schema) F “Backed into” concept/terminology support (deeper semantics) F Next: Semantics servers -- for semantic web and semantics based computing
21 Past, Present, … Future? Lots of users Lots of information systems Lots of Data Sources Users EEA DOE DoD EPA environ agriculture climate human health industry tourism soil water air textdata environ agriculture climate human health industry tourism soil water air text ambiente agricultura tiempo salud hunano industria turismo tierra agua aero textdata environ agriculture climate human health industry tourism soil water air textdata Others... ambiente agricultura tiempo salud huno industria turismo tierra agua aero textdata
22 Data Standards F Avoid a combinatorial explosion of data content, description, and metadata arrangements for information access and exchange. Data standards and metadata registries can help.
23 Data Element Concept Afghanistan Belgium China Denmark Egypt France Germany ………… Data Elements AFG BEL CHN DNK EGY FRA DEU ………… ISO 3166 English Name ISO Numeric Code ………… ISO Alpha Code Afghanistan Belgium China Denmark Egypt France Germany ………… Name: Context: Definition: Unique ID: 4572 Value Domain: Maintenance Org.: Steward: Classification: Registration Authority: Others Name: Context: Definition: Unique ID: 3820 Value Domain: Maintenance Org.: Steward: Classification: Registration Authority: Others Name: Context: Definition: Unique ID: 1047 Value Domain: Maintenance Org.: Steward: Classification: Registration Authority: Others Name: Country Identifiers Context: Definition: Unique ID: 5769 Conceptual Domain: Maintenance Org.: Steward: Classification: Registration Authority: Others
24 AFG BEL CHN DNK EGY FRA DEU ………… ………… Afghanistan Belgium China Denmark Egypt France Germany …………
25 State Laws CAA CWA RCRA TSCA “ State Regs Fed Air Reg Fed Water Reg Fed RCRA Reg Fed TSCA Reg “ Separate Data Repositories Regulated Facility Separate Regs/ Procedures Separate Environmental Media Legislation Then there is one point of access to our environmental data resources: Complete Warehouse Repository Regulated Facility Public/ Environmental Regulators/ Environmental Community June 1996
26 Data and Semantics Management Dictionary Keyword Ontology Terms Terms Data Elements Thesaurus DBMS/XML/ Documents Semantic Web Concepts
27 ISO/IEC Metadata Registries Metadata Registry Terminology Thesaurus Themes Data Standards Ontology GEMET Structured Metadata Evolving toward stronger semantics management
Metadata Registries Companies Universities Agencies Data Services Semantic Services Others Users September 2004 Environmental Data Grid Environmental Computer Grid High Performance, cluster, Personal Environmental Semantics Grid Terminology Thesaurus Ontology Taxonomy Structured Metadata Computation Services Software: Models, Visualization, Analysis Agent systems Semantic Based Computing Data Standards
29 Metadata Registries Companies Universities Agencies Data Services Semantic Services Others Users September 2004 Environmental Data Grid Environmental Computer Grid High Performance, cluster, Personal Terminology Thesaurus Ontology Taxonomy Structured Metadata Computation Services Software: Models, Visualization, Analysis Agent systems Semantic Based Computing Data Standards Environmental Semantics Grid
30 What is it ? [2] Semantics based computing: Applications that take the meaning of data into account to direct the processing. F Establish linkage between concepts referenced in text and related data in databases F Semantic Web F Support agent-based development of actionable data, for informed decision making.
31 Some Challenges F Translate the UML model into an ontology, manually. F Translate the UML model into an ontology, automated. F Identify emerging technology for building reference implementation, develop architecture F Identify test concept structures and sources F Characterize concept structures F Identify extensions needed for F Propose extensions needed for 11179
32 Manual Translation UML to an Ontology F Use Protégé tool and OWL specification F Frank will show tell all about it
33 Automated Translation UML Metamodel to an OWL Ontology Part 3 metamodel as Rational Rose UML MDL file
34 ISO/IEC Expressed as an Ontology <rdf:RDF xmlns:rdf=" xmlns:rdfs=" xmlns:owl=" xmlns=" xml:base=" <owl:cardinality rdf:datatype=" >1
35
36 Potential Standards/Technologies F DBMS u Object, XML, Relational, RDF/Graph, Logic, Text, Document, Multimedia F Knowledge Representation u Web Ontology Language (OWL) u Simple Common Logic (SCL) F Middleware/Messaging u Cocoon 2, Jini, CoABS, JMS, XMLBlaster, SOAP F XML [Semantic] Web Services u Axis, JWSDP F Agent Development u ABLE, JADE F Engines/Servers u OMS (IBM), Federator/OMS (OWI) u Jess
37 Content and Content Characterization A B C BB CCCC Directed Acyclic Graphs, Cyclic, Undirected, … Frank Olken to tell about this.
38 Preview Suggested Changes for P2/P3 (cont.) F Issue 1. Make "relation" an administered item. The relationship could be managed as part of the structure in which they are involved. Alternatively, in Clauses 4.10 and 4.11, possibly treat the subject role as an aggregate association. This is an alternative way of administering relationships, more in line with current practice.
39 Preview Suggested Changes for P2/P3 (cont.) F Issue 2. Rename the "horizontal" role and association names in Clause 4.7.3, Figure 3. E.g., Value_Domain should not have the role "representing" going in two directions. The association name "data_element_representation" may be impacted by the change in role names. Note that the role between the top two boxes is labeled "having" and "specifying", while the role between the bottom two boxes is labeled "represented_by" and "representing". The relationship between the upper two boxes and bottom two should be symmetric. Also, "having" could better be "specified_by". (We also have possible alternate proposals for labels.)
40 Preview of Issues (cont.) F Issue 3: Identify the types of correspondences between concepts. The point is to record the types of overlap between the concepts. An alternative is translation tables, which record pairs of IDs linking concepts without any more specific "type" information.
41 Preview Suggested Changes for P2/P3 F Issue 4. Directed Relationships as replacement to "association" and "related to" in Clauses 4.10 and Note that in Clause 4.11, "concept_relation" is directed, but no inverse is specified. F Issue 5. Replace the "string" value to a relation instance. This applies to u Clause 4.10: clasification_scheme_item_relationship_type_description u Clause 4.11: data_element_concept_relationship_type_description concept_relationship_type_description
42 Next Year: Proposed F Service Oriented Architecture
43 Eighth International Open Forum on Metadata Registries Semantic Interoperability: Where Meaning Meets Metadata. Open Forum 2005 April 11-14, 2005 Berlin, Germany Berlinopenforum.de